Search

Aim: To assess the image quality of deep-learning image reconstruction (DLIR) of chest computed tomography (CT) images on a mediastinal window setting in comparison to an adaptive statistical iterative reconstruction (ASiR-V).Materials and Methods: Thirty-six patients were evaluated retrospectively. All patients underwent contrast-enhanced chest CT and thin-section images were reconstructed using filtered back projection (FBP); ASiR-V (60% and 100% blending setting); and DLIR (low, medium, and high settings). Image noise, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) were evaluated objectively. Two independent radiologists evaluated ASiR-V 60% and DLIR subjectively, in comparison with FBP, on a five-point scale in terms of noise, streak artefact, lymph nodes, small vessels, and overall image quality on a mediastinal window setting (width 400 HU, level 60 HU). In addition, image texture of ASiR-Vs (60% and 100%) and DLIR-high was analysed subjectively.Results: Compared with ASiR-V 60%, DLIR-med and DLIR-high showed significantly less noise, higher SNR, and higher CNR (p<0.0001). DLIR-high and ASiR-V 100% were not significantly different regarding noise (p=0.2918) and CNR (p=0.0642). At a higher DLIR setting, noise was lower and SNR and CNR were higher (p<0.0001). DLIR-high showed the best subjective scores for noise, streak artefact, and overall image quality (p<0.0001). Compared with ASiR-V 60%, DLIR-med and DLIR-high scored worse in the assessment of small vessels (p<0.0001). The image texture of DLIR-high was significantly finer than that of ASIR-Vs (p<0.0001).Conclusions: DLIR-high improved the objective parameters and subjective image quality by reducing noise and streak artefacts and providing finer image texture. [ABSTRACT FROM AUTHOR]

• ECG data have been exploited for the detection of cardiovascular morbidity in two-dimensional space. • The sub-images have been constructed by projecting the 12-ECG signal vector as a row of the image. • Cardiovascular defects have been detected with the classical image texture methods. • Two-dimensional features in the images contain determinative indicators of various diseases. • The proposed system has provided fast and largely accurate results. Electrocardiography has been employed successfully in medicine for many years to provide vital knowledge about the cardiovascular system. Although processing and evaluation of electrocardiogram (ECG) signals provide helpful information in the detection of anomalies in the vessel, diagnosis of heart defect, and treatment of diseases, multi-channel ECG signals have been started to be employed in order to achieve higher success. Utilizing a multi-channel ECG signal instead of a one-channel ECG signal yields more adequate achievements but require higher complexity in analysis and higher computational cost. To achieve faster and accurate results in multi-channel ECG signals, an artificial intelligence-based automatic diagnosis system employing the texture features of two-dimensional images, which are constructed by projecting the ECG signal vector as a row of the image, is proposed. The hypothesis proposed in this study conjectures that these texture features in the images contain determinative indicators of various diseases (cardiovascular abnormalities/disturbance) even for the short-time intervals. Accordingly, the main contribution of this study is to expose that detection of cardiovascular defects can be done with the classical image texture methods by utilizing multi-channel biomedical signals in a sufficiently short-time-interval. The methodology has been implemented in different time intervals of a large dataset constructed from a diverse population that is labeled as one normal sinus rhythm type and eight abnormal types of ECG signals. The accuracy of this hypothesis has been proven by achieving high detection rates of identifying cardiac abnormalities and reduced the computational load of the processing system without any sacrificing accuracy. [ABSTRACT FROM AUTHOR]

Farm monitoring and operations generate heterogeneous AGRI-data from a variety of different sources that have the potential to be delivered to users ‘on the go’ and in the field to inform farm decision making. A software framework capable of interfacing with existing web mapping services to deliver in-field farm data on commodity mobile hardware was developed and tested. This raised key research challenges related to: robustness of data steaming methods under typical farm connectivity scenarios, and mapping and 3D rendering of AGRI-data in an engaging and intuitive way. The presentation of AGRI-data in a 3D and interactive context was explored using different visualisation techniques; currently the 2D presentation of AGRI- data is the dominant practice, despite the fact that mobile devices can now support sophisticated 3D graphics via programmable pipelines. The testing found that WebSockets were the most reliable streaming method for high resolution image/texture data. From our focus groups there was no single visualisation technique that was preferred demonstrating that a range of methods is a good way to satisfy a large user base. Improved 3D experience on mobile phones is set to revolutionize the multimedia market and a key challenge is identifying useful 3D visualisation methods and navigation tools that support the exploration of data driven 3D interactive visualisation frameworks for AGRI-data. [ABSTRACT FROM AUTHOR]

Aiming to identify plants, we propose to evaluate the color texture of the palisade parenchyma, from microscopic images of leaf cross-sections, using a graph based approach. Our texture analysis approach models the image texture as a graph and uses measurements computed from the shortest paths between specific vertices to provide a feasible texture signature. For a more consistent evaluation, we compared our approach to different methods for color texture analysis in a texture classification experiment. The results obtained indicate that our approach is the most suitable for this histological analysis as it surpassed all the other texture approaches using Linear Discriminant Analysis, and obtained the second best accuracy using 1-Nearest Neighbor. These results also corroborate the feasibility of using both histological (as the palisade parenchyma) and computer analysis for identification and delimitation of plant taxa. [ABSTRACT FROM AUTHOR]

There are different research and processing methods for Image Content Retrieval. This article is described the retrieval of remote sensing images method, which is based on the image texture features of the Fourier power spectrum. Firstly, a fixed size remote sensing image unit is selected from the remote sensing image, which is converted to a spectral image by FFT, and extract the eigenvalues in the frequency domain to constitute a searchable image eigenvalue database. Secondly, the similarity between the feature and the database features are examined by the Angular Cosine Algorithm. Lastly, similarity of the resulting image are sorted to select the best match. This method converts the remote sensing image into a spectral image, and makes the selection of the eigenvalue more simple. The Angle Cosine similarity detection method with weight is used to make the similarity detection more accurate. [ABSTRACT FROM AUTHOR]

Nanomechanical properties of fruit determine macroscopic firmness. Young's modulus (YM) of apple tissue obtained by atomic force microscopy (AFM) at cellular level and its correlation with other macroscopic physical parameters are used to evaluate ripening when apples were stored at 25 °C during 40 days. The YM of the tissue decreased from 0.96 ± 0.42 MPa (first day) to 0.11 ± 0.06 MPa (last day). The best linear correlation of YM was obtained with firmness (F) which decreased from 21.15 ± 0.79 N to 12.74 ± 0.34 N. Also various physical parameters obtained as peel color difference (ΔE) from 57.80 ± 0.97 to 66.51 ± 0.71, environmental electron microscopy (ESEM) such as entropy: Ent (from 8.67 ± 0.12 to 9.60 ± 0.17) and fractal dimension: FD (from 2.67 ± 0.05 to 2.75 ± 0.03) changed as well. Significant correlations ( P < 0.05) were found between YM, F, ΔE, Ent and FD using Pearson analysis. Predictive models to evaluate YM from F, ΔE, Ent, FD were obtained by multiple linear regression, R 2 > 0.95 was found. Industrial relevance The study of the nanomechanical properties of fruit cells by AFM may provide insight into internally fruit properties and how changes in these properties over time influence the quality of the fruit. The determination of cell/tissue mechanics could be used to follow changes in the nanomechanical properties that occur during the processing and storage of climacteric fruits such as apples. The study of the nanomechanical properties of plant cells as well as their correlation with other useful food properties such as firmness, peel color and image texture could lead to a better understanding of the ripening process. Mathematical models to predict cell mechanical properties at the nanometric level through food quality parameters could be an innovation in food engineering and be a novel tool in evaluating the quality of apples. Benefits of the methods herein could be extended to address current issues, such as extending the storage life of other climacteric fruits and predicting nanostructural modifications to the cells when they are physically modified or chemically treated. [ABSTRACT FROM AUTHOR]

Characterizing the dynamics of food oral breakdown is of interest to understand the temporal perception of food products. The present work aimed at studying the possible contribution of artificial vision for studying bolus formation. Four emulsion-filled gels were prepared from two concentrations of agar and gelatin. By combining two different layers of these gels, four samples of homogeneous composition and 6 samples of heterogeneous composition were prepared. The layers were colored independently in order to study their breakdown and mixing during oral processing. Images of spat out boluses were collected at different stages of the chewing process and studied by different image analysis methods: gray-level histograms, histogram of shape area, mathematical morphology and gray level co-occurrence matrix. Methods were compared for their ability in discriminating boluses as function of homogeneous gel composition and mastication time. Three methods were found to be relevant and mathematical morphology provided the best results. Using this method, we further analyzed the impact of heterogeneous gels composition on the evolution of boluses properties. Results showed that when two gel layers of different composition were combined, the agar layer dominated bolus properties and that the presence of gelatin impacted the dynamics of gel breakdown. The results were in agreement with results obtained previously when characterizing the physical properties of the boluses. This study showed that artificial vision provides reliable tools for evaluating the dynamics of bolus formation, which is complementary to the methods commonly used in literature while avoiding extensive manipulation of boluses. [ABSTRACT FROM AUTHOR]

In recent years, single image dehazing methods have achieved good performance. But there still exists some problems, such as color distortion, excessive smoothing of image texture details and image over-saturation. For solving these problems, this paper proposes an improved single image dehazing algorithm based on Color Balancing and Quad-Decomposition (CBQD). In the pre-processing stage, we define an indicator for detecting image unbalance color channel, and correct color of images that are prone to color distortion in some special environments (e.g., sandstorms). In the dehazing phase of the algorithm, we apply the improved quad-decomposition technique for eliminating the interference of non-atmospheric light source and certain bright objects, thus can estimate atmospheric light accurately. Meanwhile, we obtain the image depth information by the linear model and estimate the initial image transmission map by fused depth map, and then refine the transmission map using a gradient domain-guided filter with edge-preserving characteristics. Finally, we can restore the scene radiance, and the dehazed images have clear edges, rich texture, no halo artifacts, and natural color. Extensive experiments verifies the superiority of CBQD over state-of-the-art image dehazing techniques in terms of both the recovery quality and efficiency. [ABSTRACT FROM AUTHOR]

• Embedding multiple watermarks method based on inter-block correlation is proposed. • The texture complexity method is designed to adaptively select adjacent blocks. • Multiple watermarks are embedding in the same region of the image in parallel. • Two methods of adjacent block selection are proposed for image adaptive selection. • The watermark is embedded alternately in the middle frequency and low frequency. People pay more and more attention to copyright protection and digital watermarking technology is a reliable method for copyright identification. In this paper, we propose the parallel multiple watermarking method using adaptive inter-block correlation. Considering the image texture characteristics, the texture complexity method is designed, based on which, the Circular-shaped Adjacent Block Selection and the Arch-shaped Adjacent Block Selection will be adaptively selected for adjacent block selection. The multiple watermarks are alternately embedded into the low and middle frequency band in DCT domain of the image blocks to good imperceptibility and high robustness. By adjusting the difference of projection of embedding coefficients and reference coefficients on the spreading vector, the multiple watermarks can be embedded, so the embedding capacity increases. In addition, we improve the multiple watermarks embedding based on orthogonal spreading vectors so that it can be adapted to embedding multiple watermarks in parallel based on inter-block correlation, saving embedding and extraction time. We theoretically analyze the rationality of parallel orthogonal embedding of multiple watermarks and experimentally verify the high efficiency of parallelism. Cover images of different texture complexity are included in the test dataset, and various attacks have been simulated in the experiments. Experimental results on imperceptibility, robustness, time cost and watermark capacity show the satisfied performance of the proposed scheme. [ABSTRACT FROM AUTHOR]

Alzheimer's disease (AD) is a neurodegenerative disease characterized by cognitive and behavioral impairment that significantly interferes with social and occupational functioning. Mild cognitive impairment (MCI) is a relatively broad clinical condition involving a slight memory deficit, which in many cases represents a transitional state between a cognitively normal (CN) condition and AD. Structural magnetic resonance (sMR) imaging has been widely used in studies related to AD because it provides images with excellent anatomical details and information about structural and contrast changes induced by the disease in the brain. Many published studies restrict their analysis to a few particular regions of the brain and search for structural changes caused by the disease. Recent studies start looking for new AD biomarkers using multiple brain regions and focusing on subtle texture changes in the image. Therefore, this study proposes a new technique for MR image classification in AD diagnosis using graph kernels constructed from texture features extracted from sMR images. In our method, we first segment the MR brain images into multiple regions with the FreeSurfer. Then, we extract 22 texture features using three methods and define the graph-node attributes as the probability distributions of the extracted features. Next, for each texture feature, we build a graph and define its edge weights as the distances between pairs of node attributes using three distance metrics. After that, we use a threshold-based approach for graph edges removal and create the graph-kernels matrices. Finally, we perform image classification using Support Vector Machines (SVMs) with two graph-kernels. Results of our method have shown better performances for the CN × AD (AUC = 0.92) and CN × MCI (AUC = 0.81) classifications, and worse for the MCI × AD case (AUC = 0.78). This trend is consistent with other published results and makes sense if we consider the concept of Alzheimer's disease continuum from pathophysiological, biomarker and clinical perspectives. Besides allowing the use of different texture attributes for the diagnosis of Alzheimer's, our method uses the graph-kernel approach to represent texture features from different regions of the brain image, which considerably facilitates the image classification task via SVMs. Our results were promising when compared to the state-of-the-art in graph-based AD classification. • Alzheimer's disease classification based on graph kernel Support Vector Machines. • Evaluation of image texture patterns of multiple brain regions in Alzheimer. • Assessment of three distance metrics to measure node attributes differences. • Threshold based method for graph edge removal to obtain discriminative graphs. • Classification results comparable to other proposed graph-based models. [ABSTRACT FROM AUTHOR]

• The proposed novel RED-MAM model for ultrasound image denoising, which can obtain richer features from images, and maintain the diversity of image feature information. • The proposed multi-attention fusion attention block is composed of multiple attention modules to focus on the importance of information on multiple feature domains by fusing different attention features. In the multi-attention fusion attention block, jump connections are used multiple times to avoid the loss of key feature information during reconstruction. • The denoising performance of the RED-MAM model is validated on the FH, HC18 and CAMUS ultrasound datasets. The result proves that the improved denoising model can effectively remove noise from ultrasound images. The proposed method has better performance on the fetal cardiac ultrasound dataset in noise suppression and structure preservation compared with other classic image denoising algorithms like BM3D, K-SVD, CNN10, etc. Compared with the state of art method RDN10, our method improved by 0.1 dB on the average peak signal-to-noise ratio (PSNR), improved by 0.5 dB compared with BM3D. The dependence of speckle noise in ultrasound images on image data makes the research of ultrasound image denoising tasks a great challenge. Several deep learning-based ultrasound image denoising methods have been proposed, but almost all of them suffer from the inability to focus on the importance of feature information on multiple domains at the same time. Therefore, this paper proposes a residual encoder-decoder based on multi-attention fusion attention module (RED-MAM) for ultrasound image denoising, which consists of five convolution layers, five deconvolution layers and two multi-attention fusion attention blocks. In order to make the denoising model better adapted to the properties of speckle noise, a multi-attention fusion attention block is constructed by several different attention modules. The multi-attention fusion attention block is introduced into the residual encoder-decoder denoising network to make the model focus on the ultrasound image texture structure information on multiple domains, thus enhancing the denoising effect of the model on ultrasound images. The performance of our proposed model is evaluated on three ultrasound image datasets. In terms of quantitative metrics, RED-MAM shows substantial improvement in Peak Signal-to-Noise Ratio (PSNR), Structural Similarity (SSIM) and Root Mean Square Error (RMSE). The qualitative results show that RED-MAM has significantly improved denoising performance and has good results in noise suppression as well as structure retention. Our method achieves state-of-the-art performance on all three ultrasound datasets. [ABSTRACT FROM AUTHOR]

A World Health Organization (WHO) report estimates that in 2015, at least 561 thousand women will die of breast cancer. Although breast cancer is considered a disease of the developed world, nearly 50% of the cases and 58% of the deaths occur in the less developed countries. A mammogram is a way of discovering not just the palpable tumors that cause cancer but also other lesions that are not perceived during the physical examination performed by the expert physician or during self-exams; however, it is known that this exam is targeted for women after the age of 40 because age is one of the factors that can cause great variations in sensitivity during the exam. Besides the patient’s age, the expert’s experience and the quality of the images obtained during the exam are decisive factors in the detection of breast cancer. This work presents two novelties. The first is the use of Local Binary Patterns (LBPs) to generate a representation of a Region of Interest (ROI) image. Over this representation, we generate other representations using techniques such as image histograms, gray-level co-occurrence matrices (GLCMs) and gray-level run-length matrices (GLRLMs). These representations allow texture analysis through several perspectives. The second novelty uses these representations as input to the application of indexes adapted from ecology (Shannon, McIntosh, Simpson, Gleason and Menhinick) as texture descriptors. Based on this strategy, we analyze mammographic image textures to classify regions of these images as benign or malignant using a Support Vector Machine (SVM). The best result achieved was of 88.31% accuracy, 85% sensitivity, 91.89% specificity, a positive probability ratio of 10.48, a negative probability ratio of 0.16, and an area under the Receiver Operating Characteristic (ROC) curve of 0.88, obtained through the Shannon index. We believe that the proposed method, with some adaptations, may also be used for image texture analysis of several different lesions such as lung nodules, glaucoma and prostates. This belief is based on the achieved results and the method’s simplicity. [ABSTRACT FROM AUTHOR]

• Effects of colours on the surface texture features are studied. • Colour of incident light significantly affects surface texture features. • Predicted results are in good correlation with the measured values. • Correlation between measured and predicted values is higher under blue illumination. A machine vision system is developed to capture and store the images of a 3-D printed work piece under different colours of incident light. The main objective of the present work is to study the effects of colours of lights on the surface roughness of the work piece. The surface roughness and image texture features such as correlation, contrast, energy, homogeneity and entropy of the work piece are the key parameters in this study. A stylus type instrument is used to measure the surface roughness of the 3-D printed work piece. Gray level co-occurrence matrix (GLCM) is introduced to extract the image texture features and an artificial neural network (ANN) is used to predict the surface roughness. The results are compared and the correlations are estimated. [ABSTRACT FROM AUTHOR]

• An image texture measurement method using differential box counting based fractal dimension (FD) is proposed. • It uses a new formula for computing the number of boxes to measure the grid surface roughness. • It solves the over-counting and under-counting problems along the z and xy-directions respectively. • The above two problems are solved using a partitioning-shifting-partitioning mechanism. Texture analysis methods have been used in various image processing applications, such as segmentation, classification, and recognition, shape analysis. The performance of computer vision and image processing algorithms depend on visual texture patterns. However, it is difficult to define. Fractal Dimension (FD) helps to measure the level of roughness present in an image. A significant amount of work is available in literature to measure FD. Differential Box Counting (DBC) is one such successful method exploited in image texture analysis study, because it is simple and easy to understand. DBC is modified several times to yield a better FD value. However, most of the state-of-the-art methods suffer from the over-counting and under-counting of boxes, larger box-height, less number of grid sizes, inappropriate line fitting method. This work introduces a gray-level shift-invariant DBC method, which uses a new formula for counting boxes along z -direction to solve over-counting of boxes, a partitioning-shifting-partitioning mechanism to fix under-counting of boxes along xy -direction, a smaller box-height to enhance the FD value and robust least squares regression to estimate a line accurately. All the experiments are performed on synthesized Fractal Brownian Motion image dataset and four real datasets. The obtained results illustrate that the proposed DBC method outperforms state-of-the-art DBC methods. [ABSTRACT FROM AUTHOR]

• Super resolution reconstruction algorithm of infrared image based on pseudo texture transfer is proposed. • The proposed method mainly includes two parts: pseudo infrared image generation and infrared image reconstruction based on pseudo texture transfer model. • A pseudo image generation network is proposed to convert high-resolution visible images into pseudo infrared images. • By constructing the model PT2SR, the restoration ability of image texture details generated by super-resolution network is improved. High-quality infrared images are limited by optical diffraction limits, device material, and technology with a high acquisition cost. Improving image quality with super-resolution reconstruction is an important way to obtain high-resolution infrared images. Based on pseudo texture transfer, an infrared image super-resolution reconstruction algorithm is proposed in this paper. First, an image-to-image transfer learning framework is used to generate pseudo-infrared images from the high-resolution visible image set. Second, the feature maps of pseudo-infrared and low-resolution infrared images are extracted simultaneously, and the feature matching module is designed to calculate the similarity of feature maps with different layers between pseudo-infrared and low-resolution infrared images. The local feature maps with the maximum similarity replace the feature map of low-resolution infrared images. Finally, the original low-resolution image and the corresponding exchanged feature map are used in the proposed feature transfer model to gradually reconstruct the super-resolution image from high level to low level. Experimental results prove that the proposed method has improved performance in terms of quantitative indicators and visual quality. [ABSTRACT FROM AUTHOR]

Automated ultrasound breast imaging would overcome most of the limitations that precludes conventional hand-held echography to be an effective screening method for breast cancer diagnosis. If a three dimensional (3D) ultrasound dataset is acquired without manual intervention of the technician, repeatability and patient follow-up could be improved. Furthermore, depending on the system configuration, resolution and contrast could be enhanced with regard to conventional echography, improving lesion detectability and evaluation. Having multiple modalities is another major advantage of these automated systems, currently under development by several research groups. Because of their circular structure, some of them include through-transmission measurements that allow constructing speed of sound and attenuation maps, which adds complementary information to the conventional reflectivity B-Mode image. This work addresses the implementation of the Acoustic Radiation Force Impulse (ARFI) imaging technique in a Full Angle Spatial Compound (FASC) automated breast imaging system. It is of particular interest because of the high specificity of ARFI for breast cancer diagnosis, by representing tissue elasticity differences rather than acoustic reflectivity. First, the image formation process is analyzed and a compounding strategy is proposed for ARFI-FASC. Then, experimental results with a prototype system and two gelatin phantoms are presented: Phantom A with a hard inclusion in a soft background, and phantom B with three soft inclusions in a hard background and with three steel needles. It is demonstrated that the full angle composition of ARFI images improves image quality, enhancing Contrast to Noise Ratio (CNR) from 4.9 to 20.6 and 3.6 to 13.5 in phantoms A and B respectively. Furthermore, this CNR increase improved detectability of small structures (needles) with regard to images obtained from a single location, in which image texture masked their presence. [ABSTRACT FROM AUTHOR]

Studies on iterative reconstruction techniques on computed tomographic (CT) scanners show reduced noise and changed image texture. The purpose of this study was to address the possibility of dose reduction and improved conspicuity of lesions in a liver phantom for different iterative reconstruction algorithms. An anthropomorphic upper abdomen phantom, specially designed for receiver operating characteristic analysis was scanned with 2 different CT models from the same vendor, GE CT750 HD and GE Lightspeed VCT. Images were obtained at 3 dose levels, 5, 10, and 15mGy, and reconstructed with filtered back projection (FBP), and 2 different iterative reconstruction algorithms; adaptive statistical iterative reconstruction and Veo. Overall, 5 interpreters evaluated the images and receiver operating characteristic analysis was performed. Standard deviation and the contrast to noise ratio were measured. Veo image reconstruction resulted in larger area under curves compared with those adaptive statistical iterative reconstruction and FBP image reconstruction for given dose levels. For the CT750 HD, iterative reconstruction at the 10mGy dose level resulted in larger or similar area under curves compared with FBP at the 15mGy dose level (0.88-0.95 vs 0.90). This was not shown for the Lightspeed VCT (0.83-0.85 vs 0.92). The results in this study indicate that the possibility for radiation dose reduction using iterative reconstruction techniques depends on both reconstruction technique and the CT scanner model used. [ABSTRACT FROM AUTHOR]

The main purpose of this work is to explore the usefulness of fractal descriptors estimated in multi-resolution domains to characterize biomedical digital image texture. In this regard, three multi-resolution techniques are considered: the well-known discrete wavelet transform (DWT) and the empirical mode decomposition (EMD), and; the newly introduced; variational mode decomposition mode (VMD). The original image is decomposed by the DWT, EMD, and VMD into different scales. Then, Fourier spectrum based fractal descriptors is estimated at specific scales and directions to characterize the image. The support vector machine (SVM) was used to perform supervised classification. The empirical study was applied to the problem of distinguishing between normal and abnormal brain magnetic resonance images (MRI) affected with Alzheimer disease (AD). Our results demonstrate that fractal descriptors estimated in VMD domain outperform those estimated in DWT and EMD domains; and also those directly estimated from the original image. [ABSTRACT FROM AUTHOR]

This paper proposes an idea that treats multi-beam bathymetric sonar data as images for underwater terrain aided navigation. Bathymetric measurements are converted into the form of digital images by interpolation at the beginning of the process. Next, image texture features are obtained and integrated into terrain navigability parameters to define an underwater terrain feature vector. Then, by scanning the underwater terrain map using real-time data, the best match between their feature vectors are considered as the position estimation. Simulation results suggest that this method is reliable when signal to noise ratio is under 2 dB. In addition, the method is shown to be robust to image rotation as large as 20° and scale variance as small as 50×50 pixels, which means it is not sensitive to direction errors and the distance between the underwater vehicle and the bottom may vary. When image resolution is 1 m 2 /pixel and scanning step is 1 m/move, accuracy of the method can be as high as 1 m in suitable terrains. In comparison with terrain contour matching algorithm, performance of the proposed method is more stable when direction error is bigger than 10° and the accuracy is approximately 50% better while noise and scale vary in the real-time images. [ABSTRACT FROM AUTHOR]

As the digital images produce a lot of information about the pixels, there is a need to find alternative methods to reduce the image feature dataset for faster and automatic classification of plants through digital leaf images. In the present work, the leaf image texture features have been extracted through Gabor based techniques and then subjecting them to PSO-CFS based search method for identifying the best set of features from the complete feature set and then classifying them using four classification algorithms like KNN, J48, CART and RF. Another objective of this work is to utilize the two faces available on the plant leaves (Dorsal and Ventral), instead of one (i.e. Dorsal) for classification of plants on the basis of digital leafimages and to analyse the effects on classification accuracy values for dorsal and ventral sides of leaf images. [ABSTRACT FROM AUTHOR]

Wavelet thresholding is an important branch of the image denoising field. A key parameter in the algorithms is noise level. As a novel tool of image analysis, quaternion wavelet owns some superior properties compared to discrete wavelets, such as nearly shift-invariant wavelet coefficients and phase based texture presentation. We aim to propose an easy and efficient method to estimate the noise level accurately via quaternion wavelet and further improve the denoising performance. We find that the variance sum of high frequency coefficients of quaternion wavelet is approximately equal to the noise level. However, with the advent of strong edges and/or less smooth regions, this metric would overestimate the noise level. Phases in the quaternion wavelet domain can represent the image texture information. On the premise of detecting smooth regions via phases operation, i.e. without many textures, the proposed noise level estimation method is also suitable to images with complex scenes. The performance of the proposed noise level estimation algorithm is demonstrated superior to classical algorithms. Also, the proposed algorithm can enhance those noise level dependent techniques to improve the denoising performances which are competitive to the state-of-art denoising algorithms. [ABSTRACT FROM AUTHOR]

Atomic force microscopy (AFM) was used to evaluate the mechanical properties of apple tissue ( Malus domestica var. Golden Delicious) and isolated cells. Young's modulus (E) obtained for the tissue was 0.86 ± 0.81 MPa, while for isolated cells an average value of 0.63 ± 0.42 MPa was found. The fractal dimension of the distribution of E (FD E ) was estimated to be 1.81 for tissue and 2.47 for isolated cells, which indicates more anisotropic structures in single cell topography in comparison with the tissue. Finally, an image texture analysis was performed and different values were obtained; for roughness Ra = 49.70 ± 23 and 18.00 ± 3, entropy Ent = 5.26 ± 0.66 and 5.70 ± 0.23 and for the fractal dimension of the texture image FD T = 2.55 ± 0.02 and 2.62 ± 0.03, for tissue and isolated cells respectively. Industrial relevance AFM is rarely used to evaluate the structural and mechanical properties of food materials. AFM studies could provide insight on food properties. For example, the nanoindentation of cells by AFM could be used to follow-up changes to the mechanical properties that occur during the processing and storage of fruits with industrial relevance such as apple. Study of the nanostructure and nanomechanical properties of vegetable cells could lead to a more in-depth understanding of the relationship structure–functionality in foodstuffs. This could be applied to address several issues, such as extending the storage life of climacteric fruits, verifying the structural modifications of the tissues when they are chemically treated and having a better control of mechanical and structural properties. Currently, fractal and image analysis is gaining attention in the food science industry. Their application allows one to obtain quantitative information from nanoindentation data and AFM images which can be related to structural modifications which occur during ripening and processing of climacteric fruits. This methodology could be extended to other food materials. [ABSTRACT FROM AUTHOR]

This paper presents an integrated system for the automatic analysis of mammograms to assist radiologists in confirming their diagnosis in mammography screening. The proposed automated confirmatory system (ACS) can process a digitalized mammogram online, and generates a high quality filtered segmentation of an image for biological interpretation and a texture-feature based diagnosis. We use a serial of image pre-processing and segmentation techniques, including 2D median filtering, seeded region growing (SRG) algorithm, image contrast enhancement, to remove noise, delete radiopaque artifacts and eliminate the projection of the pectoral muscle from a digitalized mammogram. We also develop an entire-image texture-feature based classification method, by combining a Rough-set approach to extract five fundamental texture features from images, and then an Artificial Neural Network technique to classify a mammogram as: normal; indicating the presence of a benign lump; or representing a malignant tumor. Here, 222 random images from the Mammographic Image Analysis Society (MIAS) database are used for the offline ACS training. Once the system is tuned and trained, it is ready for the automated use for the analysis and diagnosis of new mammograms. To test the trained system, a separate set of 100 random images from the MIAS and another set of 100 random images from the independent BancoWeb database are selected. The proposed ACS is shown to be successful in confirming diagnosis of mammograms from the two independent databases. [ABSTRACT FROM AUTHOR]

Rationale and Objectives: (3)He magnetic resonance imaging (MRI) can be used to quantify functional responses to asthma therapy and provocation. Ventilation imaging offers quantitative information beyond ventilation defects that have not yet been exploited. Therefore, our objective was to evaluate hyperpolarized (3)He MRI ventilation defect percent (VDP) and compare this and pulmonary function measurements to ventilation image texture features and their changes post-bronchodilator administration in patients with asthma. Materials and Methods: Volunteers with a diagnosis of asthma provided written informed consent to an ethics board-approved protocol and underwent pulmonary function tests and MRI before and after salbutamol inhalation. MR images were analyzed using VDP, and their texture was evaluated via gray-level run-length matrices. These texture classifiers were compared to VDP in responders to bronchodilation based on VDP (VDP responders) and forced expiratory volume in 1 s (FEV1) (FEV1 responders). Results: In total, 47 patients with asthma (18 males 39 ± 13 years, FEV1 = 79 ± 21%) reported significantly improved FEV1, FEV1/forced vital capacity (FVC), residual volume (RV)/total lung capacity (TLC) (all P = .0001) and VDP (P = .01) post-salbutamol. Post-salbutamol, VDP responders and nonresponders to salbutamol were significantly different for coarse-texture features including long-run emphasis (LRE) and long-run, low gray-level emphasis (LRLGE, both P < .05) and for FEV1 responders to salbutamol, there was significantly different long-run, high gray-level emphasis (LRHGE, P = .04). There were significant relationships for VDP with LRE (R = .50, P = .0003), LRLGE (R = .34, P = .02), and LRHGE (R = .56, P = .0001). Receiver operating characteristic curves showed VDP with the strongest performance (AUC = .92), followed by coarse-texture classifier LRHGE (AUC = .83), FEV1 (AUC = .80), LRE (AUC = .66), FVC (AUC = .58), and LRLGE (AUC = .42). Conclusions: In patients with asthma, differences in ventilation patchiness post-salbutamol can be quantified using coarse-texture classifiers that are significantly different in bronchodilator responders. [ABSTRACT FROM AUTHOR]

This paper presents two methods to deal with the problem that traditional image denoising algorithms may easily neglect image texture details. The first one is global adaptive fractional integral algorithm (GAFIA) which deals with common noises. It selects the optimal integral order of each pixel based on the local average gradient. The second is image denoising and enhancement algorithm based on adaptive fractional calculus of small probability strategy (AFC-SPS) which deals with salt & pepper noise. It regards the appearance of noise points as small probability events, divides them, and segments the image edges and weak textures by the improved two-dimensional Otsu algorithm. Then, the function of adaptive fractional order is constructed. Experimental results show that, both of the methods have good image denoising effect, and the AFC-SPS algorithm has a better effect than other methods in enhancing the edge and preserving the texture. [ABSTRACT FROM AUTHOR]

In this paper, a method for on-machine tool condition monitoring by processing the turned surface images has been proposed. Progressive monitoring of cutting tool condition is inevitable to maintain product quality. Thus, image texture analyses using gray level co-occurrence matrix, Voronoi tessellation and discrete wavelet transform based methods have been applied on turned surface images for extracting eight useful features to describe progressive tool flank wear. Prediction of cutting tool flank wear has also been performed using these eight features as predictors by utilizing linear support vector machine based regression technique with a maximum 4.9% prediction error. [ABSTRACT FROM AUTHOR]

We propose a hierarchical (BV, G ) variational decomposition model for multiscale texture extraction in this paper, which can offers a hierarchical, separated representation of image texture in different scales. The proposed hierarchical decomposition is obtained by replacing the fixed scale parameter of the A 2 BC model with a varying sequence. Some properties of this hierarchical decomposition are presented and its convergence is proved. We adopt Euclidean projection algorithm to solve this hierarchical decomposition model numerically. In addition, we use this hierarchical decomposition to achieve the multiscale texture extraction. The performance of the proposed model is demonstrated with both synthetic and real images. [ABSTRACT FROM AUTHOR]

Background Food microstructure can be visualised by a wide range of microscopic techniques, however these methods are usually destructive and require sample preparation. X-ray micro-computed tomography (μCT) provides an alternative as it is non-invasive, non-destructive and requires no sample preparation. It characterises structures three-dimensionally, allowing evaluation of microstructural changes at resolutions as high as a few hundred nanometres. After the discovery of X-rays in 1895, X-ray computed tomography (CT) was developed and introduced into clinical practices in the 1970s. The first X-ray μCT food application, to detect the maturity of green tomatoes, followed in 1991. Scope and approach This review aims to provide an overview of the basic principles of X-ray μCT, the different systems, image processing and analysis as well as image texture analysis. Food applications are highlighted and the review concludes with future trends of X-ray μCT. Key findings and conclusions The controlled production and stability of microstructure is of great interest to the food industry. Both laboratory μCT and synchrotron systems are becoming more common and thus will lead to imaging in three dimensions at a micron scale playing a much bigger role in future food studies. Limitations include operator dependency, time and cost constraints and imaging artefacts. Technological and computational progress, however, encourages the growth of this technique in food science. [ABSTRACT FROM AUTHOR]

In this paper, we investigate the effects that different preprocessing techniques have on the performance of features extracted from Haralick's co-occurrence matrix, one of the best known methods for analyzing image texture. In addition, we compare and combine different strategies for extracting descriptors from the co-occurrence matrix. We propose an ensemble of different preprocessing methods, where, for each descriptor, a given Support Vector Machine (SVM) classifier is trained. The set of classifiers is then combined by weighted sum rule. The best result is obtained by combining the extracted descriptors using the following preprocessing methods: wavelet decomposition, local phase quantization, orientation, and the Weber law descriptor. Texture descriptors are extracted from the entire co-occurrence matrix, as well as from sub-windows, and evaluated at multiple scales. We validate our approach on eleven image datasets representing different image classification problems using the Wilcoxon signed rank test. Results show that our approach improves the performance of standard methods. All source code for the approaches tested in this paper will be available at: https://www.dei.unipd.it/node/2357 [ABSTRACT FROM AUTHOR]

Interpretation of long-term Electroencephalography (EEG) records is a tiresome task for clinicians. This paper presents an efficient, low cost and novel approach for patient-specific classification of long-term epileptic EEG records. We aim to achieve this with the minimum supervision from the neurologist. To accomplish this objective, first a novel feature extraction method is proposed based on the mapping of EEG signals into two dimensional space, resulting into a texture image. The texture image is constructed by mapping and scaling EEG signals and their associated frequency sub-bands into the gray-level image domain. Image texture analysis using gray level co-occurrence matrix (GLCM) is then applied in order to extract multivariate features which are able to differentiate between seizure and seizure-free events. To evaluate the discriminative power of the proposed feature extraction method, a comparative study is performed, against other dedicated feature extraction methods. The comparative performance evaluations show that the proposed feature extraction method can outperform other state-of-art feature extraction methods with a low computational cost. With a training rate of 25%, the overall sensitivity of 70.19% and specificity of 97.74% are achieved in the classification of over 163 h of EEG records using support vector machine (SVM) classifiers with linear kernels and trained by the stochastic gradient descent (SGD) algorithm. [ABSTRACT FROM AUTHOR]

With recent improvements in instrumentation and computer infrastructure, machine vision technologies have produced innovative approaches for controlling and monitoring mineral flotation process. In order to efficiently analyze froth image, froth texture is used to accurately and rapidly extract froth characteristics based on the statistics of image pixels. The characterization and identification of texture require a method that can express the context surrounding each pixel by combining local and global texture characteristics. To extract the distinctive froth texture features in different production states, a novel complex networks-based texture extraction and classification method for froth imaging is proposed in this paper. A network model is constructed by expressing pixels as network nodes and similarities between pixels as network links. This method automatically sets the optimal algorithm parameters for the complex network modeling of the froth images according to bubble sizes by using the Minkowski distance. Energy and entropy measurements are used to quantify the properties of the connectivity and topology of the froth-image network model. Copper froth images at different production states extracted from the flotation monitoring system in a flotation plant are used to test the froth-image network model. The experimental results show that the proposed method accurately describes the froth image texture and also robustly classifies the different production states. [ABSTRACT FROM AUTHOR]

The successful launch of the 30-m Landsat-8 Operational Land Imager (OLI) pushbroom sensor offers a new primary data source necessary for aboveground biomass (AGB) estimation, especially in resource-limited environments. In this work, the strength and performance of Landsat-8 OLI image derived texture metrics (i.e. texture measures and texture ratios) in estimating plantation forest species AGB was investigated. It was hypothesized that the sensor’s pushbroom design, coupled with the presence of refined spectral properties, enhanced radiometric resolution (i.e. from 8 bits to 12 bits) and improved signal-to-noise ratio have the potential to provide detailed spectral information necessary for significantly strengthening AGB estimation in medium-density forest canopies. The relationship between image texture metrics and measurements of forest attributes can be used to help characterize complex forests, and enhance fine vegetation biophysical properties, a difficult challenge when using spectral vegetation indices especially in closed canopies. This study examines the prospects of using Landsat-8 OLI sensor derived texture metrics for estimating AGB for three medium-density plantation forest species in KwaZulu Natal, South Africa. In order to achieve this objective, three unique data pre-processing techniques were tested (analysis I: Landsat-8 OLI raw spectral-bands vs. raw texture bands; analysis II: Landsat-8 OLI raw spectral-band ratios vs. texture band ratios and analysis III: Landsat-8 OLI derived vegetation indices vs. texture band ratios). The landsat-8 OLI derived texture parameters were examined for robustness in estimating AGB using linear regression, stepwise-multiple linear regression and stochastic gradient boosting regression models. The results of this study demonstrated that all texture parameters particularly band texture ratios calculated using a 3 × 3 window size, could enhance AGB estimation when compared to simple spectral reflectance, simple band ratios and the most popular spectral vegetation indices. For instance, the use of combined texture ratios yielded the highest R 2 values of 0.76 (RMSE = 9.55 t ha −1 (18.07%) and CV-RMSE of 0.18); 0.74 (RMSE = 12.81 t ha −1 (17.72%) and CV-RMSE of 0.08); 0.74 (RMSE = 12.67 t ha −1 (06.15%) and CV-RMSE of 0.06) and 0.53 (RMSE = 20.15 t ha −1 (14.40%) and CV-RMSE of 0.15) overall for Eucalyptus dunii , Eucalyptus grandis , Pinus taeda individually and all species, respectively. Overall, the findings of this study provide the necessary insight and motivation to the remote sensing community, particularly in resource constrained regions, to shift towards embracing various texture metrics obtained from the readily-available and cheap multispectral Landsat-8 OLI sensor. [ABSTRACT FROM AUTHOR]

• Asphalt mixture segregation detection using image texture features. • Otsu algorithm combined with Canny edge detection to obtain aggregate distribution image. • The two-dimensional entropy is used to calculate aggregate distribution uniformity. • A new indicator "U" is proposed to indicate the degree of asphalt mixture segregation. Surface segregation in asphalt mixture caused by the homogeneity of aggregates seriously influences the service life and road performance of asphalt pavements. However, existing segregation detection procedures are cumbersome and the detecting results are heavily influenced by the moisture content of the surface asphalt mixture. Therefore, this paper innovatively proposes two-dimensional (2D) entropy of the chunking image in calculating uniformity complexity (U) to evaluate surface segregation of bituminous mixtures with digital image processing (DIP). Before image chunking, Otsu algorithm and Canny edge detection are properly combined to obtain a more accurate aggregate distribution image. Furthermore, the feasibility of this detecting method is verified through a simulation experiment, a grading comparison experiment, and a noise interference experiment. These experimental results demonstrated that this study presents an efficient approach for the segregation detection of asphalt mixture with high accuracy. [ABSTRACT FROM AUTHOR]

This paper presents alternative solutions for classifying concrete spall severity based on computer vision approaches. Extreme Gradient Boosting Machine (XGBoost) and Deep Convolutional Neural Network (DCNN) are employed for categorizing image samples into two classes: shallow spall and deep spall. To delineate the properties of a concrete surface subject to spall, texture descriptors including local binary pattern, center symmetric local binary pattern, local ternary pattern, and attractive repulsive center symmetric local binary pattern (ARCS-LBP) are employed as feature extraction methods. In addition, the prediction performance of XGBoost is enhanced by Aquila optimizer metaheuristic. Meanwhile, DCNN is capable of performing image classification directly without the need for texture descriptors. Experimental results with a dataset containing real-world concrete surface images and 20 independent model evaluations point out that the XGBoost optimized by the Aquila metaheuristic and used with ARCS-LBP has achieved an outstanding classification performance with a classification accuracy rate of roughly 99%. • Propose a computer vision method for classifying concrete spall severity. • Extreme Gradient Boosting Machine and deep neural networks are employed. • Use local binary pattern variants as image texture descriptors. • The boosting machine is finetuned by the Aquila optimizer. • The proposed method achieves a high classification accuracy rate = 0.99. [ABSTRACT FROM AUTHOR]

Magnetic resonance image texture analysis (MRI-TA) has been already demonstrated as a powerful tool to extract image information better than visual examination of grey levels. Human muscular dystrophy is a disease of which diagnosis is rather easy but it remains a lack of atraumatic methods for its follow-up as well as for the management of treatments presently in progress. MRI appears to be a potential tool for disease follow-up and treatment monitoring. The present study is part of the European project MYO-MRI ( http://myo-mri.eu/ ) aimed at improving diagnosis and understanding of muscle pathology. This paper focuses on the use of Legendre and Zernike moment-based texture analysis of MRI T2-weighted images to follow from 2 to 14 months 5 healthy and 5 Golden Retriever Muscular Dystrophy Dogs (GRMD), acknowledged to be the most accurate animal model available for human Duchenne Muscular Dystrophy. Regions-Of-Interest (ROIs) were characterized by moments and the features so extracted were analyzed by Principal Component Analysis and classified by means of Support Vector Machine. The preliminary results show relevant discriminations between dogs populations in reference to disease status and evolution. [ABSTRACT FROM AUTHOR]

Image texture enhancement is an important topic in computer graphics, computer vision and pattern recognition. By applying the fractional derivative to analyze texture characteristics, a new fractional differential operator mask with adaptive non-integral step and order is proposed in this paper to enhance texture images. A non-regular self-similar support region is constructed based on a local texture similarity measure, which can effectively exclude pixels with low correlation and noise. Then, through applying sub-pixel division and introducing a local linear piecewise model to estimate the gray value in between the pixels, the resulting non-integral steps can improve the characterization of self-similarity that is inherent in many image types. Moreover, with in-depth understanding of the local texture pattern distribution in the support region, adaptive selection of the fractional derivative order is also performed to deal with complex texture details. Finally, the non-regular fractional differential operator mask which incorporates adaptive non-integral step and order is constructed. Experimental results show that, for images with rich texture contents, the effective characterization of the degree of self-similarity in the texture patterns based on our proposed approach leads to improved image enhancement results when compared with conventional approaches. [ABSTRACT FROM AUTHOR]

In uncooled long-wave infrared (LWIR) imaging systems, nonlinear behaviors of infrared detectors and lack of column cross-calibration generate obvious strip non-uniformity in the captured infrared images. Human observers are particularly sensitive to the resulting high-frequency Fixed Pattern Noise (FPN). In this paper, we propose to learn characteristics of such strip-type FPN through a set of thermal calibration experiments. Our thermal calibration experiments discover that a polynomial curve model can be used to approximate the relationship between infrared data and strip noise of sensor detectors within a column. The derived noise behavioral model allows us to distinguish high-contrast components caused by image texture and strip noise. An effective single-image based processing algorithm is proposed to remove strip-type non-uniformity in infrared images without causing undesired blurring effects. The performance of the proposed technique is thoroughly investigated, and is compared to the state-of-the-art strip denoising algorithm using realistic infrared images. [ABSTRACT FROM AUTHOR]

The aim of this work was to evaluate the effect of HHP treatment and PAST on the microstructure of red Lamuyo-type sweet peppers using image analysis and to determine the parameters that allow characterizing the changes observed on the structure using different magnifications (100 ×, 200 ×, and 350 ×). The results show that all the preservation treatments evaluated caused structural modifications on the microstructure of red sweet pepper, but HHP at 500 MPa and PAST had less impact. Fractal dimension texture, contrast, inverse difference moment, and entropy are texture features that are appropriate for characterizing the effect of HHP and PAST on red pepper texture. In this context, it is important to consider the magnification at which red pepper texture is evaluated because cell damage caused by treatments is best observed at low magnification. Consequently, image analysis could be used in future studies to relate microstructure to the functionality of products subject to HHP. Industrial relevance Red sweet peppers ( Capsicum annuum ) are an excellent source of essential nutrients and bioactive compounds. High hydrostatic pressure (HHP) applied during food processing can improve the retention of food quality attributes and nutritional and organoleptic properties better than pasteurization (PAST). Image analysis is a non-invasive technique that allows to provide objective evaluations from digitalized images. There are no studies that quantify the effect of HHP and PAST using image texture parameters and that evaluate the effect of the magnification used on these texture features. These features are critical factors that determine food acceptance or rejection by the consumers. Thus, texture measurement has gained much attention from food science and industry. Therefore, it would be interesting to study the effect of these treatments on the microstructure of red sweet pepper tissue using image analysis. Thereby, it would be possible to relate the image information to structural modifications and to the extractability of bioactive compounds or acceptance of preservation processes by consumers. [ABSTRACT FROM AUTHOR]

Fatty Liver Disease (FLD) is a progressively prevalent disease that is present in about 15% of the world population. Normally benign and reversible if detected at an early stage, FLD, if left undetected and untreated, can progress to an irreversible advanced liver disease, such as fibrosis, cirrhosis, liver cancer and liver failure, which can cause death. Ultrasound (US) is the most widely used modality to detect FLD. However, the accuracy of US-based diagnosis depends on both the training and expertise of the radiologist. US-based Computer Aided Diagnosis (CAD) techniques for FLD detection can improve accuracy, speed and objectiveness of the diagnosis, and thereby, reduce operator dependability. In this paper, we first review the advantages and limitations of different diagnostic methods which are currently available to detect FLD. We then review the state-of-the-art US-based CAD techniques that utilize a range of image texture based features like entropy, Local Binary Pattern (LBP), Haralick textures and run length matrix in several automated decision making algorithms. These classification algorithms are trained using the features extracted from the patient data in order for them to learn the relationship between the features and the end-result (FLD present or absent). Subsequently, features from a new patient are input to these trained classifiers to determine if he/she has FLD. Due to the use of such automated systems, the inter-observer variability and the subjectivity of associated with reading images by radiologists are eliminated, resulting in a more accurate and quick diagnosis for the patient and time and cost savings for both the patient and the hospital. [ABSTRACT FROM AUTHOR]

Finding mucosal abnormalities (e.g., erythema, blood, ulcer, erosion, and polyp) is one of the most essential tasks during endoscopy video review. Since these abnormalities typically appear in a small number of frames (around 5% of the total frame number), automated detection of frames with an abnormality can save physician's time significantly. In this paper, we propose a new multi-texture analysis method that effectively discerns images showing mucosal abnormalities from the ones without any abnormality since most abnormalities in endoscopy images have textures that are clearly distinguishable from normal textures using an advanced image texture analysis method. The method uses a "texton histogram" of an image block as features. The histogram captures the distribution of different "textons" representing various textures in an endoscopy image. The textons are representative response vectors of an application of a combination of Leung and Malik (LM) filter bank (i.e., a set of image filters) and a set of Local Binary Patterns on the image. Our experimental results indicate that the proposed method achieves 92% recall and 91.8% specificity on wireless capsule endoscopy (WCE) images and 91% recall and 90.8% specificity on colonoscopy images. [ABSTRACT FROM AUTHOR]

We describe a framework for deriving sequences of digital elevation models (DEMs) for the analysis of active lava flows using oblique stereo-pair time-lapse imagery. A photo-based technique was favoured over laser-based alternatives due to low equipment cost, high portability and capability for network expansion, with images of advancing flows captured by digital SLR cameras over durations of up to several hours. However, under typical field scale scenarios, relative camera orientations cannot be rigidly maintained (e.g. through the use of a stereo bar), preventing the use of standard stereo time-lapse processing software. Thus, we trial semi-automated DEM-sequence workflows capable of handling the small camera motions, variable image quality and restricted photogrammetric control that result from the practicalities of data collection at remote and hazardous sites. The image processing workflows implemented either link separate close-range photogrammetry and traditional stereo-matching software, or are integrated in a single software package based on structure-from-motion (SfM). We apply these techniques in contrasting case studies from Kilauea volcano, Hawaii and Mount Etna, Sicily, which differ in scale, duration and image texture. On Kilauea, the advance direction of thin fluid lava lobes was difficult to forecast, preventing good distribution of control. Consequently, volume changes calculated through the different workflows differed by ∼10% for DEMs (over ∼30 m 2 ) that were captured once a minute for 37 min. On Mt. Etna, more predictable advance (∼3 m h −1 for ∼3 h) of a thicker, more viscous lava allowed robust control to be deployed and volumetric change results were generally within 5% (over ∼500 m 2 ). Overall, the integrated SfM software was more straightforward to use and, under favourable conditions, produced results comparable to those from the close-range photogrammetry pipeline. However, under conditions with limited options for photogrammetric control, error in SfM-based surfaces may be difficult to detect. [ABSTRACT FROM AUTHOR]

One quarter of osteoporotic fractures occur in men. TBS, a gray-level measurement derived from lumbar spine DXA image texture, is related to microarchitecture and fracture risk independently of BMD. Previous studies reported the ability of spine TBS to predict osteoporotic fractures in women. Our aim was to evaluate the ability of TBS to predict clinical osteoporotic fractures in men. Methods 3620 men aged ≥ 50 (mean 67.6years) at the time of baseline DXA (femoral neck, spine) were identified from a database (Province of Manitoba, Canada). Health service records were assessed for the presence of non-traumatic osteoporotic fracture after BMD testing. Lumbar spine TBS was derived from spine DXA blinded to clinical parameters and outcomes. We used Cox proportional hazard regression to analyze time to first fracture adjusted for clinical risk factors (FRAX without BMD), osteoporosis treatment and BMD (hip or spine). Results Mean followup was 4.5years. 183 (5.1%) men sustain major osteoporotic fractures (MOF), 91 (2.5%) clinical vertebral fractures (CVF), and 46 (1.3%) hip fractures (HF). Correlation between spine BMD and spine TBS was modest (r=0.31), less than correlation between spine and hip BMD (r=0.63). Significantly lower spine TBS were found in fracture versus non-fracture men for MOF (p<0.001), HF (p<0.001) and CVF (p=0.003). Area under the receiver operating characteristic curve (AUC) for incident fracture discrimination with TBS was significantly better than chance (MOF AUC=0.59, p<0.001; HF AUC=0.67, p<0.001; CVF AUC=0.57, p=0.032). TBS predicted MOF and HF (but not CVF) in models adjusted for FRAX without BMD and osteoporosis treatment. TBS remained a predictor of HF (but not MOF) after further adjustment for hip BMD or spine BMD. Conclusion We observed that spine TBS predicted MOF and HF independently of the clinical FRAX score, HF independently of FRAX and BMD in men. Studies with more incident fractures are needed to confirm these findings. [ABSTRACT FROM AUTHOR]

Einkorn, durum wheat and Kamut® are rich of carotenoids, antioxidants with beneficial effects on human health. In the present study the effect of flour granule size on carotenoid content and colour was assessed; furthermore, the suitability of two colorimeters (Minolta Chroma meter CR-210 and Minolta Chroma meter II Reflectance), a spectrophotometer (Jasco V-650 with integrating sphere) and image analysis to define colour and determine carotenoid concentration in wheat flours of different granulometry was tested. Carotenoid content did not vary across flours of diverse size, except in the finest einkorn fraction (< 80 μm), which had lower concentration. For all instruments colour coordinate L* decreased and b* increased with flour size growth, while a* varied with the different devices. A Principal Components Analysis (PCA), performed considering carotenoid content and all colorimetric indices, distinguished einkorn from durum and Kamut®, and divided samples of different granulometry; a similar result was achieved by a PCA performed on the absorbance spectra from the integrating sphere. The PCA on image texture data classified the samples following flour size. In conclusion, flour colour is determined not only by carotenoid content but also by flour particle size: therefore, direct colour measurement seems not suitable to predict flour carotenoid content. [ABSTRACT FROM AUTHOR]

Waterlogging has recently occurred in cities across China, causing serious economic and property losses. It is imperative to assess the susceptibility of cities to waterlogging, especially those in urban areas, to reduce the damages caused by waterlogging. Artificial Neural Network algorithms have been used to assess the hazards of waterlogging because of their simplicity, objectivity and efficiency. Considering the neighborhood effects, this research proposes an integrated Multiple Geograpical Units Convolutional Neural Network (MGCNN) model for assessing the waterlogging risk in the primarily urban area of Guangzhou, China. Ten spatial factors were considered to be added to the MGCNN, which are DEM, Slope, Std, NDVI, NDBI, MNDWI, Entropy, Contrast, Correlation and Energy based on a Gray-level Co-occurrence Matrix (GLCM). The results show that 1) the texture features of remote sensing image plays an important role in the identification of urban waterlogging area through significance analysis of the ten impact factors; 2) the accuracy rate of the MGCNN trained with the input data of the four risk levels is higher than 0.88, while the maximum accuracy rate of the two risk levels is only about 0.6; 3) the MGCNN has a lower cross-entropy loss value of 0.82 than the SGCNN of 1.13 after 20,000 iterations of training; 4) The Area Under the Curve (AUC) value of the MGCNN was about 0.95, highest among the MGCNN,SGCNN, SVM-BRF and SVM-Poly models; 5) Extremely high risk areas for waterlogging were mainly found in low-lying areas with a high density of urban buildings, including the Yuexiu district, the northern Liwan district, the central and southern sections of the Tianhe district, and the western Haizhu district. Therefore, fully considering neighborhood effects of the surrounding geographical units, the constructed MGCNN model improves not only the accuracy but also the interpretability of the black box. •Spatial evaluation of urban waterlogging risk can be beneficial in creating strategies for urban waterlogging mitigation. •A spatial framework integrating CNN with GIS was introduced to assess urban waterlogging risk. •Compared with two-level data, the MGCNN model trained using four-level data can further improve the accuracy of evaluation. •Image texture based on GLCM has a positive impact on assessment results. •Considering spatial correlation, and the accuracy of the MGCNN was effectively improved. [ABSTRACT FROM AUTHOR]

Creep deformation of human vertebrae accumulates under physiological levels of load and is understood to contribute to the progression toward clinically observable vertebral fracture. However, little information is available in terms of clinically measurable predictors of creep behavior in human vertebrae. In this study, creep tests were performed on 22 human cadaveric T12 vertebrae (13 male, 9 female; age 41–90). Areal and volumetric bone density parameters were measured from the same specimens using dual x-ray absorptiometry and high resolution computed tomography. Image textural analyses (which probe the organization of image intensities within the cancellous bone in low resolution clinical imaging) were performed using digital tomosynthesis (DTS) images. Multiple regression models were constructed to examine the relationship between creep properties and bone density and DTS image textural parameters. For the standard clinical imaging configuration, models including DTS derived image textural parameters alone were generally more explanatory (adjusted R2: 0.14–0.68) than those with bone density parameters forced in the models (adjusted R2: 0.17–0.61). Metrics of textural heterogeneity and anisotropy presented as the most explanatory imaging markers for creep deformation and recovery from creep. These metrics of image texture may help provide, independent from bone mass, important clinically measurable indicators of the time dependent deformation of human vertebrae. [ABSTRACT FROM AUTHOR]

Plant phenology is one of the most reliable indicators of species responses to global climate change, motivating the development of new technologies for phenological monitoring. Digital cameras or near remote systems have been efficiently applied as multi-channel imaging sensors, where leaf color information is extracted from the RGB (Red, Green, and Blue) color channels, and the changes in green levels are used to infer leafing patterns of plant species. In this scenario, texture information is a great ally for image analysis that has been little used in phenology studies. We monitored leaf-changing patterns of Cerrado savanna vegetation by taking daily digital images. We extract RGB channels from the digital images and correlate them with phenological changes. Additionally, we benefit from the inclusion of textural metrics for quantifying spatial heterogeneity. Our first goals are: (1) to test if color change information is able to characterize the phenological pattern of a group of species; (2) to test if the temporal variation in image texture is useful to distinguish plant species; and (3) to test if individuals from the same species may be automatically identified using digital images. In this paper, we present a machine learning approach based on multiscale classifiers to detect phenological patterns in the digital images. Our results indicate that: (1) extreme hours (morning and afternoon) are the best for identifying plant species; (2) different plant species present a different behavior with respect to the color change information; and (3) texture variation along temporal images is promising information for capturing phenological patterns. Based on those results, we suggest that individuals from the same species and functional group might be identified using digital images, and introduce a new tool to help phenology experts in the identification of new individuals from the same species in the image and their location on the ground. [ABSTRACT FROM AUTHOR]

Abstract: Recent approaches to crater detection have been inspired by face detection’s use of gray-scale texture features. Using gray-scale texture features for supervised machine learning crater detection algorithms provides better classification of craters in planetary images than previous methods. When using Haar features it is typical to generate thousands of numerical values from each candidate crater image. This magnitude of image features to extract and consider can spell disaster when the application is an entire planetary surface. One solution is to reduce the number of features extracted and considered in order to increase accuracy as well as speed. Feature subset selection provides the operational classifiers with a concise and denoised set of features by reducing irrelevant and redundant features. Feature subset selection is known to be NP-hard. To provide an efficient suboptimal solution, four genetic algorithms are proposed to use greedy selection, weighted random selection, and simulated annealing to distinguish discriminate features from indiscriminate features. Inspired by analysis regarding the relationship between subset size and accuracy, a squeezing algorithm is presented to shrink the genetic algorithm’s chromosome cardinality during the genetic iterations. A significant increase in the classification performance of a Bayesian classifier in crater detection using image texture features is observed. [Copyright &y& Elsevier]

Abstract: For an in-depth understanding of the failure of structural materials the study of deformation mechanisms in the material bulk is fundamental. In situ synchrotron computed laminography provides 3D images of sheet samples and digital volume correlation yields the displacement and strain fields between each step of experimental loading by using the natural contrast of the material. Difficulties arise from the lack of data, which is intrinsic to laminography and leads to several artifacts, and the little absorption contrast in the 3D image texture of the studied aluminum alloy. To lower the uncertainty level and to have a better mechanical admissibility of the measured displacement field, a regularized digital volume correlation procedure is introduced and applied to measure localized displacement and strain fields. [Copyright &y& Elsevier]