Search

The task of approaching and capturing a free-tumbling satellite on-orbit presents open challenges for autonomous guidance and control strategies. One of these is to robustly predict the satellite’s...

Multi-view learning has become a significant research topic in image processing, data mining and machine learning due to the proliferation of multi-view data. Considering the difficulty in obtaining labeled data in many real applications, we focus on the multi-view unsupervised feature selection problem. Most existing multi-view feature selection introduce an identical similarity matrix among different views, which cannot preserve the specific correlation between each single view. Also, some of these methods just consider either global or local structures. In this paper, we propose an embedding method, Adaptive Similarity Embedding for Unsupervised Multi-View Feature Selection (ASE-UMFS). This method reduces the high-dimensional data to the low dimensions and unifies different views to a combination weight matrix. We also use parameters to constraint the similarity matrix for the local structure, where the regularization term is used to add a prior of uniform distribution; taking into account of the independence in projection matrix among different views, optimization of the similarity matrix is further improved. To confirm the effectiveness of ASE-UMFS, comparisons are made with benchmark algorithm on real-world data sets. The experimental results demonstrate that the proposed algorithm outperforms several state-of-the-art methods in multi-view learning.

With the development of video and image processing technology, the field of video tampering forensics is facing enormous challenges. Specifically, as the fundamental basis of judicial forensics, passive forensics for object removal video forgery is particularly essential. To extract tampering traces in video more sufficiently, the author proposed a spatiotemporal trident network based on the spatial rich model (SRM) and 3D convolution (C3D), which provides three branches and can theoretically improve the detection and localization accuracy of tampered regions. Based on the spatiotemporal trident network, a temporal detector and a spatial locator were designed to detect and locate the tampered regions in the temporal and spatial domains of videos. For the temporal detector, 3D CNNs were employed in three branches as the encoders and a bidirectional long short-term memory (BiLSTM) as the decoder. For the spatial locator, a backbone network named C3D-ResNet12 was designed as the encoder of the three branches, and the region proposal networks (RPNs) were employed as the decoders in three branches. In addition, we optimized the loss functions of the above two algorithms based on focal loss and GIoU loss. The experimental results revealed the effectiveness of spatiotemporal detection and localization algorithms: for temporal forgery detection, the accuracy of the frame classification increased to 99+%; for spatial forgery localization, the successful localization rate of the tampered regions in forged frames reached 96+%, and the mean intersection over union of the located tampered regions and the real tampered regions reached 62+%.

This paper uses wavelet neurons instead of traditional neurons, and uses wavelet multi-resolution analysis to decompose the FPGA image scan of the ship engine. Because the neural network has the approximation ability of arbitrary functions, the wavelet transform is connected to the neural network to form a wavelet neural network. To test ship engines. The hardware design of GigE image acquisition and processing system based on FPGA was started. FPGA was used as the main control chip and Gigabit Ethernet was used as the transmission medium. The hardware circuit of the image data acquisition and image processing system was designed. It mainly includes the FPGA main control circuit and the FPGA Peripheral circuits. The high-speed image acquisition, transmission, storage, and display module circuit design is realized. Real-time monitoring and fault analysis of the engine's condition is performed by the FPGA image scanning method, and data of the engine's running state is pre-processed with the help of step tracking technology to make it a standard signal. The data is transmitted to the computer through NI's data acquisition card. Combining feature extraction such as information entropy, Fourier transform, EMD and wavelet neural network technology. The accuracy of the diagnosis results and the actual fault state is improved. It can enable the staff to monitor the running status of the engine in real time, improve the efficiency of engine fault diagnosis, reduce labour costs and maintenance costs, and thus realize intelligent, real-time and accurate status monitoring of the engine.

Image processing with any potential quantum computing hardware requires a quantum color model capable of capturing and manipulating color information in images. In this study, a quantum hue, saturation, and lightness (QHSL) model is proposed as a first attempt to encode perceptually relevant triplet color components using the properties of quantum mechanics (i.e., entanglement and parallelism). The proposed color model was used to define a representation of two-dimensional QHSL images for storage and transformation with fewer computing resources. The configuration of color-assignment attributes within this QHSL representation offer useful applications for image analysis. Specifically, a pseudocolor technique with flexible gray depth divisions is presented for highlighting fine visual details.

We report on a simple approach of time-delayed reservoir computing (RC) based on a two-element phased laser array for image identification. Here the phased laser array with optical feedback and injection is trained according to the representative characteristics extracted through histograms of oriented gradients. These characteristic vectors are multiplied by a random mask signal to form input data, which are subsequently trained in the reservoir. By optimizing the parameters of the RC, we achieve an identification accuracy of 97.44% on the MNIST dataset and 85.46% on the Fashion-MNIST dataset. These results indicate that our proposed RC indeed allows accurate classification of handwritten digit and fashion production. Moreover, we further forecast an RC scheme based on a larger-scale phased laser array, which is expected to tackle more complex tasks at a high speed. Our work offers a possibility for advanced image processing using highly integrated neuromorphic photonic systems.

Face detection is an intelligent approach used in a variety of applications that identifies human faces in digital images. This work presents a new method which composes of a neural network and Techebycheve transforms for face detection. For feature extraction, Tchebychev transform was applied, in which a discrete Tchebychev transform is given for different sampling patterns and several samples here were performed on color images. A Levenberg-Marquardt backpropagation neural network was applied to the transformed image to find faces in the face detection dataset and FDDB benchmarked database. Model performance was measured based on its accuracy and the best result from the newly proposed method was 98.9%. Simulation results showed that the proposed method handles face detection efficiently.

The geometric and color features of agricultural material along with related physical properties are critical to characterize and express its physical quality. The experiments were conducted to classify the physical characteristics like size, shape, color and texture and then workout the relationship between manual observations and using image processing techniques for weight and volume of the four wheat refractions i.e. sound, damaged, shriveled and broken grains of wheat variety PBW 725. A flatbed scanner was used to acquire the images and digital image processing method was used to process the images and output of image analysis was compared with the actual measurements data using digital vernier caliper. A linear relationship was observed between the axial dimensions of refractions between manual measurement and image processing method with R2 in the range of 0.798–0.947. The individual kernel weight and thousand grain weight of the refractions were observed to be in the range of 0.021–0.045 and 12.56–46.32 g respectively. Another linear relationship was found between individual kernel weight and projected area estimated using image processing methodology with R2 in the range of 0.841–0.920. The sphericity of the refractions varied in the range of 0.52–0.71. Analyses of the captured images suggest ellipsoid shape with convex geometry while the same observation was recorded by physical measurements also. A linear relationship was observed between the volume of refractions derived from measured dimensions and calculated from image with R2 in the range of 0.845–0.945. Various color and grey level co-variance matrix texture features were extracted from acquired images using the open-source Python programming language and OpenCV library which can exploit different machine and deep learning algorithms to properly classify these refractions.

Artificial intelligence (AI) has significant potential to positively impact and advance medical imaging, including positron emission tomography (PET) imaging applications. AI has the ability to enhance and optimize all aspects of the PET imaging chain from patient scheduling, patient setup, protocoling, data acquisition, detector signal processing, reconstruction, image processing, and interpretation. AI poses industry-specific challenges which will need to be addressed and overcome to maximize the future potentials of AI in PET. This article provides an overview of these industry-specific challenges for the development, standardization, commercialization, and clinical adoption of AI and explores the potential enhancements to PET imaging brought on by AI in the near future. In particular, the combination of on-demand image reconstruction, AI, and custom-designed data-processing workflows may open new possibilities for innovation which would positively impact the industry and ultimately patients.

To quantitatively measure the connectedness robustness of a complex network, a sequence of values that record the remaining connectedness of the network after a sequence of node- or edge-removal attacks can be used. However, it is computationally time-consuming to measure the network connectedness robustness by attack simulations for large-scale networked systems. In the present paper, an efficient method based on convolutional neural network (CNN) is proposed to train for estimating the network connectedness robustness. The new approach is motivated by the facts that 1) the adjacency matrix of a network can be converted to a gray-scale image and CNN is very powerful for image processing, and 2) CNN has proved very effective in predicting the controllability robustness of complex networks. Extensive experimental studies on directed and undirected, as well as synthetic and real-world networks suggest that: 1) the proposed CNN-based methodology performs excellently in the prediction of the connectedness robustness of complex networks as a process; 2) it performs fairly well as the indicator for the connectedness robustness, compared to other predictive measures.

The pandemic of Covid-19 has caused a shift of paradigm of education, from face-to-face to e-learning. E-learning leads to an escalation in digitalization of handwritten documents because it requires submission of homework and assignments through online. To help teachers in checking digitalized handwritten homework, this paper proposes an automatic checking system based on a convolutional neural network (CNN) for handwritten numeral recognition. The CNN is used to recognize four arithmetic operations in mathematical questions consisting of addition, deduction, multiplication and division. The performance CNN in handwritten numeral recognition have been optimized in terms of activation function and gradient descent algorithm. The proposed CNN is also trained and tested with the MNIST handwritten data set. The experimental results show that the recognition accuracy the improved CNN improves to a certain extent as compared to before optimization.

Diabetic retinopathy is a retinal condition that affects people with diabetes and is the leading cause of blindness in the elderly. It's an asymptomatic illness characterized by abnormalities in blood vessels that might cause them to bleed or leak fluid, resulting in visual distortion. As a result, blood vessel extraction is critical in assisting ophthalmologists in detecting this illness at an early stage and preventing vision loss. Diabetes Retinopathy (DR) is a debilitating chronic illness that is one of the primary causes of blindness and vision impairment in diabetic individuals in industrialized nations. According to studies, the majority of instances may be avoided with early identification and treatment. Physicians utilize retinal imaging to detect lesions associated with this illness during eye screening. The amount of pictures that must be manually examined is getting expensive because of the rising number of diabetics.. In this research, we used Image Processing to offer a technique for automatically classifying diabetic retinopathy disease based on retina fundus pictures. For this, we combined a feature extraction approach based on a pre-trained deep neural network model with a machine learning-based support vector machine classification algorithm. In MATLAB software, the proposed system is examined and analyzed.

Image Super-Resolution (SR) is an important class of image processing techniques to enhance the resolution of images and videos in computer vision. Recent years have witnessed remarkable progress of image super-resolution using deep learning techniques. This article aims to provide a comprehensive survey on recent advances of image super-resolution using deep learning approaches. In general, we can roughly group the existing studies of SR techniques into three major categories: supervised SR, unsupervised SR, and domain-specific SR. In addition, we also cover some other important issues, such as publicly available benchmark datasets and performance evaluation metrics. Finally, we conclude this survey by highlighting several future directions and open issues which should be further addressed by the community in the future., Comment: Accepted by IEEE TPAMI

The demand to improve image display in TFT-LCD, implementation of design for image processing is important. In order to meet the specific requirements of low-end Thin Film Transistor-Liquid-Crystal-Display (TFT-LCD) image display. This paper adopts a novel algorithm to conduct subsequent processing of the medical image after SCALER scaling, including contrast adjustment, gamma correction and dithering. Dithering algorithm is the focus of our research. After the study of some classical video image processing algorithms, and considering the real-time requirements, an intelligent algorithm is implemented for hardware implementation and improvement. For each part, MATLAB language is firstly used for advanced simulation to verify its feasibility, and then Right-To-Left (RTL) hardware language description is carried out. The characteristics extraction from images is performed implementing RGB standard and grayscale images. The pixel intensity is analyzed for each RGB component and the variance is calculated. When a panel displays a variation of 6% related with their reference values, the panel is rejected. The results obtained from classification shows a 95.24% accuracy rate in the detection of defects. The results of the two simulations show that the design achieves the expected goal, and the processing time is shorter.

In a criminal investigation, along with processing forensic evidence, different investigative techniques are used to identify the perpetrator of the crime. It includes collecting and analyzing unconstrained face images, mostly with low resolution and various qualities, making identification difficult. Since police organizations have limited resources, in this paper, we propose a novel method that utilizes off-the-shelf solutions (Dlib library Histogram of Oriented Gradients-HOG face detectors and the ResNet faces feature vector extractor) to provide practical assistance in unconstrained face identification. Our experiment aimed to establish which one (if any) of the basic image enhancement techniques should be applied to increase the effectiveness. Results obtained from three publicly available databases and one created for this research (simulating police investigators’ database) showed that resizing the image (especially with a resolution lower than 150 pixels) should always precede enhancement to improve face detection accuracy. The best results in determining whether they are the same or different persons in images were obtained by applying sharpening with a high-pass filter, whereas normalization gives the highest classification scores when a single weight value is applied to data from all four databases.

Non-alcoholic fatty liver disease (NAFLD) is one of the most common causes of cardiometabolic diseases in overweight individuals. While liver biopsy is the current gold standard to diagnose NAFLD and magnetic resonance imaging (MRI) is a non-invasive alternative still under clinical trials, the former is invasive and the latter costly. We demonstrate electrical impedance tomography (EIT) as a portable method for detecting fatty infiltrate. We enrolled 19 overweight subjects to undergo liver MRI scans, followed by EIT measurements. The MRI images provided the a priori knowledge of the liver boundary conditions for EIT reconstruction, and the multi-echo MRI data quantified liver proton-density fat fraction (PDFF%) to validate fat infiltrate. Using the EIT electrode belts, we circumferentially injected pairwise current to the upper abdomen, followed by acquiring the resulting surface-voltage to reconstruct the liver conductivity. Pearson’s correlation analyses compared EIT conductivity or MRI PDFF with body mass index, age, waist circumference, height, and weight variables. We reveal that the correlation between liver EIT conductivity or MRI PDFF with demographics is statistically insignificant, whereas liver EIT conductivity is inversely correlated with MRI PDFF (R = −0.69, p = 0.003, n = 16). As a pilot study, EIT conductivity provides a portable method for operator-independent and cost-effective detection of hepatic steatosis.

Scene change detection is an image processing problem related to partitioning pixels of a digital image into foreground and background regions. Mostly, visual knowledge-based computer intelligent systems, like traffic monitoring, video surveillance, and anomaly detection, need to use change detection techniques. Amongst the most prominent detection methods, there are the learning-based ones, which besides sharing similar training and testing protocols, differ from each other in terms of their architecture design strategies. Such architecture design directly impacts on the quality of the detection results, and also in the device resources capacity, like memory. In this work, we propose a novel Multiscale Cascade Residual Convolutional Neural Network that integrates multiscale processing strategy through a Residual Processing Module, with a Segmentation Convolutional Neural Network. Experiments conducted on two different datasets support the effectiveness of the proposed approach, achieving average overall $\boldsymbol{F\text{-}measure}$ results of $\boldsymbol{0.9622}$ and $\boldsymbol{0.9664}$ over Change Detection 2014 and PetrobrasROUTES datasets respectively, besides comprising approximately eight times fewer parameters. Such obtained results place the proposed technique amongst the top four state-of-the-art scene change detection methods., Published in: 2020 33rd SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI)

Hepatocellular carcinoma (HCC) is the most common liver cancer with 1 million cases globally. A current clinical challenge is to determine which patients will respond to transarterial chemoembolization (TACE) as effective delivery of the embolic material may be influenced by the tumor vascular supply. The purpose of this study is to develop a novel image processing algorithm for improved quantification of tumor microvascular morphology features using contrast-enhanced ultrasound (CEUS) images and to predict the TACE response based on these biomarkers before treatment. A temporal sequence of CEUS images was corrected from rigid and non-rigid motion artifacts using affine and free form deformation models. Subsequently, a principal component analysis based singular value filter was applied to remove the clutter signal from each frame. A maximum intensity projection was created from high-resolution images. A multiscale vessel enhancement filter was first utilized to enhance the tubular structures as a preprocessing step before segmentation. Morphological image processing methods are used to extract the morphology features, namely, number of vessels (NV) and branching points (NB), vessel-to-tissue ratio (VR), and the mean vessel length (VL), tortuosity (VT), and diameter (VD) from the tumor vascular network. Finally, a support vector machine (SVM) is trained and validated using leave-one-out cross-validation technique. The proposed image analysis strategy was able to predict the patient outcome with 90% accuracy when the SVM was trained with the three features together (NB, NV, VR). Experimental results indicated that morphological features of tumor microvascular networks may be significant predictors for TACE response. Reliable prediction of the TACE therapy response may help provide effective therapy planning.

Yield estimation can provide critical information for high-value crop growers. It helps them to plan and coordinate the logistics of harvesting operations. Most existing yield estimation techniques are based on manual sampling and statistical approaches, which are complex, labor-intensive, and time-consuming. In recent years, advances in image processing and the ubiquitous use of autonomous platforms, such as unmanned aerial platforms, have provided an opportunity for faster and more accurate yield estimation techniques for different crops. A new rapid, efficient, and accurate image processing algorithm was proposed in this study. The image segmentation technique was able to separate mature citrus trees and natural background in CIE (Commission International Eclairage) L*a*b* (i.e., Lab) color space. The effect of camera distance to object, affected resolution, and detection accuracy were evaluated and discussed. The results showed that the distance between 1.524 meters (5 feet) and 2.134 meters (7 feet) could provide high detection accuracy. The number of correctly counted fruit reached 91.69% at the average processing time of 1.1 s per picture for 132 pictures. A correlation coefficient (R2) of 0.98 was obtained between the citrus counting algorithm and counting performed through human observation of 66 sample trees. The comparative analysis of the proposed method showed a higher accuracy rate and robustness compared to other similar studies.

This paper presents an approach for development of costeffective custom video / image processing systems. The approach utilizes the concept of temporal partitioning of resources in the partially reconfigurable FPGA devices. Paper proposes architecture of the multi-mode video-stream processor with cyclically reconfigurable structure. The cost-effectiveness of the proposed approach has been analyzed on the basis of experiments conducted on Multi-mode Adaptive Reconfigurable System (MARS) platform that was developed for that purpose. The video-processing cores associated with stereo-vision algorithms have been developed, tested and analyzed. The experiments have shown that the cost-effectiveness of the systems based on proposed approach can be better than the traditional approaches based on large statically configured FPGAs.

This project details the design and implementation of an image processing pipeline that targets real time video-stitching for semi-panoramic video synthesis. The scope of the project includes the analysis of possible approaches, selection of processing algorithms and procedures, design of experimental hardware set-up (including the schematic capture design of a custom catadioptric panoramic imaging system) and firmware/software development of the vision processing system components. The goal of the project is to develop a frame-stitching IP module as well as an efficient video registration algorithm capable for synthesis of a semi-panoramic video-stream at 30 frames-per-second (fps) rate with minimal FPGA resource utilization. The developed components have been validated in hardware. Finally, a number of hybrid architectures that make use of the synergy between the CPU and FPGA section of the ZYNQ SoC have been investigated and prototyped as alternatives to a complete hardware solution. Keyword: Video stitching, Panoramic vision, FPGA, SoC, vision system, registration

Previous approaches to texture analysis and segmentation use multichannel filtering by applying a set of filters in the frequency domain or a set of masks in the spatial domain. This paper presents two new texture segmentation algorithms based on multichannel filtering in conjunction with neural networks for feature extraction and segmentation. The features extracted by Gabor filters have been applied for image segmentation and analysis. Suitable choices of filter parameters and filter bank coverage in the frequency domain to optimize the filters are discussed. Here we introduce two methods to optimize Gabor filter bank. First, a Gabor filter bank with a flat response is implemented and the optimal feature dimension is extracted by competitive networks. Second, a subset of Gabor filter bank is selected to compose the best discriminative filters, so that each filter in this small set can discriminate a pair of textures in a given image. In both approaches, multilayer perceptrons are employed to segment the extracted features. The comparisons of segmentation results generated using the proposed methods and previous research using Gabor, discrete cosine transform (DCT), and Laws filters are presented. Finally, the segmentation results generated by applying the optimized filter banks to textured images are presented and discussed.

In this document, we propose the recognition of powdery mildew in cucumber leaves based on image processing. Two cucumber cycles were established and infested with powdery mildew. As the disease developed, photos were taken to perform the analysis. Two hundred photographs were manually preprocessed eliminating the background and leaving only leaves infested with the disease. The images were segmented using three threshold binarization techniques: gray scale binarization, RGB binarization and K-means algorithm with initially located centroids. The results were compared between the different methods. The gray scale binarization as well as the RGB binarization allowed locating the disease based on a percentage of the lighter shades, although the latter method analyzes each one of the different color layers. The K-means algorithm identified groups with similar characteristics around provided centers. A false positive detection test was also performed with 25 previously processed photographs. The experimental results show that the proposed gray scale binarization method better results for the recognition of the disease than the other methods.

Lung cancer has been a major contribution to mortality rates world-wide for many years now. There is a need for early diagnosis of lung cancer which if implemented, will help in reducing mortality rates. Recently, image processing techniques have been widely applied in various medical facilities for accurate detection and diagnosis of abnormality in the body images like in various cancers such as brain tumour, breast tumour and lung tumour. This paper is a development of an algorithm based on medical image processing to segment the lung tumour in CT images due to the lack of such algorithms and approaches used to detect tumours. The work involves the application of different image processing tools in order to arrive at the desired result when combined and successively applied. The segmentation system comprises different steps along the process. First, Image preprocessing is done where some enhancement is done to enhance and reduce noise in images. In the next step, the different parts in the images are separated to be able to segment the tumour. In this phase threshold value was selected automatically. Then morphological operation (Area opening) is implemented on the thresholded image. Finally, the lung tumour is accurately segmented by subtracting the opened image from the thresholded image. Support Vector Machine (SVM) classifier is used to classify the lung tumour into 4 different types: Adenocarcinoma(AC), Large Cell Carcinoma(LCC) Squamous Cell Carcinoma(SCC), and No tumour (NT). Keywords: Lung tumour; image processing techniques; segmentation; thresholding; image enhancement; Support Vector Machine; Machine learning

It is of a great importance in modern agriculture to study fast, automatic, inexpensive and accurate method of diagnosing plant diseasesTherefore, timely and accurately diagnosis of the disease in the fields is one of the most important factors in dealing with plant diseases. For this reason, in the present study, the image processing method study, has been examined for diagnosing the two important diseases of rice and tomato, brown spots and leaf blasts. In this study, firstly the data section is treated using improved k-means segmentation, after preprocessing. Secondly, comprehensive features are extracted and the disease areas are demarcated. An improved genetic algorithm is used in the feature selection step. Finally, images are categorized using the k-nearest neighbor’s algorithm (k-NN) classifier. The accuracy of the proposed method for the rice data set is 99.12 and for the tomato data set is 97.29, which shows a very good performance compared to other methods.

Rearing of silkworm is highly dependent on environmental variations. To have a healthy cocoon production, it is necessary to have a proper temperature and humidity controlled house for silkworm rearing. Temperature, humidity and fresh air should be managed to get a wonderful silk product. An ideal temperature of 23°C to 28°C and humidity in between 65% to 85% is to be maintained. IoT based silkworm rearing house consists of sensors and actuators, which are interfaced with a low power controllers. The Sericulture unit can be equipped with a wireless sensor node to sense the real time Temperature and Humidity [1], also necessary actuators to control these environmental parameters. The color change in the body of the worms indicates the different stages and the light yellowish indicates that they have reached to the cocoon stage and the morphological changes in silkworm structure can be used to detect abnormal worms[2].The proposed framework introduces an Internet of Things (IoT) empowered Wireless Personal Area Network (WPAN) system. The received image is first segregated into two classes as diseased or healthy by analyzing the histogram of the background removed image based on thresholding. Again the diseased class will be sub classified into 2 diseases as either Flacherie or Pebrine by applying suitable mask for extracting worm and obtaining the histogram of the worm and analyzing it. The result will be sent to the farmer via E-mail. The proposed system could be a probable solution for productivity in silkworms. View Article DOI: 10.47856/ijaast.2021.v08i9.004

The goal -is to explore ways of using Earth remote sensing data for efficient land use. Methods - detailed information on current location of certain types of agricultural crops in the study areas has been summarized, which opens up opportunities for the effective use of cultivated areas. It was revealed that the basis of the principle of the method under consideration is the relationship between the state and structure of vegetation types with its reflective ability. It has been determined that information on the spectral reflective property of the vegetation cover in the future can help replace more laborious methods of laboratory analysis. For classification of farmland, satellite images of medium spatial resolution with a combination of channels in natural colors were selected. Results - a method for identifying agricultural plants by classification according to the maximum likelihood algorithm was considered. The commonly used complexes of geoinformation software products with modules for special image processing allow displaying indicators in the form of raster images. It is shown that the use of Earth remote sensing data is the most relevant solution in the field of crop recognition and makes it possible to simplify the implementation of such types of work as the analysis of the intensity of land use, the assessment of the degree of pollution with weeds and determination of crop productivity. Conclusions - the research results given in the article indicate that timely information on the current location of certain types of agricultural crops in the studied territories significantly simplifies the implementation of the tasks and increases the resource potential of agricultural lands. In turn, the timing of the survey and the state of environment affect the spectral reflectivity of vegetation.

Synthetic aperture radars (SARs) provide a lot of images that can be used for numerous applications. A problem with acquired images is that they are corrupted by speckle which is a noise-like phenomenon with multiplicative nature. In addition, speckle is non-Gaussian and it is often spatially correlated. A typical task in SAR image processing is despeckling and many methods have been already proposed. However, most of them do not take noise spatial correlation into account during denoising. In this paper, we show how this can be done in despeckling based on discrete cosine transform. The use of frequency-dependent thresholds leads to sufficient improvement of denoising efficiency in terms of visual quality metrics. Moreover, we consider quite complex structure texture images for which noise removal is usually problematic and can lead to information loss. Comparison to the well-known local statistic Lee and Frost filters, extended DCT-based filter is carried out for different remote sensing systems including Sentinel-1 and Sentinel-2.

The latest generation of convolution neural networks (CNNs) has achieved impressive results in the field of image classification. This paper is concerned with a new approach to the development of fruit/plant disease detection model, based on leaf image processing and classification, by the use of ANN. Novel way of training and the methodology used facilitate a quick and easy system implementation in practice. The developed model is able to recognize 13 different types of plant diseases out of healthy leaves, with the ability to distinguish plant leaves from their surroundings. According to our knowledge, this method for plant disease recognition has been proposed for the first time. All essential steps required for implementing this disease recognition model are fully described throughout the paper, starting from gathering images in order to create a database, assessed by agricultural experts. Caffe, a deep learning framework developed by Berkley Vision and Learning Centre, was used to perform the deep CNN training

Historical Islamic ornaments include a fantastic treasury of geometric and mathematical algorithms. Inevitably, restoration of these ornaments in periodic patterns consisting of repeated elements has been faced following and substituting the other available similar ingredients instead of vanished parts. Still, the prediction of parametric, quasi, or non-periodic patterns, where components are not identical, needs to be carried out in a more challenging process than the periodic ones due to shape, scale, or angle of rotation alteration. Intelligent restoration could facilitate the forecasting of damaged parts in such geometric patterns that an algorithm has changed their geometric characteristics. In some architectural heritage, geometric patterns include a parametric algorithm like parametric patterns in the ceiling of Sheikh Lotfollahmosque in Isfahan, Iran, and the dominant structure of Persian domes Karbandi. In this article, the aim is to propose a new method for the smart restoration of the parametric geometric patterns in which, by having access to the image of the existing patterns, the vanished parts could be reconstructed spontaneously. Our approach is based on image processing by detecting boundaries of deterioration, finding every individual element, and extracting features of detected individual patterns via Zernike moments. The order of individual patterns starts from the farthest pattern to detected deterioration. Then by creating a time series, the Back-propagation neural network would be trained by extracted features, and the vanished patterns’ features could be predicted and reconstructed. Eventually, the reconstructed and real patterns are compared to determine differences between them by mean-squared error and to evaluate the performance of our method. To validate the process, a parametric geometric pattern is designed by the assumption that some parts are disappeared. The proposed method’s results, in this case, hold an efficient performance with the accuracy of 92.99%. Furthermore, Sheikh Lotfollah’s patterns and Naseredin Mirza mansion’s patterns as two real cases are tested by the proposed method, representing reliable and suitable performance results.

BACKGROUND: Identification of implant model from primary knee arthroplasty in pre-op planning of revision surgery is a challenging task with added delay. The direct impact of this inability to identify the implants in time leads to the increase in complexity in surgery. Deep learning in the medical field for diagnosis has shown promising results in getting better with every iteration. This study aims to find an optimal solution for the problem of identification of make and model of knee arthroplasty prosthesis using automated deep learning models. METHODS: Deep learning algorithms were used to classify knee arthroplasty implant models. The training, validation and test comprised of 1078 radiographs with a total of 6 knee arthroplasty implant models with anterior–posterior (AP) and lateral views. The performance of the model was calculated using accuracy, sensitivity, and area under the receiver-operating characteristic curve (AUC), which were compared against multiple models trained for comparative in-depth analysis with saliency maps for visualization. RESULTS: After training for a total of 30 epochs on all 6 models, the model performing the best obtained an accuracy of 96.38%, the sensitivity of 97.2% and AUC of 0.985 on an external testing dataset consisting of 162 radiographs. The best performing model correctly and uniquely identified the implants which could be visualized using saliency maps. CONCLUSION: Deep learning models can be used to differentiate between 6 knee arthroplasty implant models. Saliency maps give us a better understanding of which regions the model is focusing on while predicting the results.

Superpixel segmentation is a popular image preprocessing technology in image processing. Among the various methods used to calculate uniform superpixel, the performance of linear spectral clustering (LSC) is better than the state-of-the-art superpixel segmentation algorithms. However, this method is slow on images and has low accuracy on non-convex images. In order to improve this problem, we propose a subsampled clustering method that can accelerate LSC. Meanwhile, this paper presents an improved distance measurement method based on non-convex image features and Manhattan distance, which can achieve high accuracy on non-convex images. The proposed method is evaluated on the BSDS500 dataset. The experimental results confirmed that this method runs faster than LSC, and at the same time produces almost the same superpixel segmentation accuracy on the images. In addition, the proposed method improves the accuracy of superpixel segmentation on non-convex images.

123I thyroid scintigraphy can be performed with either a low energy or medium energy collimator. The high-energy photon emissions from 123I cause septal penetration with scattered photons, which deteriorate image quality. The aim of this study is to evaluate the impact of collimator choice on 123I thyroid scintigraphy in clinical practice. Methods: Forty seven patients who received thyroid planar scintigraphy with both a low energy high resolution (LEHR) collimator and a medium energy (ME) collimator were prospectively recruited using the same imaging protocol. Image quality, collimator sensitivity, and estimation of thyroid size were assessed between LEHR and ME collimators, and were compared with thyroid ultrasonography as the gold standard. Results: Images acquired with the ME collimator demonstrate less scattered background noise, improved thyroid to background contrast, and increased sensitivity in the thyroid gland compared to images acquired by the LEHR collimator. Manual measurement of the thyroid length is more accurate by using the ME collimator. Automatic estimation of the thyroid area size by using the same thyroid cutoff threshold is larger in ME collimator images than in LEHR collimator images. Conclusion:123I thyroid scintigraphy using the ME collimator generates cleaner images with less background noise and has higher collimator sensitivity for thyroid imaging compared to the LEHR collimator. Different thyroid cutoff threshold should be used to estimate the thyroid area size and volume between low and medium energy collimators.

To mitigate financial loss and follow the recommended sanitary measures due to the COVID-19 pandemic, self-reading, a method in which a consumer reads and reports his own energy consumption, has been presented as an efficient alternative for power companies. In such context, this work presents a solution for self-reading via chatbot in chatting applications. This solution is under development as part of a research and development (R&D) project. It is integrated with a method based on image processing that automatically reads the energy consumption and recognizes the identification code of a meter for validation purposes. Furthermore, all processes utilize cognitive services from the IBM Watson platform to recognize intentions in the dialog with the consumers. The dataset used to validate the proposed method for self-reading contains examples of analogical and digital meters used by Equatorial Energy group. Preliminary results presented accuracies of 77.20% and 84.30%, respectively, for the recognition of complete reading sequences and identification codes in digital meters and accuracies of 89% and 95.20% in the context of analogical meters. Considering both meter types, the method obtains an accuracy per digit of 97%. The proposed method was also evaluated with UFPR-AMR public dataset and achieves a result comparable to the state of the art.

Sign language is a physical language that enables people with disabilities to communicate using hand and facial gestures. For this reason, it is very important for people with disabilities to expre...

Image Restoration is one of the challenging and essential milestones in the image processing domain. Digital image processing is a technique for manipulating digital images using a variety of computer algorithms. The process of transforming the degraded or damaged image to the original image can be known as Image Restoration. The image restoration process improves image quality by converting the degraded image into the original clean image. The techniques for image restoration are comprised of predefined parameters through which digital image gets processed for refinements. The purpose of restoration is to start with the acquired image and then estimate the original image as accurately as possible. A degraded image can be contaminated by any of a blur or noise or both. Many factors can contribute to image degradation, including poor capture, poor lighting, and poor eyesight. Medical science, defensive sensor systems, forensic detections, and astrology all rely on image restoration for accuracy. This paper discusses various image restoration techniques using recent trends for performance improvements.

The article is devoted to the questions of cultural heritage preservation by creating the digital collection of book monuments. The original documents are monuments of book culture and their dilapidated state requires careful handling, splitting of documents for scanning is extremely undesirable. The market does not present the equipment for contactless scanning of books without embroidering, therefore an algorithm that allows digitalizing book monuments in a contactless way has been developed. The technique has been constructed using an algorithm based on the projection of the light grid on the object scanned. The authors propose a sequence of actions consisting of image processing and comparing the results between two images. The first snapshot determines the initial parameters of the grid; the second snapshot determines the actual distortion of the test snapshot. Subsequent mathematical processing allows getting scanned images without absence of geometric distortions of the scanned page due to the system of using the two-dimensional array of corrections. The application of the system has been modeled on the example of «The legend of the destruction of Siberian cities of Tara and Tyumen by the lesser Tatars / / Collection of moral stories, words, lives and other articles [hand.]». The evaluation parameters of the simulation result have been the following: text distinctness, absence of geometric distortions, color quality, uniformity of document scanning quality within a single book, etc., as checked and recognized as high by the experts.The experience described opens possibilities of book monuments digitization using the new algorithm. The development of the system is aimed at expanding the database of objects of material culture to be digitized, perfecting the software, improving the quality of digital images, as well as the capabilities of image recognition and search for the document itself and information it contains.

Breast cancer is one of the most common malignancies. Pathological image processing of breast has become an important means for early diagnosis of breast cancer. Using medical image processing to assist doctors to detect potential breast cancer as early as possible has always been a hot topic in the field of medical image diagnosis. In this paper, a breast cancer recognition method based on image processing is systematically expounded from four aspects: breast cancer detection, image segmentation, image registration, and image fusion. The achievements and application scope of supervised learning, unsupervised learning, deep learning, CNN, and so on in breast cancer examination are expounded. The prospect of unsupervised learning and transfer learning for breast cancer diagnosis is prospected. Finally, the privacy protection of breast cancer patients is put forward.

Deep learning algorithms have begun to be used in medical image processing studies, especially in the last decade. MRI is used in the diagnosis of Alzheimer’s disease, a type of dementia disease, which is the 7th among the diseases that cause death in the world. Alzheimer’s disease has no known cure in the literature, so it is important to attempt treatment before starting the irreversible path by diagnosing the pre-illness stages. In this study, the previous stages of Alzheimer’s disease were classified as normal, mild cognitive impairment, and Alzheimer’s disease through brain MRIs. Different models using CNN architecture were used to classify 2182 image objects obtained from the ADNI database. The study was presented in a very comprehensive comparison framework, and the performances of 29 different pre-trained models on images were evaluated. The accuracy values of each model and the precision, specificity, and sensitivity rates of each class were determined. In the study, the EfficientNetB0 model provided the highest accuracy at the test stage with an accuracy rate of 92.98%. In the comparative evaluation stage with the confusion matrix, the highest rates of precision, sensitivity, and specificity values of the Alzheimer’s disease class were achieved by EfficientNetB3 (89.78%), EfficientNetB2 (94.42%), and EfficientNetB3 (97.28%) models, respectively. The results of the study showed that among the pre-trained models, EfficientNet models achieved a high rate of classification performance as the models with the highest performance. This study will contribute to clinical studies in early prevention by detecting Alzheimer’s disease before it occurs.

It is crucial to ensure proper surveillance for the safety and security of people and their assets. The development of an aerial surveillance system might be very effective in catering to the challenges in surveillance systems. Current systems are expensive and complex. A cost-effective and efficient solution is required, which is easily accessible to anyone with a moderate budget. In aerial surveillance, quadcopters are equipped with state-of-the-art image processing technology that captures detailed photographs of every object underneath. A quadcopter-based solution is proposed to monitor desired premises for any unusual activities, like the movement of persons with weapons and face detection to achieve the desired surveillance. After detection of any unusual activity, the proposed system generates an alert for security personals. The proposed solution is based on quadcopter surveillance and video streaming for anomaly detection in the received video streams through deep learning models. A well-known FasterRCNN algorithm is modified for fast learning with feature reduction in the initial feature extraction step. Five different kinds of CNNs were evaluated for their ability to identify objects of interest in surveillance images. ResNet-50–based FasterRCNN with the highest average precision performed as an excellent solution for threat detection. The average precision of the system is 79% across all categories achieved.

A fully event-based image processing pipeline containing neuromorphic vision sensors and spiking neural network has the potential to achieve high throughput, low latency and high dynamic range vision processing. In this work, we present an end-to-end SNN unsupervised learning inference framework to achieve near-real-time processing performance. The design uses fully event-driven operations that significantly improve learning and inference speed: over $100\times $ increase of inference throughput on CPU and near-real-time inference on GPU for neuromorphic vision sensors can be achieved. The event-driven processing method supports unsupervised spike-timing-dependent plasticity learning of convolutional SNN. When labels are limited, it achieves higher accuracy than supervised training approaches. In addition, the proposed method improves robustness for low-precision SNN as it reduces spiking activity distortion and achieves higher learning accuracy than regular discrete-time simulated low-precision networks.

The increase in computational power in recent years has opened a new door for image processing techniques. Three-dimensional object recognition, identification, pose estimation, and mapping are becoming popular. The need for real-world objects to be mapped into three-dimensional spatial representation is greatly increasing, especially considering the heap jump we obtained in the past decade in virtual reality and augmented reality. This paper discusses an algorithm to convert an array of captured images into estimated 3D coordinates of their external mappings. Elementary methods for generating three-dimensional models are also discussed. This framework will help the community in estimating three-dimensional coordinates of a convex-shaped object from a series of two-dimension images. The built model could be further processed for increasing the resemblance of the input object in terms of its shapes, contour, and texture.

The most basic feature of an image is edge, which is the junction of one attribute area and another attribute area in the image. It is the most uncertain place in the image and the place where the image information is most concentrated. The edge of an image contains rich information. So, the edge location plays an important role in image processing, and its positioning method directly affects the image effect. In order to further improve the accuracy of edge location for multidimensional image, an edge location method for multidimensional image based on edge symmetry is proposed. The method first detects and counts the edges of multidimensional image, sets the region of interest, preprocesses the image with the Gauss filter, detects the vertical edges of the filtered image, and superposes the vertical gradient values of each pixel in the vertical direction to obtain candidate image regions. The symmetry axis position of the candidate image region is analyzed, and its symmetry intensity is measured. Then, the symmetry of vertical gradient projection in the candidate image region is analyzed to verify whether the candidate region is a real edge region. The multidimensional pulse coupled neural network (PCNN) model is used to synthesize the real edge region after edge symmetry processing, and the result of edge location of the multidimensional image is obtained. The results show that the method has strong antinoise ability, clear edge contour, and precise location.

Background: Accurate quantification of bone, muscle, and their components is still an unmet need in the musculoskeletal field. Current methods to quantify tissue volumes in 3D images are expensive, labor-intensive, and time-consuming; thus, a reliable, valid, and quick application is highly needed.Methods: Tissue Compass is a standalone software for semiautomatic segmentation and automatic quantification of musculoskeletal organs. To validate the software, cross-sectional micro-CT scans images of rat femur (n=19), and CT images of hip and abdomen (n=100) from the Osteoporotic Fractures in Men (MrOS) Study were used to quantify bone, hematopoietic marrow (HBM), and marrow adipose tissue (MAT) using commercial manual software as a comparator. Also, abdominal CT scans (n=100) were used to quantify psoas muscle volumes and intermuscular adipose tissue (IMAT) using the same software. We calculated Pearson's correlation coefficients, individual intra-class correlation coefficients (ICC), and Bland-Altman limits of agreement together with Bland-Altman plots to show the inter- and intra-observer agreement between Tissue Compass and commercially available software.Results: In the animal study, the agreement between Tissue Compass and commercial software was r>0.93 and ICC>0.93 for rat femur measurements. Bland-Altman limits of agreement was -720.89 (-1.5e+04, 13074.00) for MAT, 4421.11 (-1.8e+04, 27149.73) for HBM and -6073.32 (-2.9e+04, 16388.37) for bone. The inter-observer agreement for QCT human study between two observers was r>0.99 and ICC>0.99. Bland-Altman limits of agreement was 0.01 (-0.07, 0.10) for MAT in hip, 0.02 (-0.08, 0.12) for HBM in hip, 0.05 (-0.15, 0.25) for bone in hip, 0.02 (-0.18, 0.22) for MAT in L1, 0.00 (-0.16, 0.16) for HBM in L1, 0.02 (-0.23, 0.27) for bone in L1. The intra-observer agreement for QCT human study between two applications was r>0.997 and ICC>0.99. Bland-Altman limits of agreement was 0.03 (-0.13, 0.20) for MAT in hip, 0.05 (-0.08, 0.18) for HBM in hip, 0.05 (-0.24, 0.34) for bone in hip, -0.02 (-0.34, 0.31) for MAT in L1, -0.14 (-0.44, 0.17) for HBM in L1, -0.29 (-0.62, 0.05) for bone in L1, 0.03 (-0.08, 0.15) for IMAT in psoas, and 0.02 (-0.35, 0.38) for muscle in psoas. Conclusion: Compared to a conventional application, Tissue Compass demonstrated high accuracy and non-inferiority while also facilitating easier analyses. Tissue Compass could become the tool of choice to diagnose tissue loss/gain syndromes in the future by requiring a small number of CT sections to detect tissue volumes and fat infiltration.

Geologists infer many issues associated with the formation of the Earth, depositional history and weathering processes based on rock assessment. Preparation of the thin sections of rocks and investigating these sections is one of the common methods in studying rocks. Since all further studies on the sections require the segmentation of existing minerals in thin sections, the accurate and automatic identification of minerals is very important. In this research, for the intelligent identification of minerals, digital images were prepared in ordinary and then polarized light of the minerals. The minerals were then classified by color features, and texture using support vector machine classification techniques. The result of applying the proposed algorithm on 75 thin sections consisting of 42 minerals show an accuracy of 99.25% in minerals segmentation using the polynomial-based support vector machine classification. Also, our proposed algorithm shows the best performance with other techniques by fivefold and tenfold cross-validation. The results indicate that the proposed method outperforms the state-of-the-art method in mineral thin section identification. In addition to the high accuracy, the proper speed of this method also shows the significant efficiency of the proposed method for the automatic identification of minerals.

Image processing is a very broad field containing various areas, including image super-resolution (ISR) which re-represents a low-resolution image as a high-resolution one through a certain means of image transformation. The problem with most of the existing ISR methods is that they are devised for the condition in which sufficient training data is expected to be available. This article proposes a new approach for sparse data-based (rather than sufficient training data-based) ISR, by the use of an ANFIS (Adaptive Network-based Fuzzy Inference System) interpolation technique. Particularly, a set of given image training data is split into various subsets of sufficient and sparse training data subsets. Typical ANFIS training process is applied for those subsets involving sufficient data, and ANFIS interpolation is employed for the rest that contains sparse data only. Inadequate work is available in the current literature for the sparse data-based ISR. Consequently, the implementations of the proposed sparse data-based approach, for both training and testing processes, are compared with the state-of-the-art sufficient data-based ISR methods. This is of course very challenging, but the results of experimental evaluation demonstrate positively about the efficacy of the work presented herein.