Search

This study presents a method for aligning the geometric parameters of images in multi-channel imaging systems based on the application of pre-processing methods, machine learning algorithms, and a calibration setup using an array of orderly markers at the nodes of an imaginary grid. According to the proposed method, one channel of the system is used as a reference. The images from the calibration setup in each channel determine the coordinates of the markers, and the displacements of the marker centers in the system's channels relative to the coordinates of the centers in the reference channel are then determined. Correction models are obtained as multiple polynomial regression models based on these displacements. These correction models align the geometric parameters of the images in the system channels before they are used in the calculations. The models are derived once, allowing for geometric calibration of the imaging system. The developed method is applied to align the images in the channels of a module of a multispectral imaging polarimeter. As a result, the standard image alignment error in the polarimeter channels is reduced from 4.8 to 0.5 pixels. [ABSTRACT FROM AUTHOR]

In grid intelligent inspection systems, automatic registration of infrared and visible light images in power scenes is a crucial research technology. Since there are obvious differences in key attributes between visible and infrared images, direct alignment is often difficult to achieve the expected results. To overcome the high difficulty of aligning infrared and visible light images, an image alignment method is proposed in this paper. First, we use the Sobel operator to extract the edge information of the image pair. Second, the feature points in the edges are recognised by a curvature scale space (CSS) corner detector. Third, the Histogram of Orientation Gradients (HOG) is extracted as the gradient distribution characteristics of the feature points, which are normalised with the Scale Invariant Feature Transform (SIFT) algorithm to form feature descriptors. Finally, initial matching and accurate matching are achieved by the improved fast approximate nearest-neighbour matching method and adaptive thresholding, respectively. Experiments show that this method can robustly match the feature points of image pairs under rotation, scale, and viewpoint differences, and achieves excellent matching results. [ABSTRACT FROM AUTHOR]

Generating naturally stitched images is a challenging task due to the presence of the parallax in images. In this paper, we propose a semantic-based method to handle the distortions and artifacts generated in the process of stitching. Firstly, a semantic-based random sample consensus algorithm is proposed to obtain the corresponding feature points, which can effectively eliminate mismatch points. The obtained corresponding features are used to align the image and establish the global structure information. Then, we propose semantics-preserving warps based on mesh optimization. The global structure information, semantic information and the corresponding features are combined to warp the images. Experimental results show that our algorithm can provide accurate alignment results while effectively reducing image distortions in both overlapping and non-overlapping regions. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Computer Engineering & Applications is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Image alignment is an important problem in computer vision, which can be solved by tensor based methods that are robust to noise and have satisfactory performance. However, these methods face two common challenges: (1) they have high computational cost when dealing with large-scale tensor data; (2) they ignore the local structures within and across images. To overcome these challenges, we propose an efficient data-driven tensor dictionary learning (DTDL) model for image alignment. In our DTDL model, we factorize the underlying third order tensor into a coefficients tensor and three dictionary matrices of smaller sizes, which reduces the dimensionality and complexity of the problem. We also exploit the generalized hyper-Laplacian regularization to preserve the local structures that are embedded in the underlying tensor and represented by the dictionary framework. Furthermore, we prove that our proximal linearized alternating direction method of multipliers algorithm can generate a sequence that converges to a Karush–Kuhn–Tucker point under very mild conditions. We conduct experiments on image alignment and face recognition tasks, and show that our method outperforms state-of-the-art methods in terms of performance and efficiency. [ABSTRACT FROM AUTHOR]

Crop canopy temperature can often be acquired using the thermal imager. Non-contact and non-destructive automated detection can be expected to achieve for crop water stress status. Automatic image alignment can be used to treat the fuzzy edge distribution, strong noise, as well as shape and texture information lacking in thermal infrared images, according to the information complementarity between visible light and thermal infrared images. The automated extraction can be realized on the crop canopy temperature. This study aims to solve the problems of differences in the radiation, shape, and texture between visible light images and thermal infrared images, leading to the low align images of different modalities. Multimodal image registration was also proposed to integrate the improved brain storm optimization (BSO) and Powell algorithm. Firstly, the original visible light image was downsampled and cropped, according to the normalized cross-correlation value. The area with the most similarity region was obtained in the thermal infrared image under the same resolution; Then, the target area was extracted from the cropped image. The target area image and the original thermal infrared image were decomposed by wavelet transform, where the multilayered low-frequency information was retained; Thirdly, the primitive affine transformation matrix was obtained by the image moments in the low-resolution layer; At the same time, the global search was used to optimize the affine transform matrix in the low-resolution layer using the improved BSO; Fourthly, the optimization was used as the initial point of the Powell algorithm. The optimization was performed in the high-resolution layer; Lastly, the optimization in the previous step was input into the Powell algorithm again. The original image layer was optimized again to obtain the final affine transformation matrix. The original BSO optimization was improved for the optimal affine transformation matrix in the image alignment task. The specific improvements included the following five aspects: The BSO population distribution was initialized using a chaotic mapping function; The mutation range of new individual was modified; The number of K-means clusters was dynamically adjusted in the BSO by the elbow; The chaotic local search was incorporated into the strategy of individual variation; and the probability parameters were dynamically adjusted, according to the different BSO in the early and late stages. Mutual information (MI), normalized mutual information (NMI), root mean square error (RMSE) and mean structure similarity index measure (MSSIM) were taken as the evaluation indexes. A comparison was made with Powell optimization, genetic algorithm (GA) and BSO_Powell algorithm. Specifically, MI indexes were improved by 0.054 2, 0.076 9, 0.040 5, respectively; NMI indexes were improved by 0.015 9, 0.023 1, 0.052 7, respectively; RMSE indexes were reduced by 15.02, 13.03, 27.08, respectively; and MSSIM indexes were improved by 0.0523, 0.0488, 0.1224, respectively, in greenhouse data; In field data, MI indexes were improved by 0.064 2, 0.066 7, 0.035 5, respectively; NMI indexes were improved by 0.007 7, 0.0125, 0.0124, respectively; RMSE indexes were reduced by 14.06, 10.57, 15.40, respectively; and MSSIM indexes were improved by 0.047 1, 0.038 1, 0.042 9, respectively. The strong robustness can accurately achieved in the multimodal image registration tasks for potatoes under complex environments. [ABSTRACT FROM AUTHOR]

Featured Application: This paper is applied in the industrial monitoring of the stamping process, where vibrations are common, necessitating the use of image registration and homography estimation methods to align template images with test images. By employing machine-vision and image-processing techniques, the process is monitored in real-time to detect any anomalies, ultimately aiming to protect the stamping molds. Homography estimation is a crucial task in aligning template images with target images in stamping monitoring systems. To enhance the robustness and accuracy of homography estimation against random vibrations and lighting variations in stamping environments, this paper proposes an improved unsupervised homography estimation model. The model takes as input the channel-stacked template and target images and outputs the estimated homography matrix. First, a specialized deformable convolution module and Group Normalization (GN) layer are introduced to expand the receptive field and enhance the model's ability to learn rotational invariance when processing large, high-resolution images. Next, a multi-scale, multi-stage unsupervised homography estimation network structure is constructed to improve the accuracy of homography estimation by refining the estimation through multiple stages, thereby enhancing the model's resistance to scale variations. Finally, stamping monitoring image data is incorporated into the training through data fusion, with data augmentation techniques applied to randomly introduce various levels of perturbation, brightness, contrast, and filtering to improve the model's robustness to complex changes in the stamping environment, making it more suitable for monitoring applications in this specific industrial context. Compared to traditional methods, this approach provides better homography matrix estimation when handling images with low texture, significant lighting variations, or large viewpoint changes. Compared to other deep-learning-based homography estimation methods, it reduces estimation errors and performs better on stamping monitoring images, while also offering broader applicability. [ABSTRACT FROM AUTHOR]

This paper introduces a method aimed at diagnosing the presence or absence of lesions by detecting anomalies in Chest X-ray images. The proposed approach is based on the PatchCore anomaly detection method, which extracts a feature vector containing location information of an image patch from normal image data and calculates the anomaly distance from the normal vector. However, applying PatchCore directly to medical image processing presents challenges due to the possibility of diseases occurring only in specific organs and the presence of image noise unrelated to lesions. In this study, we present an image alignment method that utilizes affine transformation parameter prediction to standardize already captured X-ray images into a specific composition. Additionally, we introduce a region-specific abnormality detection method that requires affine-transformed chest X-ray images. Furthermore, we propose a method to enhance application efficiency and performance through feature map hard masking. The experimental results demonstrate that our proposed approach achieved a maximum AUROC (Area Under the Receiver Operating Characteristic) of 0.774. Compared to a previous study conducted on the same dataset, our method shows a 6.9% higher performance and improved accuracy. [ABSTRACT FROM AUTHOR]

Image alignment represents a crucial and essential subject in computer vision applications for image analysis. Getting spatial transformation to align a moving image with a reference image is the aim of image alignment. Deep learning techniques, which have been more and more popular recently, provide good outcomes when applied to alignment challenges in addition to many other computer vision problems. In this work, a supervised DL technique has been used in order to estimate the spatial transformation parameter. The spatial transformation model is based on the stiff technique. To convert moving images to a fixed image, rigid transformation parameters are estimated using a supervised convolutional neural network (CNN). The primary contribution of the presented research is to use a model to handle input images with quality degradation to carry out supervised rigid image alignment with the regression model of the CNNs. In the study, many parameters have been examined in an attempt to ascertain the impact of noise in each image and the parameters that yield the optimal outcomes for the problem. [ABSTRACT FROM AUTHOR]

Ensuring a safe nighttime environmental perception system relies on the early detection of vulnerable road users with minimal delay and high precision. This paper presents a sensor-fused nighttime environmental perception system by integrating data from thermal and RGB cameras. A new alignment algorithm is proposed to fuse the data from the two camera sensors. The proposed alignment procedure is crucial for effective sensor fusion. To develop a robust Deep Neural Network (DNN) system, nighttime thermal and RGB images were collected under various scenarios, creating a labeled dataset of 32,000 image pairs. Three fusion techniques were explored using transfer learning, alongside two single-sensor models using only RGB or thermal data. Five DNN models were developed and evaluated, with experimental results showing superior performance of fused models over non-fusion counterparts. The late-fusion system was selected for its optimal balance of accuracy and response time. For real-time inferencing, the best model was further optimized, achieving 33 fps on the embedded edge computing device, an 83.33% improvement in inference speed over the system without optimization. These findings are valuable for advancing Advanced Driver Assistance Systems (ADASs) and autonomous vehicle technologies, enhancing pedestrian detection during nighttime to improve road safety and reduce accidents. [ABSTRACT FROM AUTHOR]

This paper presents a new deep-learning architecture designed to enhance the spatial synchronization between CMOS and event cameras by harnessing their complementary characteristics. While CMOS cameras produce high-quality imagery, they struggle in rapidly changing environments—a limitation that event cameras overcome due to their superior temporal resolution and motion clarity. However, effective integration of these two technologies relies on achieving precise spatial alignment, a challenge unaddressed by current algorithms. Our architecture leverages a dynamic graph convolutional neural network (DGCNN) to process event data directly, improving synchronization accuracy. We found that synchronization precision strongly correlates with the spatial concentration and density of events, with denser distributions yielding better alignment results. Our empirical results demonstrate that areas with denser event clusters enhance calibration accuracy, with calibration errors increasing in more uniformly distributed event scenarios. This research pioneers scene-based synchronization between CMOS and event cameras, paving the way for advancements in mixed-modality visual systems. The implications are significant for applications requiring detailed visual and temporal information, setting new directions for the future of visual perception technologies. [ABSTRACT FROM AUTHOR]

Weed control has a significant impact on crop yield during cultivation. In this research, semantic segmentation is used to detect morning glories in soybean fields. By removing morning glory earlier in the growing season, the decrease in soybean crop yield can be minimized. However, it is difficult to create annotated images necessary for semantic segmentation at the growing season because soybeans and morning glories are both green and similar in color, making it difficult to distinguish them. This research assumes that morning glory colonies, once located at the growing season, remain stationary during the harvest season. The colonies of the morning glory at the growing season are identified by aligning the orthomosaic image from the growing season with the orthomosaic image from the harvest season because the leaves of the soybeans wither and turn brown during the harvest season. The proposed method trains a model of morning glory at the growing season based on its location at the harvest season and estimates the colonies of morning glory on the harvest season orthomosaic image. In this research, we investigated the accuracy of a deep learning-based morning glory detection model and discovered that the performance of the model varied depending on the proportion of morning glory areas on each image in the training dataset. The model demonstrated an optimal performance when only 3.5% of the proportion of the morning glory areas achieved an F2 score of 0.753. [ABSTRACT FROM AUTHOR]

Existing homography and optical flow methods are erroneous in challenging scenes, such as fog, rain, night, and snow because the basic assumptions such as brightness and gradient constancy are broken. To address this issue, we present an unsupervised learning approach that fuses gyroscope into homography and optical flow learning. Specifically, we first convert gyroscope readings into motion fields named gyro field. Second, we design a self-guided fusion module (SGF) to fuse the background motion extracted from the gyro field with the optical flow and guide the network to focus on motion details. Meanwhile, we propose a homography decoder module (HD) to combine gyro field and intermediate results of SGF to produce the homography. To the best of our knowledge, this is the first deep learning framework that fuses gyroscope data and image content for both deep homography and optical flow learning. To validate our method, we propose a new dataset that covers regular and challenging scenes. Experiments show that our method outperforms the state-of-the-art methods in both regular and challenging scenes. The code and dataset are available at https://github.com/lhaippp/GyroFlowPlus. [ABSTRACT FROM AUTHOR]

Supervised homography estimation methods face a challenge due to the lack of adequate labeled training data. To address this issue, we propose DMHomo, a diffusion model-based framework for supervised homography learning. This framework generates image pairs with accurate labels, realistic image content, and realistic interval motion, ensuring that they satisfy adequate pairs. We utilize unlabeled image pairs with pseudo labels such as homography and dominant plane masks, computed from existing methods, to train a diffusion model that generates a supervised training dataset. To further enhance performance, we introduce a new probabilistic mask loss, which identifies outlier regions through supervised training, and an iterative mechanism to optimize the generative and homography models successively. Our experimental results demonstrate that DMHomo effectively overcomes the scarcity of qualified datasets in supervised homography learning and improves generalization to real-world scenes. The code and dataset are available at GitHub (https://github.com/lhaippp/DMHomo). [ABSTRACT FROM AUTHOR]

This paper presents a novel approach for natural stitching multiple images by integrating thin-plate spline (TPS), cylindrical projection, and plumb-line constraint. Firstly, the homography estimated under plumb-line constraint is used to transform each image to keep the scene in the image upward as much as possible, so as to suppress the accumulation of image projection deformation and make the transformed images approximately available for cylindrical projection. Then, by introducing cylindrical projection into TPS as a global transformation, a multiple image alignment framework called cylindrical projection thin-plate spline (CP-TPS) is established to accurately align the transformed images. In this step, the virtual control points (VCP) are set in the non-overlapping area of images so that the CP-TPS can produce desired deformation in the final stitched image. Finally, a seam-line intersecting the significant structure in the aligned image is automatically generated by combining TPS, dynamic programming matching, and control points triangulation. In this step, the seam-line itself is used to estimate CP-TPS parameter. Experiments were conducted on four public image sets. The results show that the proposed approach can realize the natural stitching of public multiple image sets and has the best performance, compared with Autostitch, APAP, NISwGSP, ELA, and GES-GSP. [ABSTRACT FROM AUTHOR]

In grid intelligent inspection systems, automatic registration of infrared and visible light images in power scenes is a crucial research technology. Since there are obvious differences in key attributes between visible and infrared images, direct alignment is often difficult to achieve the expected results. To overcome the high difficulty of aligning infrared and visible light images, an image alignment method is proposed in this paper. First, we use the Sobel operator to extract the edge information of the image pair. Second, the feature points in the edges are recognised by a curvature scale space (CSS) corner detector. Third, the Histogram of Orientation Gradients (HOG) is extracted as the gradient distribution characteristics of the feature points, which are normalised with the Scale Invariant Feature Transform (SIFT) algorithm to form feature descriptors. Finally, initial matching and accurate matching are achieved by the improved fast approximate nearest-neighbour matching method and adaptive thresholding, respectively. Experiments show that this method can robustly match the feature points of image pairs under rotation, scale, and viewpoint differences, and achieves excellent matching results.

In this paper, we study the problem of a batch of linearly correlated image alignment, where the observed images are deformed by some unknown domain transformations, and corrupted by additive Gaussian noise and sparse noise simultaneously. By stacking these images as the frontal slices of a third-order tensor, we propose to utilize the tensor factorization method via transformed tensor-tensor product to explore the low-rankness of the underlying tensor, which is factorized into the product of two smaller tensors via transformed tensor-tensor product under any unitary transformation. The main advantage of transformed tensor-tensor product is that its computational complexity is lower compared with the existing literature based on transformed tensor nuclear norm. Moreover, the tensor ℓ p (0 < p < 1) norm is employed to characterize the sparsity of sparse noise and the tensor Frobenius norm is adopted to model additive Gaussian noise. A generalized Gauss-Newton algorithm is designed to solve the resulting model by linearizing the domain transformations, and a proximal Gauss-Seidel algorithm is developed to solve the corresponding subproblem. Furthermore, the convergence of the proximal Gauss-Seidel algorithm is established according to Kurdyka-Łojasiewicz property, whose convergence rate is also analyzed. Extensive numerical examples on real-world image datasets are carried out to demonstrate the superior performance of the proposed method as compared to several state-of-the-art methods in both accuracy and computational time. [ABSTRACT FROM AUTHOR]

Medical professionals in thoracic medicine routinely analyze chest X-ray images, often comparing pairs of images taken at different times to detect lesions or anomalies in patients. This research aims to design a computer-aided diagnosis system that enhances the efficiency of thoracic physicians in comparing and diagnosing X-ray images, ultimately reducing misjudgments. The proposed system encompasses four key components: segmentation, alignment, comparison, and classification of lung X-ray images. Utilizing a public NIH Chest X-ray14 dataset and a local dataset gathered by the Chiayi Christian Hospital in Taiwan, the efficacy of both the traditional methods and deep-learning methods were compared. Experimental results indicate that, in both the segmentation and alignment stages, the deep-learning method outperforms the traditional method, achieving higher average IoU, detection rates, and significantly reduced processing time. In the comparison stage, we designed nonlinear transfer functions to highlight the differences between pre- and post-images through heat maps. In the classification stage, single-input and dual-input network architectures were proposed. The inclusion of difference information in single-input networks enhances AUC by approximately 1%, and dual-input networks achieve a 1.2–1.4% AUC increase, underscoring the importance of difference images in lung disease identification and classification based on chest X-ray images. While the proposed system is still in its early stages and far from clinical application, the results demonstrate potential steps forward in the development of a comprehensive computer-aided diagnostic system for comparative analysis of chest X-ray images. [ABSTRACT FROM AUTHOR]

Summary: Feature extraction is a fundamental step in the feature matching task. A lot of studies are devoted to feature extraction. Recent researches propose to extract features by pre‐trained neural networks, and the output is used for feature matching. However, the quality and the quantity of the features extracted by these methods are difficult to meet the requirements for the practical applications. In this article, we propose a two‐stage object‐aware‐based feature matching method. Specifically, the proposed object‐aware block predicts a weighted feature map through a mask predictor and a prefeature extractor, so that the subsequent feature extractor pays more attention to the key regions by using the weighted feature map. In addition, we introduce a state‐of‐the‐art model estimation algorithm to align image pair as the input of the object‐aware block. Furthermore, our method also employs an advanced outlier removal algorithm to further improve matching quality. Experimental results show that our object‐aware‐based feature matching method improves the performance of feature matching compared with several state‐of‐the‐art methods. [ABSTRACT FROM AUTHOR]

Image alignment is a highly researched topic in computer vision, which aligns a pair of images due to image changes. Despite the numerous studies conducted on this topic, large object transformation and huge illumination changes between a pair of images are still commonly encountered in real-world scenes, making the task of image alignment very challenging. In this paper, a novel image alignment algorithm is proposed. By inputting a pair of images that need to be aligned into the correlated siamese neural network, a series of blocks are extracted in feature layers from the reference image, and those blocks are correlated in the feature layers of the target image. Finally, the homography parameters between images are then regressed from the correlate layers. Compared with the classical image alignment algorithms, supervised deep homography, and unsupervised deep homography, the experimental results of our method demonstrate a superior performance on the image alignment tasks involving illumination changes, camera translation, and rotation. [ABSTRACT FROM AUTHOR]

Dorsal hand vein patterns have gained prominence as a reliable biometric modality for verifying and identifying individuals. They offer a significant alternative to other well-known biometric modalities such as face, iris, or fingerprints, providing unique advantages in terms of security and privacy. While multiple researchers are engaged in this biometric research area, dorsal hand vein-based recognition systems, involving knuckle alignment and matching algorithms, exhibit a distinct set of limitations that motivated our work. Thus, this paper introduces an innovative knuckle alignment method tailored for processing near-infrared (NIR) imaging of dorsal hand vein images, aiming to enhance the precision and robustness of such a biometric system. Our newly proposed efficient knuckle alignment method addresses the variability in hand positioning during capturing. It utilizes the entire dorsal hand vein image without necessitating the extraction of specific regions of interest. Our knuckle alignment method ensures greater consistency across varied samples in terms of alignment accuracy and reduces intra-class variability, which significantly enhances the accuracy and reliability of the resulting biometric system. To evaluate the efficacy of our proposed approach, we conduct comparative analyses among automated, computer vision-based, and traditional manual-based alignment methods. Experimental results demonstrate that by standardizing image orientation, the proposed automated approach improves system performance. In addition, the proposed automated approach is expected to offer substantial benefits in terms of scalability and operational efficiency without compromising the high precision typically associated with manual techniques. Our method yields a verification accuracy score of 99.07% on the JLU dataset and 99.90% on the DHV dataset, which is higher than the competition. These findings underscore the potential of our proposed knuckle alignment method to serve as a valuable tool in the ongoing development and optimization of dorsal hand vein authentication systems.

Homography estimation is a crucial task in aligning template images with target images in stamping monitoring systems. To enhance the robustness and accuracy of homography estimation against random vibrations and lighting variations in stamping environments, this paper proposes an improved unsupervised homography estimation model. The model takes as input the channel-stacked template and target images and outputs the estimated homography matrix. First, a specialized deformable convolution module and Group Normalization (GN) layer are introduced to expand the receptive field and enhance the model’s ability to learn rotational invariance when processing large, high-resolution images. Next, a multi-scale, multi-stage unsupervised homography estimation network structure is constructed to improve the accuracy of homography estimation by refining the estimation through multiple stages, thereby enhancing the model’s resistance to scale variations. Finally, stamping monitoring image data is incorporated into the training through data fusion, with data augmentation techniques applied to randomly introduce various levels of perturbation, brightness, contrast, and filtering to improve the model’s robustness to complex changes in the stamping environment, making it more suitable for monitoring applications in this specific industrial context. Compared to traditional methods, this approach provides better homography matrix estimation when handling images with low texture, significant lighting variations, or large viewpoint changes. Compared to other deep-learning-based homography estimation methods, it reduces estimation errors and performs better on stamping monitoring images, while also offering broader applicability.

Abstract Background Optical coherence tomography (OCT) is the most important and commonly utilized imaging modality in ophthalmology and is especially crucial for the diagnosis and management of macular diseases. Each OCT volume is typically only available as a series of cross-sectional images (B-scans) that are accessible through proprietary software programs which accompany the OCT machines. To maximize the potential of OCT imaging for machine learning purposes, each OCT image should be analyzed en bloc as a 3D volume, which requires aligning all the cross-sectional images within a particular volume. Methods A dataset of OCT B-scans obtained from 48 age-related macular degeneration (AMD) patients and 50 normal controls was used to evaluate five registration algorithms. After alignment of B-scans from each patient, an en face surface map was created to measure the registration quality, based on an automatically generated Laplace difference of the surface map–the smoother the surface map, the smaller the average Laplace difference. To demonstrate the usefulness of B-scan alignment, we trained a 3D convolutional neural network (CNN) to detect age-related macular degeneration (AMD) on OCT images and compared the performance of the model with and without B-scan alignment. Results The mean Laplace difference of the surface map before registration was 27 ± 4.2 pixels for the AMD group and 26.6 ± 4 pixels for the control group. After alignment, the smoothness of the surface map was improved, with a mean Laplace difference of 5.5 ± 2.7 pixels for Advanced Normalization Tools Symmetric image Normalization (ANTs-SyN) registration algorithm in the AMD group and a mean Laplace difference of 4.3 ± 1.4.2 pixels for ANTs in the control group. Our 3D CNN achieved superior performance in detecting AMD, when aligned OCT B-scans were used (AUC 0.95 aligned vs. 0.89 unaligned). Conclusions We introduced a novel metric to quantify OCT B-scan alignment and compared the effectiveness of five alignment algorithms. We confirmed that alignment could be improved in a statistically significant manner with readily available alignment algorithms that are available to the public, and the ANTs algorithm provided the most robust performance overall. We further demonstrated that alignment of OCT B-scans will likely be useful for training 3D CNN models.

In the field of automated fabric defect detection, commonly used machine vision methods rely on manually designed features, making it difficult to detect complex defects. And deep learning methods are prone to interference from fabric patterns, resulting in lower detection accuracy. Therefore, a feature fusion network model was proposed based on Cascade R-CNN. Speeded Up Robust Feature algorithm was applied to generate a difference image of feature without patterns. This image was fused with original defect image to obtain more features; deform-able convolution was applied to strengthen the feature extraction ability of network. Through experimental analysis, the model shows a 10.8% improvement in detection accuracy compared to Cascade R-CNN. [ABSTRACT FROM AUTHOR]

Underwater image stitching is a technique employed to seamlessly merge images with overlapping regions, creating a coherent underwater panorama. In recent years, extensive research efforts have been devoted to advancing image stitching methodologies for both terrestrial and underwater applications. However, existing image stitching methods, which do not utilize detector information, heavily rely on matching feature pairs and tend to underperform in situations where underwater images contain regions with blurred feature textures. To address this challenge, we present an improved scale-invariant feature transform (SIFT) underwater image stitching method. This method enables the stitching of underwater images with arbitrarily acquired images featuring blurred feature contours and that do not require any detector information. Specifically, we perform a coarse feature extraction between the reference and training images, and then we acquire the target image and perform an accurate feature extraction between the reference and target images. In the final stage, we propose an improved fade-in and fade-out fusion method to obtain a panoramic underwater image. The experimental results show that our proposed method demonstrates enhanced robustness, particularly in scenarios where detecting feature points is challenging, when compared to traditional SIFT methods. Additionally, our method achieves higher matching accuracy and produces higher-quality results in the stitching of underwater images. [ABSTRACT FROM AUTHOR]

Premise: Previously we described methods for generating three‐dimensional (3D) virtual reconstructions of plant tissues from transverse thin sections. Here, we report the applicability of longitudinal sections and improved image‐processing steps that are simpler to perform and utilize free applications. Methods: In order to obtain improved digital images and a virtual 3D object (cuboid), GIMP 2.10 and ImageJ 2.3.0 running on a laptop computer were used. Sectional views of the cuboid and 3D visualization were realized with use of the plug‐ins "Volume Viewer" and "3D Viewer" in ImageJ. Results: A 3D object was constructed and sectional views along several cutting planes were generated. The 3D object consisted of selected tissues inside the cuboid that were extracted and visualized from the original section data, and an animated video of the 3D construct was also produced. Discussion: Virtual cuboids can be constructed by stacking longitudinal images along the transverse depth direction or stacking transverse images vertically along the organ axis, with both generating similar 3D objects. Which to use depends on the purpose of the investigation: if the vertical cell structures need close examination, the former method may be better, but for more general spatial evaluations or for evaluation of organs over longer tissue distances than can be accommodated with longitudinal sectioning, the latter method should be chosen. [ABSTRACT FROM AUTHOR]

The integration of digital media art into the public art design of the smart city can enhance people’s participation and sense of identity in the art design of public space and maximize the attractiveness and cultural connotation of the city itself. Based on this, this paper builds a fusion technology framework mainly based on digital projection technology and digital interaction technology, proposes an interactive projection technology integrating digital projection and Kinect, and combines the dynamic scene image splicing and alignment fusion technology to realize the interaction between the user and the digital scene. In the case analysis of Qianjiang New City, the number of visitors reached 31.55 million in 2022. The evaluation score of landscape facilities in the public evaluation of architectural landscape is under 3.5. The overall evaluation grade is only “general,” and the evaluation mean value of the degree of intelligence of landscape facilities is only 3.089. The comprehensive evaluation satisfaction of the digital landscape includes multimedia landscape, architectural landscape, and landscape facilities, and the average value of the evaluation of the degree of intelligence of the landscape facilities is only 3.089. The satisfaction value of the multimedia landscape, landscape of buildings and structures, and water landscape in the evaluation of satisfaction is 4.36, 4.159, and 4.027, respectively, which reaches the level of “good” satisfaction, while the satisfaction value of landscape facilities is 3.479, which is less than 3.5. The satisfaction level only reaches the level of “general.” The satisfaction value of landscape facilities is 3.479, which is less than 3.5, and the satisfaction level is only “fair.”

In this paper, we conduct a preliminary study on the current development status in medical image alignment and build up a basic framework for image alignment. The feature space, search space, similarity measure, and search strategy of cardiac angiography images are calculated and studied. The DGVF model is utilized to process the traditional snake model for optimization search and is combined with B-splines to construct the B-spline DGVF model. Optimize the traditional MsFCM algorithm by using the PSO algorithm and propose an MsFCM-PSO image segmentation method. It is applied together with the B-spline DGVF model to segment the vascular lumen in cardiac angiography ultrasound images. Finally, the model of this paper is analyzed in terms of segmentation performance, alignment stability, and evaluation of alignment results. The mean values of Dice, IoU, and HD of this paper’s MsFCM-PSO model in image segmentation of cardiac vessels are 94.27%, 92.60%, and 1.06, respectively (all optimal performances). In the ablation experiments, the MsFCMPSO model in this paper shows an increase of 6.02% and 5.47% in the mean values of Dice and IoU compared to the benchmark model. The stability calibration percentage of this paper’s MsFCM-PSO algorithm is 31.13% when the Gaussian factor is 0.5, which is significantly better than other algorithms. The algorithm in this paper is better than other methods in terms of alignment stability and alignment results.

The economy and society have entered a highly transformative digital age. Digital Intelligence promotes the change of social and economic forms and promotes the integration and innovation of technology, economy and society. In this paper, we formulate the goal of training talents with digital intelligence thinking and the model of training talents with digital intelligence thinking, and propose a strategy for implementing talent training. Using big data technology (cloud computing resource scheduling method, virtual scene dynamic splicing technology) to assist in the implementation of teaching strategies. Conduct simulation experiments to analyze the performance of the cloud resource scheduling algorithm proposed in this paper on different numbers of virtual machines. Evaluate the effect of virtual scene splicing technology on image alignment. Discuss the influences factors from the three aspects of cultivation objectives, curriculum arrangement, and scientific research training in light of the satisfaction of the implementation of the mathematical and intellectual talent cultivation model. The cultivation objectives for numerical intelligence ability are viewed by 40.61% of the students as average, and 13.75% of them feel completely satisfied. Only 4.31% of the students agreed on scientific research and innovation. It can be seen that most of the students did not carry out any interdisciplinary research in depth. It shows that the implementation of the strategy for cultivating mathematical and intellectual talents still needs improvement and strengthening.

Background: Optical coherence tomography (OCT) is the most important and commonly utilized imaging modality in ophthalmology and is especially crucial for the diagnosis and management of macular diseases. Each OCT volume is typically only available as a series of cross-sectional images (B-scans) that are accessible through proprietary software programs which accompany the OCT machines. To maximize the potential of OCT imaging for machine learning purposes, each OCT image should be analyzed en bloc as a 3D volume, which requires aligning all the cross-sectional images within a particular volume. Methods: A dataset of OCT B-scans obtained from 48 age-related macular degeneration (AMD) patients and 50 normal controls was used to evaluate five registration algorithms. After alignment of B-scans from each patient, an en face surface map was created to measure the registration quality, based on an automatically generated Laplace difference of the surface map–the smoother the surface map, the smaller the average Laplace difference. To demonstrate the usefulness of B-scan alignment, we trained a 3D convolutional neural network (CNN) to detect age-related macular degeneration (AMD) on OCT images and compared the performance of the model with and without B-scan alignment. Results: The mean Laplace difference of the surface map before registration was 27 ± 4.2 pixels for the AMD group and 26.6 ± 4 pixels for the control group. After alignment, the smoothness of the surface map was improved, with a mean Laplace difference of 5.5 ± 2.7 pixels for Advanced Normalization Tools Symmetric image Normalization (ANTs-SyN) registration algorithm in the AMD group and a mean Laplace difference of 4.3 ± 1.4.2 pixels for ANTs in the control group. Our 3D CNN achieved superior performance in detecting AMD, when aligned OCT B-scans were used (AUC 0.95 aligned vs. 0.89 unaligned). Conclusions: We introduced a novel metric to quantify OCT B-scan alignment and compared the effectiveness of five alignment algorithms. We confirmed that alignment could be improved in a statistically significant manner with readily available alignment algorithms that are available to the public, and the ANTs algorithm provided the most robust performance overall. We further demonstrated that alignment of OCT B-scans will likely be useful for training 3D CNN models. [ABSTRACT FROM AUTHOR]

We present a template matching algorithm based on local descriptors for aligning two geospatial products of different modalities with a large area asymmetry. Our system is generic with regards to the modalities of the geospatial products and is applicable to the self-localization of aerial devices such as drones and missiles. This algorithm consists in finding a superposition such that the average dissimilarity of the superposed points is minimal. The dissimilarity of two points belonging to two different geospatial products is the distance between their respective local descriptors. These local descriptors are learned. We performed experiments consisting in estimating a translation between optical (Pléiades) and SAR (Miranda) images onto vector data (OpenStreetMap), onto optical images (DOP) and onto SAR images (KOMPSAT-5). Each remote sensing image to be aligned covered 0.64 km2, and each reference geospatial product spanned over 225 km2. We conducted a total of 381 alignment experiments, with six unique modality combinations. In aggregate, the precision reached was finer than 10 m with 72% probability and finer than 20 m with 96% probability. This is considerably more than with traditional methods such as normalized cross-correlation and mutual information. [ABSTRACT FROM AUTHOR]

To solve the problem of high-resolution image alignment time overhead, an SIFT-based fast image alignment algorithm is presented. The overlap region of images is computed in detail by phase correlation algorithm to avoid a lot of useless calculations of non-overlapping region. After the distribution of feature points determined in difference of Gaussian through formula derivation, the total number of feature points is limited. The more stable spatially distributed for the feature points is obtained due to the expanded detection range of extreme points and added non-maximum suppression. It is noteworthy that the range of the descriptor is calculated by the method of down-sampling. And the circular descriptor is constructed with only 56-dimensional in the feature point descriptor generation stage, which makes the time of the descriptor generation and feature point matching shorter. This indicates that the total descriptor calculation is faster in lower dimensions by the new algorithm. In addition, experimental results show that the average time (9.60s, 13.46s, and 15.81s) of the proposed algorithm is 0.86%, 0.43%, and 0.10% of the SIFT algorithm, respectively. The overall speed is 2–3 orders of magnitude faster than the SIFT algorithm, which indicates that the new algorithm can solve the problem of high-resolution image alignment time overhead. The new algorithm provides a good stitching quality and shows an excellent performance for high-resolution image compared with several existing image stitching algorithms at the current. It indicates that the algorithm has potential application value in real-time image stitching.

Image stitching is a technique that is often employed in image processing and computer vision applications. The feature points in an image provide a significant amount of key information. Image stitching requires accurate extraction of these features since it may decrease misalignment flaws in the final stitched image. In recent years, a variety of feature detectors and descriptors that may be utilized for image stitching have been presented. However, the computational cost and correctness of feature matching restrict the utilization of these techniques. To date, no work compared feature detectors and descriptors for image stitching applications, i.e., no one has considered the effect of detectors and descriptors on the generated final stitched image. This paper presents a detailed comparative analysis of commonly used feature detectors and descriptors proposed previously. This study gives various contributions to the development of a general comparison of feature detectors and descriptors for image stitching applications. These detectors and descriptors are compared in terms of number of matched points, time taken and quality of stitched image. After analyzing the obtained results, it was observed that the combination of AKAZE with AKAZE can be preferable almost in all possible situations. [ABSTRACT FROM AUTHOR]

Printing defects are extremely common in the manufacturing industry. Although some studies have been conducted to detect printing defects, the stability and practicality of the printing defect detection has received relatively little attention. Currently, printing defect detection is susceptible to external environmental interference such as illuminance and noise, which leads to poor detection rates and poor practicality. This research develops a printing defect detection method based on scale-adaptive template matching and image alignment. Firstly, the research introduces a convolutional neural network (CNN) to adaptively extract deep feature vectors from templates and target images at a low-resolution version. Then, a feature map cross-correlation (FMCC) matching metric is proposed to measure the similarity of the feature map between the templates and target images, and the matching position is achieved by a proposed location refinement method. Finally, the matching image and the template are both sent to the image alignment module, so as to detect printing defects. The experimental results show that the accuracy of the proposed method reaches 93.62%, which can quickly and accurately find the location of the defect. Simultaneously, it is also proven that our method achieves state-of-the-art defect detection performance with strong real-time detection and anti-interference capabilities. [ABSTRACT FROM AUTHOR]

The essence of image super-resolution reconstruction technology is to break through the limitation of hardware conditions, and reconstruct a high-resolution image from a low-resolution image which contains less infor-mation through the image super-resolution reconstruction algorithms. With the development of the technology on deep learning, deep learning has been introduced into the image super-resolution reconstruction field. This paper summarizes the image super-resolution reconstruction algorithms based on deep learning, classifies, analyzes and compares the typical algorithms. Firstly, the model framework, upsampling method, nonlinear mapping learning module and loss function of single image super-resolution reconstruction method are introduced in detail. Secondly, the reference-based super-resolution reconstruction method is analyzed from two aspects: pixel alignment and Patch matching. Then, the benchmark datasets and image quality evaluation indices used for image super-resolution recon-struction algorithms are summarized, the characteristics and performance of the typical super-resolution recons-truction algorithms are compared and analyzed. Finally, the future research trend on the image super-resolution reconstruction algorithms based on deep learning is prospected.

Image alignment with parallax is a challenging computer vision problem. While existing methods employing local-varying smooth warps have enhanced alignment accuracy for complex spacial deformations compared to global homography estimation, they fall short in adequately representing the discontinuous transformations evident in images with large parallax. In this paper, we propose an image alignment method grounded in the estimation and fusion of multiple warping models. To effectively tackle the challenges posed by discontinuous deformations in different regions, our method comprises two key stages: principal region alignment and fine region refinement. In both stages, multiple warping models are initially estimated using feature correspondences and are then concurrently optimized by minimizing pixel-level photometric loss. For each pixel, we select the optimal model that minimizes warping error. Additionally, we introduce a feature grouping method based on Delaunay triangulation. Experiments on real-world images demonstrate the superior alignment accuracy achieved by our proposed method compared to other state-of-the-arts. • An image alignment method for discontinuous deformation is proposed. • A feature grouping method for images with local-varying deformation is given. • A procedure of successively aligning principal regions and fine areas is presented. [ABSTRACT FROM AUTHOR]

Underwater imaging often suffers from poor quality due to the complex underwater environment and limitations of hardware equipment, leading to images with shallow depth of field and moving objects, which pose a challenge for information fusion of image sequences from the same underwater scene. To effectively address these problems, we propose a decoupled variational Retinex method for reconstructing and fusing underwater shallow depth of field images. Specifically, we first construct a module that adopts the decoupled variational Retinex model to adjust pixel dynamic range and luminance components, enhance non-local properties' extraction with higher-order data constraints, and significantly improve image quality. Then, we develop a precision alignment strategy for image sequences by calculating and correcting control point deviations in the overlapping areas, achieving accurate registration of the image sequences, and effectively reconstructing scenes with parallax. Moreover, scenes with moving objects within the image sequence are reconstructed by redistributing overlapping areas. We design a novel cost function based on the neighborhood information of seams, which facilitates iterative optimization of these solved seams. This process improves the segmentation accuracy within these regions, achieving more precise scene reconstruction. Compared with state-of-the-art approaches, our method demonstrates superior performance in rectifying degraded image quality and reconstructing visually appealing images, with the resulting reconstructed images showing enhanced subjective visual quality. • Enhances underwater image quality and widens view angles. • Fuses image alignment with seam-driven stitching. • Refines feature points and eliminates outliers. • Reduces artifacts by adjusting moving object positions. • Effective across common underwater scenes with superior visuals. [ABSTRACT FROM AUTHOR]

Underwater image stitching is a technique employed to seamlessly merge images with overlapping regions, creating a coherent underwater panorama. In recent years, extensive research efforts have been devoted to advancing image stitching methodologies for both terrestrial and underwater applications. However, existing image stitching methods, which do not utilize detector information, heavily rely on matching feature pairs and tend to underperform in situations where underwater images contain regions with blurred feature textures. To address this challenge, we present an improved scale-invariant feature transform (SIFT) underwater image stitching method. This method enables the stitching of underwater images with arbitrarily acquired images featuring blurred feature contours and that do not require any detector information. Specifically, we perform a coarse feature extraction between the reference and training images, and then we acquire the target image and perform an accurate feature extraction between the reference and target images. In the final stage, we propose an improved fade-in and fade-out fusion method to obtain a panoramic underwater image. The experimental results show that our proposed method demonstrates enhanced robustness, particularly in scenarios where detecting feature points is challenging, when compared to traditional SIFT methods. Additionally, our method achieves higher matching accuracy and produces higher-quality results in the stitching of underwater images.

Laser speckle imaging (LSI) is a wide-field optical technique that enables superficial blood flow quantification. LSI is normally performed in a mounted configuration to decrease the likelihood of motion artifact. However, mounted LSI systems are cumbersome and difficult to transport quickly in a clinical setting for which portability is essential in providing bedside patient care. To address this issue, we created a handheld LSI device using scientific grade components. To account for motion artifact of the LSI device used in a handheld setup, we incorporated a fiducial marker (FM) into our imaging protocol and determined the difference between highest and lowest speckle contrast values for the FM within each data set (Kbest and Kworst). The difference between Kbest and Kworst in mounted and handheld setups was 8% and 52%, respectively, thereby reinforcing the need for motion artifact quantification. When using a threshold FM speckle contrast value (KFM) to identify a subset of images with an acceptable level of motion artifact, mounted and handheld LSI measurements of speckle contrast of a flow region (KFLOW) in in vitro flow phantom experiments differed by 8%. Without the use of the FM, mounted and handheld KFLOW values differed by 20%. To further validate our handheld LSI device, we compared mounted and handheld data from an in vivo porcine burn model of superficial and full thickness burns. The speckle contrast within the burn region (KBURN) of the mounted and handheld LSI data differed by