J. Imaging, Volume 8, Issue 10 (October 2022) – 39 articles

In a linear tomosynthesis scanner designed for imaging histologic samples of several centimeters size at 10 µm resolution, the mechanical instability of the scanning stage (±10 µm) exceeded the resolution of the image system, making it necessary to determine the trajectory of the stage for each scan to avoid blurring and artifacts in the images that would arise from the errors in the geometric information used in 3D reconstruction. We present a method for online calibration by attaching a layer of randomly dispersed micro glass beads or calcium particles to the bottom of the sample stage. The method was based on a parametric representation of the rigid body motion of the sample stage-marker layer assembly. The marker layer was easy to produce and proven effective in the calibration procedure. Full article

A retinal vessel analysis is a procedure that can be used as an assessment of risks to the eye. This work proposes an unsupervised multimodal approach that improves the response of the Frangi filter, enabling automatic vessel segmentation. We propose a filter that computes pixel-level vessel continuity while introducing a local tolerance heuristic to fill in vessel discontinuities produced by the Frangi response. This proposal, called the local-sensitive connectivity filter (LS-CF), is compared against a naive connectivity filter to the baseline thresholded Frangi filter response and to the naive connectivity filter response in combination with the morphological closing and to the current approaches in the literature. The proposal was able to achieve competitive results in a variety of multimodal datasets. It was robust enough to outperform all the state-of-the-art approaches in the literature for the OSIRIX angiographic dataset in terms of accuracy and 4 out of 5 works in the case of the IOSTAR dataset while also outperforming several works in the case of the DRIVE and STARE datasets and 6 out of 10 in the CHASE-DB dataset. For the CHASE-DB, it also outperformed all the state-of-the-art unsupervised methods. Full article

The pattern of expansion of endodontic lesions in the jaws has been less commonly addressed in the literature. For this reason, the aim of this study is to assess the pattern of endodontic lesions of maxillary and mandibular posterior teeth using cone-beam computed tomography (CBCT). This cross-sectional study was conducted on 317 endodontic lesions of posterior teeth on CBCT scans retrieved from a radiology center in Qazvin, Iran, from 2020 to 2022. Endodontic lesions were assessed on sagittal, coronal, and axial sections by an endodontist and dental student using the Romexis software. The largest lesion diameter was measured occluso-apically, mesiodistally, and buccolingually. Lesion size was analyzed based on age, gender, jaw, tooth type, and presence/absence of root filling by independent samples t-tests and a one-way Analysis Of Variannce (ANOVA). The largest diameter of lesions in the maxilla and mandible was recorded in the occluso-apical dimension followed by buccolingual and mesiodistal dimensions (p > 0.05). The pattern of lesions was the same in teeth with and without endodontic treatment, but it was significantly different in maxillary and mandibular endodontically treated teeth in the occluso-apical and buccolingual dimensions (p < 0.05). No significant correlation was noted with tooth type or jaw except for maxillary and mandibular first molar lesions, which were significantly different in the occluso-apical dimension (p < 0.05). Lesion size in all three dimensions was significantly greater in males than females (p < 0.05), and was the highest in the occluso-apical dimension in both genders. In the maxilla, the mean lesion size significantly decreased in the mesiodistal dimension with age (p < 0.05). In conclusion, the largest lesion diameter in the maxilla and mandible was found in the occluso-apical dimension, indicating the role of bone density in the pattern of lesions. Full article

This paper describes a new open data set, consisting of images of a chessboard collected underwater with different refractive indices, which allows for investigation of the quality of different radial distortion correction methods. The refractive index is regulated by the degree of salinity of the water. The collected data set consists of 662 images, and the chessboard cell corners are manually marked for each image (for a total of $35,748$ nodes). Two different mobile phone cameras were used for the shooting: telephoto and wide-angle. With the help of the collected data set, the practical applicability of the formula for correction of the radial distortion that occurs when the camera is submerged underwater was investigated. Our experiments show that the radial distortion correction formula makes it possible to correct images with high precision, comparable to the precision of classical calibration algorithms. We also show that this correction method is resistant to small inaccuracies in the indication of the refractive index of water. The data set, as well as the accompanying code, are publicly available. Full article

This article is devoted to the study of the improved neighbor mean interpolation (INMI) steganographic method. To date, no steganalysis of such a method of information embedding has been carried out. We implemented the INMI method of embedding messages in raster files and conducted a stegoanalysis on a set of 800 images of 225 × 225 size. Experimentally, we found that with this embedding method, the maximum container capacity is 21% and that it depends on the contents of the container. It is established that only 60 files out of 800 actually have the maximum capacity. We presented the calculation of the Type I error and the percentage of information detection in the obtained containers by the regular–singular (RS) method. The results showed that the considered steganographic algorithm is resistant to RS steganalysis and is comparable to the stegosystem of the permutation method investigated in one of our previous articles. Full article

Clipping, as a fundamental process in computer graphics, displays only the part of a scene which is needed to be displayed and rejects all others. In two dimensions, the clipping process can be applied to a variety of geometric primitives such as points, lines, polygons or curves. A line-clipping algorithm processes each line in a scene through a series of tests and intersection calculations to determine whether the entire line or any part of it is to be saved. It also calculates the intersection position of a line with the window edges so its major goal is to minimize these calculations. This article surveys important techniques and algorithms for line-clipping in 2D but it also includes some of the latest research made by the authors. The survey criteria include evaluation of all line-clipping algorithms against a rectangular window, line clipping versus polygon clipping, and our line clipping against a convex polygon, as well as all line-clipping algorithms against a convex polygon algorithm. Full article

Recent advances in object detection facilitated by deep learning have led to numerous solutions in a myriad of fields ranging from medical diagnosis to autonomous driving. However, historical research is yet to reap the benefits of such advances. This is generally due to the low number of large, coherent, and annotated datasets of historical documents, as well as the overwhelming focus on Optical Character Recognition to support the analysis of historical documents. In this paper, we highlight the importance of visual elements, in particular illustrations in historical documents, and offer a public multi-class historical visual element dataset based on the Sphaera corpus. Additionally, we train an image extraction model based on YOLO architecture and publish it through a publicly available web-service to detect and extract multi-class images from historical documents in an effort to bridge the gap between traditional and computational approaches in historical studies. Full article

Recently, learning algorithms such as Convolutional Neural Networks have been successfully applied in different stages of data processing from the acquisition to the data analysis in the imaging context. The aim of these algorithms is the dimensionality of data reduction and the computational effort, to find benchmarks and extract features, to improve the resolution, and reproducibility performances of the imaging data. Currently, no Neutron Imaging combined with learning algorithms was applied on cultural heritage domain, but future applications could help to solve challenges of this research field. Here, a review of pioneering works to exploit the use of Machine Learning and Deep Learning models applied to X-ray imaging and Neutron Imaging data processing is reported, spanning from biomedicine, microbiology, and materials science to give new perspectives on future cultural heritage applications. Full article

Early and accurate detection of scalp hair loss is imperative to provide timely and effective treatment plans to halt further progression and save medical costs. Many techniques have been developed leveraging deep learning to automate the hair loss detection process. However, the accuracy and robustness of assessing hair loss severity still remain a challenge and barrier for transitioning such a technique into practice. The presented work proposes an efficient and accurate algorithm to classify hair follicles and estimate hair loss severity, which was implemented and validated using a multitask deep learning method via a Mask R-CNN framework. A microscopic image of the scalp was resized, augmented, then processed through pre-trained ResNet models for feature extraction. The key features considered in this study concerning hair loss severity include the number of hair follicles, the thickness of the hair, and the number of hairs in each hair follicle. Based on these key features, labeling of hair follicles (healthy, normal, and severe) were performed on the images collected from 10 men in varying stages of hair loss. More specifically, Mask R-CNN was applied for instance segmentation of the hair follicle region and to classify the hair follicle state into three categories, following the labeling convention (healthy, normal and severe). Based on the state of each hair follicle captured from a single image, an estimation of hair loss severity was determined for that particular region of the scalp, namely local hair loss severity index (P), and by combining P of multiple images taken and processed from different parts of the scalp, we constructed the hair loss severity estimation ($P_{avg}$) and visualized in a heatmap to illustrate the overall hair loss type and condition. The proposed hair follicle classification and hair loss severity estimation using Mask R-CNN demonstrated a more efficient and accurate algorithm compared to other methods previously used, enhancing the classification accuracy by 4 to 15%. This performance supports its potential for use in clinical settings to enhance the accuracy and efficiency of current hair loss diagnosis and prognosis techniques. Full article

Mineral building materials suffer from weathering processes such as salt efflorescence, freeze–thaw cycling, and microbial colonization. All of these processes are linked to water (liquid and vapor) in the pore space. The degree of damage following these processes is heavily influenced by pore space properties such as porosity, pore size distribution, and pore connectivity. X-ray computed micro-tomography (µCT) has proven to be a valuable tool to non-destructively investigate the pore space of stone samples in 3D. However, a trade-off between the resolution and field-of-view often impedes reliable conclusions on the material’s properties. X-ray dark-field imaging (DFI) is based on the scattering of X-rays by sub-voxel-sized features, and as such, provides information on the sample complementary to that obtained using conventional µCT. In this manuscript, we apply X-ray dark-field tomography for the first time on four mineral building materials (quartzite, fired clay brick, fired clay roof tile, and carbonated mineral building material), and investigate which information the dark-field signal entails on the sub-resolution space of the sample. Dark-field tomography at multiple length scale sensitivities was performed at the TOMCAT beamline of the Swiss Light Source (Villigen, Switzerland) using a Talbot grating interferometer. The complementary information of the dark-field modality is most clear in the fired clay brick and roof tile; quartz grains that are almost indistinguishable in the conventional µCT scan are clearly visible in the dark-field owing to their low dark-field signal (homogenous sub-voxel structure), whereas the microporous bulk mass has a high dark-field signal. Large (resolved) pores on the other hand, which are clearly visible in the absorption dataset, are almost invisible in the dark-field modality because they are overprinted with dark-field signal originating from the bulk mass. The experiments also showed how the dark-field signal from a feature depends on the length scale sensitivity, which is set by moving the sample with respect to the grating interferometer. Full article

Nowadays, image analysis has a relevant role in most scientific and research areas. This process is used to extract and understand information from images to obtain a model, knowledge, and rules in the decision process. In the case of biological areas, images are acquired to describe the behavior of a biological agent in time such as cells using a mathematical and computational approach to generate a system with automatic control. In this paper, MCF7 cells are used to model their growth and death when they have been injected with a drug. These mammalian cells allow understanding of behavior, gene expression, and drug resistance to breast cancer. For this, an automatic segmentation method called GEMA is presented to analyze the apoptosis and confluence stages of culture by measuring the increase or decrease of the image area occupied by cells in microfluidic devices. In vitro, the biological experiments can be analyzed through a sequence of images taken at specific intervals of time. To automate the image segmentation, the proposed algorithm is based on a Gabor filter, a coefficient of variation (CV), and linear regression. This allows the processing of images in real time during the evolution of biological experiments. Moreover, GEMA has been compared with another three representative methods such as gold standard (manual segmentation), morphological gradient, and a semi-automatic algorithm using FIJI. The experiments show promising results, due to the proposed algorithm achieving an accuracy above 90% and a lower computation time because it requires on average 1 s to process each image. This makes it suitable for image-based real-time automatization of biological lab-on-a-chip experiments. Full article

Ascending thoracic aortic aneurysm is a life-threatening disease, which is difficult to detect prior to the occurrence of a catastrophe. Epidemiology patterns of ascending thoracic aortic dilations/aneurysms remain understudied, whereas the risk assessment of it may be improved. The electronic databases PubMed/Medline 1966–2022, Web of Science 1975–2022, Scopus 1975–2022, and RSCI 1994–2022 were searched. The current guidelines recommend a purely aortic diameter-based assessment of the thoracic aortic aneurysm risk, but over 80% of the ascending aorta dissections occur at a size that is lower than the recommended threshold of 55 mm. Moreover, a 55 mm diameter criterion could exclude a vast majority (up to 99%) of the patients from preventive surgery. The authors review several visualization-based and alternative approaches which are proposed to better predict the risk of dissection in patients with borderline dilated thoracic aorta. The imaging-based assessments of the biomechanical aortic properties, the Young’s elastic modulus, the Windkessel function, compliance, distensibility, wall shear stress, pulse wave velocity, and some other parameters have been proposed to improve the risk assessment in patients with ascending thoracic aortic aneurysm. While the authors do not argue for shifting the diameter threshold to the left, they emphasize the need for more personalized solutions that integrate the imaging data with the patient’s genotypes and phenotypes in this heterogeneous pathology. Full article

This study is inspired by the widely used algorithm for real-time optical flow, the sparse Lucas–Kanade, by applying a feature extractor to decrease the computational requirement of optical flow based neural networks from real-world thermal aerial imagery. Although deep-learning-based algorithms have achieved state-of-the-art accuracy and have outperformed most traditional techniques, most of them cannot be implemented on a small multi-rotor UAV due to size and weight constraints on the platform. This challenge comes from the high computational cost of these techniques, with implementations requiring an integrated graphics processing unit with a powerful on-board computer to run in real time, resulting in a larger payload and consequently shorter flight time. For navigation applications that only require a 2D optical flow vector, a dense flow field computed from a deep learning neural network contains redundant information. A feature extractor based on the Shi–Tomasi technique was used to extract only appropriate features from thermal images to compute optical flow. The state-of-the-art RAFT-s model was trained with a full image and with our proposed alternative input, showing a substantial increase in speed while maintain its accuracy in the presence of high thermal contrast where features could be detected. Full article

The pulmonary valve (PV) is the least imaged among the heart valves. However, pulmonary regurgitation (PR) and pulmonary stenosis (PS) can occur in a variety of patients ranging from fetuses, newborns (e.g., tetralogy of Fallot) to adults (e.g., endocarditis, carcinoid syndrome, complications of operated tetralogy of Fallot). Due to their complexity, PR and PS are studied using multimodality imaging to assess their mechanism, severity, and hemodynamic consequences. Multimodality imaging is crucial to plan the correct management and to follow up patients with pulmonary valvulopathy. Echocardiography remains the first line methodology to assess patients with PR and PS, but the information obtained with this technique are often integrated with cardiac magnetic resonance (CMR) and computed tomography (CT). This state-of-the-art review aims to provide an updated overview of the usefulness, strengths, and limits of multimodality imaging in patients with PR and PS. Full article

Background: Radiomics extracts quantitative image features to identify biomarkers for characterizing disease. Our aim was to characterize the ability of radiomic features extracted from magnetic resonance (MR) imaging of the liver and spleen to detect cirrhosis by comparing features from patients with cirrhosis to those without cirrhosis. Methods: This retrospective study compared MR-derived radiomic features between patients with cirrhosis undergoing hepatocellular carcinoma screening and patients without cirrhosis undergoing intraductal papillary mucinous neoplasm surveillance between 2015 and 2018 using the same imaging protocol. Secondary analyses stratified the cirrhosis cohort by liver disease severity using clinical compensation/decompensation and Model for End-Stage Liver Disease (MELD). Results: Of 167 patients, 90 had cirrhosis with 68.9% compensated and median MELD 8. Combined liver and spleen radiomic features generated an AUC 0.94 for detecting cirrhosis, with shape and texture components contributing more than size. Discrimination of cirrhosis remained high after stratification by liver disease severity. Conclusions: MR-based liver and spleen radiomic features had high accuracy in identifying cirrhosis, after stratification by clinical compensation/decompensation and MELD. Shape and texture features performed better than size features. These findings will inform radiomic-based applications for cirrhosis diagnosis and severity. Full article

The real-life scene images exhibit a range of variations in text appearances, including complex shapes, variations in sizes, and fancy font properties. Consequently, text recognition from scene images remains a challenging problem in computer vision research. We present a scene text recognition methodology by designing a novel feature-enhanced convolutional recurrent neural network architecture. Our work addresses scene text recognition as well as sequence-to-sequence modeling, where a novel deep encoder–decoder network is proposed. The encoder in the proposed network is designed around a hierarchy of convolutional blocks enabled with spatial attention blocks, followed by bidirectional long short-term memory layers. In contrast to existing methods for scene text recognition, which incorporate temporal attention on the decoder side of the entire architecture, our convolutional architecture incorporates novel spatial attention design to guide feature extraction onto textual details in scene text images. The experiments and analysis demonstrate that our approach learns robust text-specific feature sequences for input images, as the convolution architecture designed for feature extraction is tuned to capture a broader spatial text context. With extensive experiments on ICDAR2013, ICDAR2015, IIIT5K and SVT datasets, the paper demonstrates an improvement over many important state-of-the-art methods. Full article

In the beverage, food and drug industry, more and more machine vision systems are being used for the defect detection of Polyethylene Terephthalate (PET) bottle caps. In this paper, in order to address the result of cylindrical distortions that influence the subsequent defect detection in the imaging process, a very fast image stitching algorithm is proposed to generate a panorama planar image of the surface of PET bottle caps. Firstly, the three-dimensional model of the bottle cap is established. Secondly, the relative poses among the four cameras and the bottle cap in the three-dimensional space are calculated to obtain the mapping relationship between three-dimensional points on the side surface of the bottle cap and image pixels taken by the camera. Finally, the side images of the bottle cap are unfolded and stitched to generate a planar image. The experimental results demonstrate that the proposed algorithm unfolds the side images of the bottle cap correctly and very fast. The average unfolding and stitching time for 1.6-megapixel color caps image can reach almost 123.6 ms. Full article

The paper presents a model of structured objects in a grayscale or color image, described by means of optimal piecewise constant image approximations, which are characterized by the minimum possible approximation errors for a given number of pixel clusters, where the approximation error means the total squared error. An ambiguous image is described as a non-hierarchical structure but is represented as an ordered superposition of object hierarchies, each containing at least one optimal approximation in g₀ = 1, 2,..., etc., colors. For the selected hierarchy of pixel clusters, the objects-of-interest are detected as the pixel clusters of optimal approximations, or as their parts, or unions. The paper develops the known idea in cluster analysis of the joint application of Ward’s and K-means methods. At the same time, it is proposed to modernize each of these methods and supplement them with a third method of splitting/merging pixel clusters. This is useful for cluster analysis of big data described by a convex dependence of the optimal approximation error on the cluster number and also for adjustable object detection in digital image processing, using the optimal hierarchical pixel clustering, which is treated as an alternative to the modern informally defined “semantic” segmentation. Full article

Robot-assisted surgery is becoming popular in the operation room (OR) for, e.g., orthopedic surgery (among other surgeries). However, robotic executions related to surgical steps cannot simply rely on preoperative plans. Using pedicle screw placement as an example, extra adjustments are needed to adapt to the intraoperative changes when the preoperative planning is outdated. During surgery, adjusting a surgical plan is non-trivial and typically rather complex since the available interfaces used in current robotic systems are not always intuitive to use. Recently, thanks to technical advancements in head-mounted displays (HMD), augmented reality (AR)-based medical applications are emerging in the OR. The rendered virtual objects can be overlapped with real-world physical objects to offer intuitive displays of the surgical sites and anatomy. Moreover, the potential of combining AR with robotics is even more promising; however, it has not been fully exploited. In this paper, an innovative AR-based robotic approach is proposed and its technical feasibility in simulated pedicle screw placement is demonstrated. An approach for spatial calibration between the robot and HoloLens 2 without using an external 3D tracking system is proposed. The developed system offers an intuitive AR–robot interaction approach between the surgeon and the surgical robot by projecting the current surgical plan to the surgeon for fine-tuning and transferring the updated surgical plan immediately back to the robot side for execution. A series of bench-top experiments were conducted to evaluate system accuracy and human-related errors. A mean calibration error of 3.61 mm was found. The overall target pose error was 3.05 mm in translation and ${1.12}^{\circ }$ in orientation. The average execution time for defining a target entry point intraoperatively was 26.56 s. This work offers an intuitive AR-based robotic approach, which could facilitate robotic technology in the OR and boost synergy between AR and robots for other medical applications. Full article

The intrinsic features of documents, such as paper color, texture, aging, translucency, the kind of printing, typing or handwriting, etc., are important with regard to how to process and enhance their image. Image binarization is the process of producing a monochromatic image having its color version as input. It is a key step in the document processing pipeline. The recent Quality-Time Binarization Competitions for documents have shown that no binarization algorithm is good for any kind of document image. This paper uses a sample of the texture of the scanned historical documents as the main document feature to select which of the 63 widely used algorithms, using five different versions of the input images, totaling 315 document image-binarization schemes, provides a reasonable quality-time trade-off. Full article

Visual assessment based on intraoperative 2D X-rays remains the predominant aid for intraoperative decision-making, surgical guidance, and error prevention. However, correctly assessing the 3D shape of complex anatomies, such as the spine, based on planar fluoroscopic images remains a challenge even for experienced surgeons. This work proposes a novel deep learning-based method to intraoperatively estimate the 3D shape of patients’ lumbar vertebrae directly from sparse, multi-view X-ray data. High-quality and accurate 3D reconstructions were achieved with a learned multi-view stereo machine approach capable of incorporating the X-ray calibration parameters in the neural network. This strategy allowed a priori knowledge of the spinal shape to be acquired while preserving patient specificity and achieving a higher accuracy compared to the state of the art. Our method was trained and evaluated on 17,420 fluoroscopy images that were digitally reconstructed from the public CTSpine1K dataset. As evaluated by unseen data, we achieved an 88% average F1 score and a 71% surface score. Furthermore, by utilizing the calibration parameters of the input X-rays, our method outperformed a counterpart method in the state of the art by 22% in terms of surface score. This increase in accuracy opens new possibilities for surgical navigation and intraoperative decision-making solely based on intraoperative data, especially in surgical applications where the acquisition of 3D image data is not part of the standard clinical workflow. Full article

Tissue phantoms are important for medical research to reduce the use of animal or human tissue when testing or troubleshooting new devices or technology. Development of machine-learning detection tools that rely on large ultrasound imaging data sets can potentially be streamlined with high quality phantoms that closely mimic important features of biological tissue. Here, we demonstrate how an ultrasound-compliant tissue phantom comprised of multiple layers of gelatin to mimic bone, fat, and muscle tissue types can be used for machine-learning training. This tissue phantom has a heterogeneous composition to introduce tissue level complexity and subject variability in the tissue phantom. Various shrapnel types were inserted into the phantom for ultrasound imaging to supplement swine shrapnel image sets captured for applications such as deep learning algorithms. With a previously developed shrapnel detection algorithm, blind swine test image accuracy reached more than 95% accuracy when training was comprised of 75% tissue phantom images, with the rest being swine images. For comparison, a conventional MobileNetv2 deep learning model was trained with the same training image set and achieved over 90% accuracy in swine predictions. Overall, the tissue phantom demonstrated high performance for developing deep learning models for ultrasound image classification. Full article

The colorization of grayscale images can, nowadays, take advantage of recent progress and the automation of deep-learning techniques. From the media industry to medical or geospatial applications, image colorization is an attractive and investigated image processing practice, and it is also helpful for revitalizing historical photographs. After exploring some of the existing fully automatic learning methods, the article presents a new neural network architecture, Hyper-U-NET, which combines a U-NET-like architecture and HyperConnections to handle the colorization of historical black and white aerial images. The training dataset (about 10,000 colored aerial image patches) and the realized neural network are available on our GitHub page to boost further research investigations in this field. Full article

Defect inspection using imaging-processing techniques, which detects and classifies manufacturing defects, plays a significant role in the quality control of microelectromechanical systems (MEMS) sensors in the semiconductor industry. However, high-precision classification and location are still challenging because the defect images that can be obtained are small and the scale of the different defects on the picture of the defect is different. Therefore, a simple, flexible, and efficient convolutional neural network (CNN) called accurate-detection CNN (ADCNN) to inspect MEMS pressure-sensor-chip packaging is proposed in this paper. The ADCNN is based on the faster region-based CNN, which improved the performance of the network by adding random-data augmentation and defect classifiers. Specifically, the ADCNN achieved a mean average precision of 92.39% and the defect classifier achieved a mean accuracy of 97.2%. Full article

The last two years are considered the most crucial and critical period of the COVID-19 pandemic affecting most life aspects worldwide. This virus spreads quickly within a short period, increasing the fatality rate associated with the virus. From a clinical perspective, several diagnosis methods are carried out for early detection to avoid virus propagation. However, the capabilities of these methods are limited and have various associated challenges. Consequently, many studies have been performed for COVID-19 automated detection without involving manual intervention and allowing an accurate and fast decision. As is the case with other diseases and medical issues, Artificial Intelligence (AI) provides the medical community with potential technical solutions that help doctors and radiologists diagnose based on chest images. In this paper, a comprehensive review of the mentioned AI-based detection solution proposals is conducted. More than 200 papers are reviewed and analyzed, and 145 articles have been extensively examined to specify the proposed AI mechanisms with chest medical images. A comprehensive examination of the associated advantages and shortcomings is illustrated and summarized. Several findings are concluded as a result of a deep analysis of all the previous works using machine learning for COVID-19 detection, segmentation, and classification. Full article

The benefits of autonomous image segmentation are readily apparent in many applications and garners interest from stakeholders in many fields. The wide range of benefits encompass applications ranging from medical diagnosis, where the shape of the grouped pixels increases diagnosis accuracy, to autonomous vehicles where the grouping of pixels defines roadways, traffic signs, other vehicles, etc. It even proves beneficial in many phases of machine learning, where the resulting segmentation can be used as inputs to the network or as labels for training. The majority of the available image segmentation algorithmic development and results focus on visible image modalities. Therefore, in this treatment, the authors present the results of a study designed to identify and improve current semantic methods for infrared scene segmentation. Specifically, the goal is to propose a novel approach to provide tile-based segmentation of occlusion clouds in Long Wave Infrared images. This work complements the collection of well-known semantic segmentation algorithms applicable to thermal images but requires a vast dataset to provide accurate performance. We document performance in applications where the distinction between dust cloud tiles and clear tiles enables conditional processing. Therefore, the authors propose a Gray Level Co-Occurrence Matrix (GLCM) based method for infrared image segmentation. The main idea of our approach is that GLCM features are extracted from local tiles in the image and used to train a binary classifier to provide indication of tile occlusions. Our method introduces a new texture analysis scheme that is more suitable for image segmentation than the solitary Gabor segmentation or Markov Random Field (MRF) scheme. Our experimental results show that our algorithm performs well in terms of accuracy and a better inter-region homogeneity than the pixel-based infrared image segmentation algorithms. Full article

In this article, a new efficient and robust approach—the high-resolution microwave imaging system—for early breast cancer diagnosis is presented. The core concept of the proposed approach is to employ a combination of a newly proposed delay-and-sum (DAS) algorithm and the specific absorption rate (SAR) parameter to provide high image quality of breast tumors, along with fast image processing. The new algorithm enhances the tumor response by altering the parameter referring to the distance between the antenna and the tumor in the conventional DAS matrices. This adjustment entails a much clearer reconstructed image with short processing time. To achieve these aims, a high directional Vivaldi antenna is applied around a simulated hemispherical breast model with an embedded tumor. The detection of the tumor is carried out by calculating the maximum value of SAR inside the breast model. Consequently, the antenna position is relocated near the tumor region and is moved to nine positions in a trajectory path, leading to a shorter propagation distance in the image-creation process. At each position, the breast model is illuminated with short pulses of low power waves, and the back-scattered signals are recorded to produce a two-dimensional image of the scanned breast. Several simulations of testing scenarios for reconstruction imaging are investigated. These simulations involve different tumor sizes and materials. The influence of the number of antennas on the reconstructed images is also examined. Compared with the results from the conventional DAS, the proposed technique significantly improves the quality of the reconstructed images, and it detects and localizes the cancer inside the breast with high quality in a fast computing time, employing fewer antennas. Full article

Multimedia data manipulation and forgery has never been easier than today, thanks to the power of Artificial Intelligence (AI). AI-generated fake content, commonly called Deepfakes, have been raising new issues and concerns, but also new challenges for the research community. The Deepfake detection task has become widely addressed, but unfortunately, approaches in the literature suffer from generalization issues. In this paper, the Face Deepfake Detection and Reconstruction Challenge is described. Two different tasks were proposed to the participants: (i) creating a Deepfake detector capable of working in an “in the wild” scenario; (ii) creating a method capable of reconstructing original images from Deepfakes. Real images from CelebA and FFHQ and Deepfake images created by StarGAN, StarGAN-v2, StyleGAN, StyleGAN2, AttGAN and GDWCT were collected for the competition. The winning teams were chosen with respect to the highest classification accuracy value (Task I) and “minimum average distance to Manhattan” (Task II). Deep Learning algorithms, particularly those based on the EfficientNet architecture, achieved the best results in Task I. No winners were proclaimed for Task II. A detailed discussion of teams’ proposed methods with corresponding ranking is presented in this paper. Full article