Previous Image Research Articles

Temporal Subtraction Technique for Thoracic MDCT Based on Residual VoxelMorph

The temporal subtraction technique is a useful tool for computer aided diagnosis (CAD) in visual screening. The technique subtracts the previous image set from the current one for the same subject to emphasize temporal changes and/or new abnormalities. However, it is difficult to obtain a clear subtraction image without subtraction image artifacts. VoxelMorph in deep learning is a useful method, as preparing large training datasets is difficult for medical image analysis, and the possibilities of incorrect learning, gradient loss, and overlearning are concerns. To overcome this problem, we propose a new method for generating temporal subtraction images of thoracic multi-detector row computed tomography (MDCT) images based on ResidualVoxelMorph, which introduces a residual block to VoxelMorph to enable flexible positioning at a low computational cost. Its high learning efficiency can be expected even with a limited training set by introducing residual blocks to VoxelMorph. We performed our method on 84 clinical images and evaluated based on three-fold cross-validation. The results showed that the proposed method reduced subtraction image artifacts on root mean square error (RMSE) by 11.3% (p < 0.01), and its effectiveness was verified. That is, the proposed temporal subtraction method for thoracic MDCT improves the observer’s performance.

Lighting the darkness in the sea: A deep learning model for underwater image enhancement

Currently, optical imaging cameras are widely used on underwater vehicles to obtain images and support numerous marine exploration tasks. Many underwater image enhancement algorithms have been proposed in the past few years to suppress backscattering noise and improve the signal-to-noise ratio of underwater images. However, these algorithms are mainly focused on underwater image enhancement tasks in a bright environment. Thus, it is still unclear how these algorithms would perform on images acquired in an underwater scene with low illumination. Images obtained in a dark underwater scene usually include more noise and have very low visual quality, which may easily lead to artifacts during the process of enhancement. To bridge this gap, we thoroughly study the existing underwater image enhancement methods and low illumination image enhancement methods based on deep learning and propose a new underwater image enhancement network to solve the problem of serious degradation of underwater image quality in a low illumination environment. Due to the lack of ready-made datasets for training, we also propose the first dataset for low-light underwater image enhancement to train our model. Our method can be implemented to skillfully and simultaneously address low-light degradation and scattering degradation in low-light underwater images. Experimental results also show that our method is robust against different illumination levels, which greatly expands the applicable scenarios of our method. Compared with previous underwater image enhancement methods and low-light image enhancement methods, outstanding performance is achieved using our method in various low-light underwater scenes.

Regulating Cyberspace in Vietnam: Entry, Struggle, and Gain

This study explores the evolution of cyber regulation in Vietnam since its inception, that is from the events of January 1997, when cyberspace first arrived in Vietnam, to the momentous protests instigated by the Cybersecurity Draft Law in June 2018. A Vietnamese cyber regulatory regime is imagined as an analytically constructed regulatory space where different actors enter, struggle, and gain in their pursuit of regulatory interests. The study argues that cyberspace in contemporary Vietnam has aided non-state actors to participate in the law-making and regulatory processes by inducing state actors to respond with cyber laws, regulatory approaches, and measures. Moving beyond the dichotomy of cyberspace as an inevitable tool for liberation or oppression, Vietnamese cyberspace has been both an instrument for non-state actors to participate in lawmaking, and a regulatory measure for state actors to regain control. A sociological landscape in contemporary Vietnam is depicted through the evolution of a Vietnamese cyber regulatory regime, shaped by dynamic interactions between domestic actors. In sharp contrast to the previous image of an authoritarian Vietnam, cyberspace has aided contemporary Vietnam to metamorphose into a more pluralistic society where organically formed social actors co-regulate cyberspace.

SPINDLE CELL TUMOR INVOLVING THE JAW OF A YOUNG PATIENT

An 18-year-old male patient without previous comorbidities reported pain in the left side of the jaw. Clinically, 34 and 35 teeth mobility were observed. Radiographic images exhibited a poorly defined board radiolucent lesion extending from tooth 33 to 35 with tooth resorption. Antimicrobial and analgesic therapy were prescribed, and the patient was referred to specialized care. Computed tomography and resonance evaluation confirmed previous image findings bringing the diagnostic hypothesis of lateral periodontal cyst or other neoplasm of odontogenic origin. Lesion curettage was performed, complemented by margins ostectomy and cryotherapy. Teeth 33, 34, and 35 were lost. Histopathologic analysis revealed spindle shaped cells proliferation with wavy nuclei, diagnosing a fusocellular neoplasia suggestive of fibroma and/or neurofibroma. Immunohistochemistry analysis was requested. The results showed positive staining to EMA (weak and focal) and to smooth muscle actin and negative staining to AE1/AE3, desmin, CD34, S100, and ALK1. The final diagnosis was inflammatory myofibroblastic tumor. An 18-year-old male patient without previous comorbidities reported pain in the left side of the jaw. Clinically, 34 and 35 teeth mobility were observed. Radiographic images exhibited a poorly defined board radiolucent lesion extending from tooth 33 to 35 with tooth resorption. Antimicrobial and analgesic therapy were prescribed, and the patient was referred to specialized care. Computed tomography and resonance evaluation confirmed previous image findings bringing the diagnostic hypothesis of lateral periodontal cyst or other neoplasm of odontogenic origin. Lesion curettage was performed, complemented by margins ostectomy and cryotherapy. Teeth 33, 34, and 35 were lost. Histopathologic analysis revealed spindle shaped cells proliferation with wavy nuclei, diagnosing a fusocellular neoplasia suggestive of fibroma and/or neurofibroma. Immunohistochemistry analysis was requested. The results showed positive staining to EMA (weak and focal) and to smooth muscle actin and negative staining to AE1/AE3, desmin, CD34, S100, and ALK1. The final diagnosis was inflammatory myofibroblastic tumor.

Massive-Scale Aerial Photo Categorization by Cross-Resolution Visual Perception Enhancement.

Categorizing aerial photographs with varied weather/lighting conditions and sophisticated geomorphic factors is a key module in autonomous navigation, environmental evaluation, and so on. Previous image recognizers cannot fulfill this task due to three challenges: 1) localizing visually/semantically salient regions within each aerial photograph in a weakly annotated context due to the unaffordable human resources required for pixel-level annotation; 2) aerial photographs are generally with multiple informative attributes (e.g., clarity and reflectivity), and we have to encode them for better aerial photograph modeling; and 3) designing a cross-domain knowledge transferal module to enhance aerial photograph perception since multiresolution aerial photographs are taken asynchronistically and are mutually complementary. To handle the above problems, we propose to optimize aerial photograph's feature learning by leveraging the low-resolution spatial composition to enhance the deep learning of perceptual features with a high resolution. More specifically, we first extract many BING-based object patches (Cheng et al., 2014) from each aerial photograph. A weakly supervised ranking algorithm selects a few semantically salient ones by seamlessly incorporating multiple aerial photograph attributes. Toward an interpretable aerial photograph recognizer indicative to human visual perception, we construct a gaze shifting path (GSP) by linking the top-ranking object patches and, subsequently, derive the deep GSP feature. Finally, a cross-domain multilabel SVM is formulated to categorize each aerial photograph. It leverages the global feature from low-resolution counterparts to optimize the deep GSP feature from a high-resolution aerial photograph. Comparative results on our compiled million-scale aerial photograph set have demonstrated the competitiveness of our approach. Besides, the eye-tracking experiment has shown that our ranking-based GSPs are over 92% consistent with the real human gaze shifting sequences.

ImageGCN: Multi-Relational Image Graph Convolutional Networks for Disease Identification With Chest X-Rays.

Image representation is a fundamental task in computer vision. However, most of the existing approaches for image representation ignore the relations between images and consider each input image independently. Intuitively, relations between images can help to understand the images and maintain model consistency over related images, leading to better explainability. In this paper, we consider modeling the image-level relations to generate more informative image representations, and propose ImageGCN, an end-to-end graph convolutional network framework for inductive multi-relational image modeling. We apply ImageGCN to chest X-ray images where rich relational information is available for disease identification. Unlike previous image representation models, ImageGCN learns the representation of an image using both its original pixel features and its relationship with other images. Besides learning informative representations for images, ImageGCN can also be used for object detection in a weakly supervised manner. The experimental results on 3 open-source x-ray datasets, ChestX-ray14, CheXpert and MIMIC-CXR demonstrate that ImageGCN can outperform respective baselines in both disease identification and localization tasks and can achieve comparable and often better results than the state-of-the-art methods.

Computer Science Diagram Understanding with Topology Parsing

Diagram is a special form of visual expression for representing complex concepts, logic, and knowledge, which widely appears in educational scenes such as textbooks, blogs, and encyclopedias. Current research on diagrams preliminarily focuses on natural disciplines such as Biology and Geography, whose expressions are still similar to natural images. In this article, we construct the first novel geometric type of diagrams dataset in Computer Science field, which has more abstract expressions and complex logical relations. The dataset has exhaustive annotations of objects and relations for about 1,300 diagrams and 3,500 question-answer pairs. We introduce the tasks of diagram classification (DC) and diagram question answering (DQA) based on the new dataset, and propose the Diagram Paring Net (DPN) that focuses on analyzing the topological structure and text information of diagrams. We use DPN-based models to solve DC and DQA tasks, and compare the performances to well-known natural images classification models and visual question answering models. Our experiments show the effectiveness of the proposed DPN-based models on diagram understanding tasks, also indicate that our dataset is more complex compared to previous natural image understanding datasets. The presented dataset opens new challenges for research in diagram understanding, and the DPN method provides a novel perspective for studying such data. Our dataset can be available from https://github.com/WayneWong97/CSDia.

Temporal-Variation Skeleton Point Correction Algorithm for Improved Accuracy of Human Action Recognition

The technique of human action recognition can be applied in a number of fields, such as medical rehabilitation, posture in the domain of sports, and emotion perception in counselling. Deep learning action recognition models that consider the continuous changes in the human skeleton can efficiently identify action states, such as rehabilitation processes, posture, and changes in emotion. However, the keypoints of the human skeleton are susceptible to occlusion due to body movement, resulting in a lower confidence level in their positions, which in turn affects the training and execution of human action recognition models. This paper proposes a keypoint correction algorithm based on continuous-time images of temporal variation that can predict and correct the skeleton keypoints with low confidence, based on the previous and subsequent multiple images. The proposed temporal-variation skeleton keypoint correction algorithm can provide accurate feature association training for the locations of the skeleton keypoints. The proposed method is shown to improve the accuracy of human action recognition by at least 30% over three alternative models in the literature: STV-GCN, GCN-NAS, and 2S-AGCN. The proposed algorithm is also improved by at least 30% compared with the above-mentioned methods in the literature, based on an action recognition model with a Gaussian filter.

Impact of tomosynthesis on the evolution of the cancer detection rate in the French National Breast Cancer Screening Program

ObjectivesThis study aims to evaluate the cancer detection rate in French National Breast Cancer Screening Program, especially the cancer detection rate during second reading session (Reading 2) based on digital technologies used in radiology centres. Study designThis was an analytical and descriptive study. MethodsCancer detection rate was estimated by the ratio between the number of cancers detected and the number of women screened. The positive predictive value (PPV) was estimated as cancer detection rate among abnormal Reading 2. The relationship between Reading 2's PPV and the predictive factors was evaluated using multilevel mixed-effects logistic regression. ResultsA total of 1,380,006 digital mammograms were retained in the analysis between 2010 and 2019. Cancer detection rate represented 7.8‰ at first reading session (Reading 1) and 0.5‰ at Reading 2. Cancer detection rate is significantly associated with the use of tomosynthesis (P < 0.001) at Reading 1, and differences appear within different tomosynthesis brands (P = 0.007). Reading 2's PPV differs significantly according on technologies used by first Reader (P < 0.004). Nevertheless, Reading 2 has 1.9 (1.5–2.4) more likely to predict a cancer with the presence of previous mammogram compared with those without previous images. ConclusionUsing tomosynthesis technology improves cancer detection rate at Reading 1, even if differences are noticeable between brands. Using tomosynthesis technology at Reading 1 reduces Reading 2's PPV and cancer detection rate at Reading 2.

Natural Language Processing in Diagnostic Texts from Nephropathology.

Introduction: This study investigates whether it is possible to predict a final diagnosis based on a written nephropathological description—as a surrogate for image analysis—using various NLP methods. Methods: For this work, 1107 unlabelled nephropathological reports were included. (i) First, after separating each report into its microscopic description and diagnosis section, the diagnosis sections were clustered unsupervised to less than 20 diagnostic groups using different clustering techniques. (ii) Second, different text classification methods were used to predict the diagnostic group based on the microscopic description section. Results: The best clustering results (i) could be achieved with HDBSCAN, using BoW-based feature extraction methods. Based on keywords, these clusters can be mapped to certain diagnostic groups. A transformer encoder-based approach as well as an SVM worked best regarding diagnosis prediction based on the histomorphological description (ii). Certain diagnosis groups reached F1-scores of up to 0.892 while others achieved weak classification metrics. Conclusion: While textual morphological description alone enables retrieving the correct diagnosis for some entities, it does not work sufficiently for other entities. This is in accordance with a previous image analysis study on glomerular change patterns, where some diagnoses are associated with one pattern, but for others, there exists a complex pattern combination.

Influence of CT image processing on the predicted impact of pores on fatigue of additively manufactured Ti6Al4V and AlSi10Mg

AbstractPores are inherent to additively manufactured components and critical especially in technical components. Since they reduce the component's fatigue life, a reliable identification and description of pores is vital to ensure the component's performance. X‐ray computed tomography (CT) is an established and non‐destructive testing method to investigate internal defects. The CT scan process can induce noise and artefacts in the resulting images which afterwards have to be reduced through image processing. To reconstruct the internal defects of a component, the images need to be segmented in defect region and bulk material by applying a threshold. The application of the threshold as well as the previous image processing alter the geometry and size of the identified defects. This contribution aims to quantify the influence of selected commercial image processing and segmentation methods on identified pores in several additively manufactured components made of AlSi10Mg and Ti6Al4V as well as in an artificial CT scan. To that aim, gray value histograms and characteristic parameters thereof are compared for different image processing tools. After the segmentation of the processed images, particle characteristics are compared. The influence of image processing and segmentation on the predicted fatigue life of the material is evaluated through the change of the largest pore in each set of data applying Murakami's empirical ‐parameter model.

An Efficient Diffusion Approach for Chaos-Based Image Encryption and DNA Sequences

Experts and researchers in the field of information security have placed a high value on the security of image data in the last few years. They have presented several image encryption techniques that are more secure. To increase the security level of image encryption algorithms, this article offers an efficient diffusion approach for image encryption methods based on one-dimensional Logistic, three-dimensional Lorenz, DNA encoding and computing, and SHA-256. The encryption test demonstrates that the method has great security and reliability. This article, also, examines the security of encryption methods, such as secret key space analysis, key sensitivity test, histogram analysis, information entropy process, correlation examination, and differential attack. When the image encryption method described in this article is compared to several previous image encryption techniques, the encryption algorithm has higher information entropy and a lower correlation coefficient.

Сетевая концепция метадисциплинарной платформы конструирования множественных реальных и виртуальных миров

In recent years, the problem concerning mastering augmented and virtual reality has arisen in a wide range of areas of modern life. The paper analyzes the need, requirements and possibilities of metadisciplinary (i.e. non-disciplinary) construction in a generalized space that combines reality and virtuality in educational sphere. At the same time, complexity of combining the worlds, volume of data, and consideration of dynamics are contemplated as problems available for computational solution and practical implementation on the basis of a network platform. Within the framework of the proposed concept, an entrance to the network sub-space containing network imprints of a generalized real-virtual world, coordination of multiple components in it and return back to a previous image, perceived through the developed application of the mixed world is provided. The network platform includes both the network elements of ontology (general description) and representation, storage and processing of data using graph non-relational databases and tools for analyzing and visualizing networks. An architecture with a network component of data processing (networkalization) makes it possible to move forward in the direction of improving the design platform of multiple real and virtual worlds. Special emphasis on the network paradigm is due to the practical value of the ―networked‖ mixed reality (and especially in the educational field) which lies in the fact that information is presented in a well-structured form. It can be expected that tools capable of performing such structuring not only technologically (which is necessary for recognition, tracking and visualization), but also at user level, will be a good way to support educational space.

Unsupervised Appearance Map Abstraction for Indoor Visual Place Recognition With Mobile Robots

Visual Place Recognition (VPR), the task of identifying the place where an image has been taken from, is at the core of important robotic problems as relocalization, loop-closure detection or topological navigation. Even for indoors, the focus of this work, VPR is challenging for a number of reasons, including real-time performance when dealing with large image databases ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\sim10^{4}$</tex-math></inline-formula> ) (probably captured by different robots), or the avoidance of Perceptual Aliasing in environments with repetitive structures and scenes. In this letter, we tackle these issues by proposing an off-line mapping technique that abstracts a dense database of georeferenced images without particular order into a Multivariate Gaussian Mixture Model, by creating soft clusters in terms of their similarity in both pose and appearance. This abstract representation is obtained through an Expectation-Maximization algorithm and plays the role of a simplified map. Since querying this map yields a probability of being in a cluster, we exploit this “belief” within a Bayesian filter that regards previous query images and a topological map between clusters to perform more robust VPR. We evaluate our proposal in two different indoor datasets, demonstrating comparable VPR precision to querying the full database while incurring in shorter query times and handling Perceptual Aliasing for sequential navigation.

GUDCP: Generalization of Underwater Dark Channel Prior for Underwater Image Restoration

This letter introduces an underwater image enhancement method to handle low contrast and color cast of underwater images. Firstly, with the help of hierarchical searching technique, we propose a novel backscattered light estimation method. And in this procedure, a novel scoring formula is considered into our method, which comprehensively considers multiple prior knowledge. Then, we generalize underwater dark channel prior (UDCP) approach to obtain more robust transmission estimation. In addition, we also develop a white balance method to further modify the appearance of the resultant image. Extensive experiments on real-world images demonstrate that the proposed method outperforms several previous image restoration or enhancement works.

A Novel Medical Image Segmentation Model with Domain Generalization Approach

In deep learning-based computing vision for image processing, image segmentation is a prominent issue. There is promising generalisation performance in the medical image segmentation sector for approaches using domain generalisation (DG). Single domain generalisation (SDG) is a more difficult problem than conventional generalisation (DG), which requires numerous source domains to be accessible during network training, as opposed to conventional generalisation (DG). Color medical images may be incorrectly segmented because of the augmentation of the full image in order to increase model generalisation capacity. An arbitrary illumination SDG model for improving generalisation power for colour image segmentation approach for medical images through synthesizing random radiance charts is presented as a first solution to this challenge. Color medical images may be decomposed into reflectivity and illumination maps using retinex-based neural networks (ID-Nets). In order to provide medical colour images under various lighting situations, illumination randomization is used to enhance illumination maps. A new metric, TGCI, called the transfer gradient consistency index was devised to quantify the performance of the breakdown of retinal images by simulating physical lighting. Two of the existing retinal image segmentation tasks are tested extensively in order to assess our suggested system. According to the Dice coefficient, our framework surpasses previous SDGs and image improvement algorithms, outperforming the best SDGs by up to 1.7 per cent.

Shadow Generation for Composite Image in Real-World Scenes

Image composition targets at inserting a foreground object into a background image. Most previous image composition methods focus on adjusting the foreground to make it compatible with background while ignoring the shadow effect of foreground on the background. In this work, we focus on generating plausible shadow for the foreground object in the composite image. First, we contribute a real-world shadow generation dataset DESOBA by generating synthetic composite images based on paired real images and deshadowed images. Then, we propose a novel shadow generation network SGRNet, which consists of a shadow mask prediction stage and a shadow filling stage. In the shadow mask prediction stage, foreground and background information are thoroughly interacted to generate foreground shadow mask. In the shadow filling stage, shadow parameters are predicted to fill the shadow area. Extensive experiments on our DESOBA dataset and real composite images demonstrate the effectiveness of our proposed method. Our dataset and code are available at https://github.com/bcmi/Object-Shadow-Generation- Dataset-DESOBA.

Comprehensive reversible secret image sharing with palette cover images

The comprehensive reversible cover images make secret image sharing (SIS) has more applicable scenarios, such as remote-sensing images, medical diagnosis images and so on. The previous reversible secret image sharing (RSIS) methods use binary or grayscale images as cover images. In this paper, we design an effective reversible secret image sharing method using palette cover images (RSISP). Shadow pixel values are screened to keep their 4 higher bits identical with the corresponding cover image pixel value. In the recovery phase, the original cover images are easily recovered by a bit extraction operation according to the palettes. As the experiment results show, the shadow images are comprehensive and have no pixel expansion. In addition, both the secret image and cover images can be reconstructed losslessly.

Abstract 4121: Durable stable disease with nivolumab, PD-1 inhibitor, and subsequent bevacizumab, Anti-VEGF monoclonal antibody, in long standing recurrent respiratory papillomatosis with lung squamous cell carcinoma

Abstract Background: Recurrent respiratory papillomatosis (RRP) is a human papilloma virus (HPV) associated benign disease, which, in rare cases, can develop into lung squamous cell carcinoma. Treatment options are limited with no approved systemic adjuvant therapy and the disease requires frequent surgical resection to prevent airway obstruction. Therefore, expanding available treatment options remains a highly important, but unmet need in RRP. Herein, we present a case of RRP with squamous cell carcinoma of lung that demonstrated durable stable disease with nivolumab and subsequent bevacizumab. Case presentation: A 48-year-old woman with long standing RRP since teenager, who underwent multiple microsuspension laryngoscopies, debridements including laryngeal lesions, and ultimately, 5 years ago, video assisted thoracic surgery (VATS) resection of right upper lobe mass, confirming squamous cell carcinoma, has been managed with debulking surgery on tracheal lesions by ENT on regular basis. Nivolumab, a programmed death-1 blockade, 3 mg/kg every 2 weeks was started given squamous cell carcinoma of lung. Stable disease (SD) in lung lesions was achieved on nivolumab for one year but she required more frequent debulking of tracheal lesions. After multidisciplinary discussion with Otorhinolaryngology, she was started on bevacizumab, vascular endothelial growth factor inhibitor, 10 mg/kg every 6 weeks. Follow-up CT scans in 10 weeks showed a decrease in size of a left lower lobe mass to 27 mm from 32 mm in the previous image. After one year, administration of bevacizumab was spaced out to every 12 weeks. She requires less debulking surgery and the last debulking surgery was approximately 3 years ago. She maintains SD for over 4 years since starting bevacizumab. Discussion: Our patient with RRP and squamous cell carcinoma of lung demonstrated SD on nivolumab before she was switched to bevacizumab. In addition, her SD on bevacizumab of more than 4 years with 12 weeks of maintenance interval is encouraging as there is no current standard of care treatment for RRP. Despite difficulty in running clinical trials in RRP due to its rareness, additional studies evaluating the efficacy of PD-1 inhibitor or anti-VEGF monoclonal antibody on RRP are warranted. Citation Format: Horyun Choi, Jinah Kim, Yeun Ho Lee, Leeseul Kim, Young Kwang Chae. Durable stable disease with nivolumab, PD-1 inhibitor, and subsequent bevacizumab, Anti-VEGF monoclonal antibody, in long standing recurrent respiratory papillomatosis with lung squamous cell carcinoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 4121.

Formation and Schema Analysis of Oil Painting Style Based on Texture and Color Texture Features under Few Shot.

Texture has strong expressiveness in picture art, and color texture features play an important role in composition. Together with texture, they can convey the artistic connotation of portrait, especially in oil painting. Therefore, in order to make the picture form oil painting style and oil painting schema, we need to study the texture and color texture in combination with the previous oil painting art images. But now, there are few samples of good oil paintings, so it is difficult to study the texture and color texture in oil paintings. Therefore, in order to form a unique artistic style of modern oil painting and promote the development of modern oil painting art, this paper studies the texture and color texture characteristics in the environment of few oil painting works. This paper establishes a model through deep neural network to extract the image incentive and color texture of oil painting art works, which provides guidance for promoting the development of oil painting art. The experiments in this paper show that the depth neural network has high definition for the extraction of texture and color texture of small sample oil painting images, which can reach more than 85%. It has high guiding significance for the research and creation of oil painting art.