Review
Chemistry, Analytical
Amnah Aldayri, Waleed Albattah
Summary: This paper provides a detailed review of recent developments in anomaly detection methods from the perspective of computer vision, based on different available datasets. A new taxonomic organization of existing works in crowd analysis and anomaly detection is introduced. A summary of existing reviews and datasets related to anomaly detection is listed, covering an overview of different crowd concepts, types of anomalies, and surveillance systems. Additionally, research trends and future work prospects are analyzed.
Article
Biology
Roger Resmini, Lincoln Faria da Silva, Petrucio R. T. Medeiros, Adriel S. Araujo, Debora C. Muchaluat-Saade, Aura Conci
Summary: Breast cancer is the second most common cancer worldwide, and early diagnosis and treatment are crucial for patient healing. This paper proposes a hybrid computational method using dynamic and static infrared thermography for breast cancer screening and diagnosis, achieving high accuracy through machine learning techniques. The results demonstrate the potential of the proposed approaches in breast cancer screening and diagnosis.
COMPUTERS IN BIOLOGY AND MEDICINE
(2021)
Article
Engineering, Civil
Yadi Zhu, Ke Ni, Xiaohong Li, Asim Zaman, Xiang Liu, Yun Bai
Summary: This study proposes a generalized artificial intelligence (AI)-based crowd analytics model framework that can effectively calculate the flow volume, crowd density, and walking speed by analyzing and visualizing video records of high-density crowds. The model's accuracy has been validated and the study also shows differences in behavior in different scenarios.
TRANSPORTATION RESEARCH RECORD
(2023)
Article
Computer Science, Interdisciplinary Applications
Konrad Olejnik, Pawel Pelczynski, Kinga Troszczynska, Svitlana Khadzhynova
Summary: The interaction between liquids and materials is crucial in technological processes and material quality. However, assessing the wetting and penetration of liquids through heterogeneous materials is challenging. This study aimed to evaluate the suitability of texture analysis in assessing the uniformity of liquid penetration through paper structures. A new measurement method based on dynamic changes in texture parameters was proposed, using entropy and contrast. The study found correlations between liquid penetration dynamics and examined texture parameters, suggesting that certain parameters can provide information about liquid flow through porous materials.
COMPUTERS & STRUCTURES
(2023)
Article
Construction & Building Technology
Mohammad Hassan Daneshvari, Ebrahim Nourmohammadi, Mahmoud Ameri, Barat Mojaradi
Summary: The primary cause of decreasing road safety, comfort, and service life is the raveling of asphalt pavement. This study proposes and verifies a computer vision technique based on image texture features for automatic detection of asphalt pavement raveling. Two feature extraction scenarios are compared, and the results show that the second scenario offers higher prediction performance.
CONSTRUCTION AND BUILDING MATERIALS
(2023)
Article
Computer Science, Information Systems
Sonal Gore, Jayant Jagtap
Summary: Advanced biomedical texture descriptors combined with machine learning-assisted techniques can accurately identify the 1p/19q codeletion status of glioma patients, showing promising potential for personalized treatment.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2022)
Article
Geosciences, Multidisciplinary
Yuqin Shu, Guibing Zheng, Xiawan Yan
Summary: This study proposes an integrated MGCNN model to assess the waterlogging risk in the primarily urban area of Guangzhou, China. The texture features play an important role in identifying urban waterlogging areas, and the MGCNN model shows better accuracy and interpretability compared to other models.
PHYSICS AND CHEMISTRY OF THE EARTH
(2022)
Article
Remote Sensing
Sophie de Roda Husman, Joost J. van der Sanden, Stef Lhermitte, Marieke A. Eleveld
Summary: This study developed a Random Forest classifier based on multiple features of Sentinel-1 data for classifying three main ice types during river ice breakup. Results show improved classifier performance by including GLCM mean and VH intensity features, highlighting the complementary nature of texture and intensity in ice classification during breakup.
INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
(2021)
Article
Anatomy & Morphology
Zineb Bellahreche, Nesrine Semiane, Aicha Mallek, Yasmina Dahmani
Summary: A histological study on gerbils receiving a high-fat/high-carbohydrate-diet revealed that it leads to structural damage in the adrenal glands, which is time-dependent.
Article
Computer Science, Information Systems
Yingying Jiang, Yiming Miao, Bander Alzahrani, Ahmed Barnawi, Reem Alotaibi, Long Hu
Summary: The ULCM system is a novel ultralarge-scale crowd monitoring system that utilizes advanced sensing and networking technologies to provide decision makers with intelligent support for crowd management by collecting and processing real-time crowding data.
IEEE INTERNET OF THINGS JOURNAL
(2021)
Article
Engineering, Biomedical
R. Barker-Clarke, D. T. Weaver, J. G. Scott
Summary: This study aims to derive analogous texture features for graphs and networks and demonstrate their usefulness in summarizing graphs, aiding comparative graph studies, classifying biological graphs, and detecting dysregulation in cancer. The co-occurrence matrices for graphs are generated to obtain novel graph 'texture' features that reflect graph structure and node label distributions. These texture metrics are sensitive to discretization parameters and noise and can be used to classify cell line expression with high accuracy.
PHYSICS IN MEDICINE AND BIOLOGY
(2023)
Article
Engineering, Biomedical
Sharanagouda Nawaldgi, Y. S. Lalitha
Summary: Glaucoma is a significant cause of blindness globally, as it damages optic nerves in the eyes, leading to vision loss. Early detection and treatment are crucial, as the damage caused by glaucoma is irreversible. However, large-scale glaucoma screening is challenging due to a lack of skilled manpower in ophthalmology. To address this issue, automated glaucoma detection methods have been proposed. In this study, a novel method using structural and texture features from color fundus images (CFI) for automated glaucoma assessment is presented. The results show an overall efficiency of over 89%.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2022)
Article
Criminology & Penology
Lasse Suonpera Liebst, Marie Rosenkrantz Lindegaard, Wim Bernasco
Summary: The study found no support for Collins's hypothesis that emotional dominance leads to violence, but rather found evidence against it. In robbery incidents, it is the absence of emotional dominance that promotes offender violence, not the presence of it. The results suggest that Collins's situational stance could benefit from considering instrumental motivation and premeditation.
JOURNAL OF INTERPERSONAL VIOLENCE
(2021)
Article
Construction & Building Technology
Xun Zhao, Lige Xue, Feiyun Xu
Summary: This paper presents an asphalt paving segregation detection method based on image texture features, which has been validated to have high accuracy and efficiency in the classification of targets with similar texture features, achieving a diagnosis accuracy of 94% for asphalt paving segregation.
CONSTRUCTION AND BUILDING MATERIALS
(2021)
Article
Environmental Sciences
Johannes Lohse, Anthony P. Doulgeris, Wolfgang Dierking
Summary: This study investigates the inclusion of Sentinel-1 texture features in a Bayesian classifier to improve the classification of sea ice types in SAR images. Results show that texture features play a crucial role in classification, especially in the generalized separation of ice and water, as well as the classification of young ice and multi-year ice, leading to significant improvements in classification accuracy.
Article
Computer Science, Information Systems
Dongyu She, Jufeng Yang, Ming-Ming Cheng, Yu-Kun Lai, Paul L. Rosin, Liang Wang
IEEE TRANSACTIONS ON MULTIMEDIA
(2020)
Article
Computer Science, Software Engineering
Ran Song, Yonghuai Liu, Paul L. Rosin
Summary: A novel network trained in a weakly supervised manner, called CfS-CNN, has been developed to solve the difficulty of collecting vertex-level annotation for mesh saliency detection. This network outperforms existing state-of-the-art methods in mesh saliency detection and can be directly applied to scene saliency. Experimental results demonstrate the significant improvement of this approach.
IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS
(2021)
Article
Computer Science, Software Engineering
Mazlum Ferhat Arslan, Alexandros Haridis, Paul L. Rosin, Sibel Tari, Charlotte Brassey, James D. Gardiner, Asli Genctav, Murat Genctav
Summary: This paper presents the results of the SHREC'21 track on quantifying shape complexity. The goal is to evaluate the performance of shape complexity measures and investigate their relationships. The methods are evaluated by computing correlation coefficients between the produced orders and the ground truth. This study provides an improved means for evaluating shape complexity.
COMPUTERS & GRAPHICS-UK
(2022)
Article
Chemistry, Multidisciplinary
Bozhi Zhang, Meijing Gao, Paul L. Rosin, Xianfang Sun, Qiuyue Chang, Qichong Yan, Yucheng Shang
Summary: This paper studied the performance evaluation and optimization theory of thermal microscope imaging systems, analyzed the differences between thermal microscope imaging and telephoto thermal imaging, derived the signal-to-noise ratio expression, and researched the performance evaluation model. Simulation and experiments on different detectors were carried out, providing reference for the performance evaluation and optimization of thermal microscope imaging systems.
APPLIED SCIENCES-BASEL
(2021)
Editorial Material
Computer Science, Software Engineering
Silvia Biasotti, Roberto M. Dyke, Yu-Kun Lai, Paul L. Rosin, Remco Veltkamp
COMPUTERS & GRAPHICS-UK
(2022)
Article
Computer Science, Artificial Intelligence
Ran Yi, Yong-Jin Liu, Yu-Kun Lai, Paul L. Rosin
Summary: This paper introduces a unique and expressive style of art called face portrait line drawing. To automatically transform face photos into portrait drawings, the authors propose a novel method using unpaired training data and the ability to generate drawings in multiple styles and unseen styles. Through the introduction of a new quality metric and quality loss, the authors address the problem of missing important facial features in existing methods due to information imbalance. Experimental results demonstrate that their method outperforms state-of-the-art methods in portrait drawing generation.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Computer Science, Artificial Intelligence
Ran Yi, Mengfei Xia, Yong-Jin Liu, Yu-Kun Lai, Paul L. Rosin
Summary: This paper proposes a composite GAN for transforming face photos to artistic portrait drawings, addressing challenges such as highly abstract styles and different drawing techniques. By introducing novel loss terms and a classification-and-synthesis approach, the method captures the highly abstract art form and improves the line quality. Extensive experiments show that the proposed method outperforms state-of-the-art methods both qualitatively and quantitatively.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2021)
Article
Computer Science, Software Engineering
Yi-Ling Qiao, Lin Gao, Jie Yang, Paul L. Rosin, Yu-Kun Lai, Xilin Chen
Summary: This article introduces a deep learning approach to process 3D meshes, which utilizes Laplacian spectral analysis to encode mesh connectivity and employs mesh feature aggregation blocks to gather local and global information. The method outperforms state-of-the-art algorithms in shape segmentation and classification tasks on ShapeNet and COSEG datasets.
IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS
(2022)
Article
Computer Science, Artificial Intelligence
Abraham Nieva de la Hidalga, Paul L. Rosin, Xianfang Sun, Laurence Livermore, James Durrant, James Turner, Mathias Dillen, Alicia Musson, Sarah Phillips, Quentin Groom, Alex Hardisty
Summary: This paper presents a cross-validation approach to evaluate the applicability, adaptability, and portability of a semantic segmentation network in different types of natural history collections and institutions. The proposed method is tested on entomological microscope slides and herbarium sheets, and contributions include a software and ground truth sets for cross-validation.
MACHINE VISION AND APPLICATIONS
(2022)
Article
Computer Science, Artificial Intelligence
Guotao Wu, Ran Song, Mingxin Zhang, Xiaolei Li, Paul L. Rosin
Summary: This paper presents a lightweight CNN, called LiTM-Net, which generates high-quality HDR images by recovering information in saturated regions. Compared to existing methods, LiTM-Net achieves faster performance on mobile devices without sacrificing reconstruction quality.
PATTERN RECOGNITION
(2022)
Article
Environmental Sciences
Wanli Ma, Oktay Karaku, Paul L. Rosin
Summary: Land cover mapping is a widely used technique in remote sensing computational imaging that provides spatial information on various classes of physical properties on the Earth's surface. It plays a crucial role in developing solutions to environmental problems and faces challenges in integrating complementary information from multi-modal remote sensing imagery.
Article
Computer Science, Information Systems
Shu-Yu Chen, Yu-Kun Lai, Shihong Xia, Paul L. Rosin, Lin Gao
Summary: With the rapid development of VR technology, the need for bidirectional communication in immersive VR has arisen. This paper introduces a real-time system that can capture and reconstruct 3D faces wearing HMDs, and recover eye gaze effectively. It also proposes a novel method to map eye gaze directions to the 3D virtual world, offering a new interactive mode in VR. The effectiveness of the system is demonstrated through comparison with state-of-the-art techniques and live capture.
IEEE TRANSACTIONS ON MULTIMEDIA
(2023)
Article
Computer Science, Artificial Intelligence
Ran Song, Wei Zhang, Yitian Zhao, Yonghuai Liu, Paul L. Rosin
Summary: This paper proposes a framework that combines a Generative Adversarial Network and a Conditional Random Field to learn the visual saliency of 3D objects and scenes. The experimental results demonstrate that this method outperforms the existing approaches in predicting human fixations and addresses the research question about 3D visual saliency.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Review
Computer Science, Information Systems
Shu-Yu Chen, Jia-Qi Zhang, You-You Zhao, Paul L. Rosin, Yu-Kun Lai, Lin Gao
Summary: Image colorization is an important topic in computer graphics, aimed at adding color to monochromatic images. This survey presents the research history and popular algorithms in this field, highlighting recent developments in the combination of colorization with NLP and industrial applications. Various color control techniques, such as reference images and color-scribbles, are designed to improve color manipulation. The taxonomy of colorization methods based on input type (grayscale, sketch-based, and hybrid) is provided, with pros and cons discussed for each algorithm. The impact of deep learning, especially Generative Adversarial Networks (GANs), on this field is also discussed.
VISUAL INFORMATICS
(2022)
Article
Computer Science, Information Systems
Liang Fan, Xianfang Sun, Paul L. Rosin
Summary: In this paper, a novel triplet network with a spatial pyramid pooling layer and an attention model in the image space is proposed for face sketch recognition, achieving better performance on composite face photo-sketch datasets. The proposed solution reduces cross-modality differences between photo and sketch images, improving recognition accuracy by searching similar regions of the images. Particularly, the accuracy is higher than 81% for Set B in the UoM-SGFS dataset.