Article
Computer Science, Software Engineering
Sean Flynn, David Hart, Bryan Morse, Seth Holladay, Parris Egbert
Summary: This paper introduces a novel method for intelligently resizing a wide range of volumetric data, including fluids, allowing for more versatile post-processing. Additionally, a faster seam computation method is presented to improve production workflow viability.
ACM TRANSACTIONS ON GRAPHICS
(2021)
Article
Biochemical Research Methods
Felix Kallenborn, Andreas Hildebrandt, Bertil Schmidt
Summary: This study introduces a scalable error correction algorithm CARE for Illumina data using minhashing, allowing for efficient similarity search within large sequencing read collections. The algorithm shows significantly reduced false-positive error corrections while maintaining competitive true-positive numbers in performance evaluation.
Article
Computer Science, Artificial Intelligence
Dongjing Wang, Xin Zhang, Dongjin Yu, Guandong Xu, Shuiguang Deng
Summary: The article discusses a personalized music recommender system that incorporates rich content and context data in a unified and adaptive way. By proposing a method called content- and context-aware music embedding (CAME) and integrating deep learning techniques, the system is able to effectively capture the features of music pieces.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2021)
Article
Computer Science, Artificial Intelligence
Meng Lan, Jing Zhang, Zengmao Wang
Summary: In this paper, a novel model called Coherence-aware Context Aggregator (CCA) is proposed for semi-supervised video object segmentation (VOS). CCA evaluates the coherence of the predicted result and updates the temporal context to improve segmentation performance. Experimental results demonstrate that CCA achieves a better trade-off between efficiency and accuracy compared to state-of-the-art methods.
PATTERN RECOGNITION
(2023)
Article
Computer Science, Information Systems
Yiji Zhao, Youfang Lin, Zhihao Wu, Yang Wang, Haomin Wen
Summary: Dynamic networks are widely used in various scientific fields, but existing network distance measures are designed for static networks and ignore valuable context structure information. To address this issue, a context-aware distance paradigm is proposed, and a context-aware spectral distance is given as an instance. In node-aligned dynamic networks, the context helps the context-aware spectral distance outperform traditional spectral distance.
ACM TRANSACTIONS ON THE WEB
(2022)
Article
Chemistry, Analytical
Chuanzhen Hu, Xianli Wang, Ling Liu, Chuanhai Fu, Kaiqin Chu, Zachary J. Smith
Summary: Compressive imaging strategy combined with context-aware image prior has significantly improved the speed and accuracy of Raman hyperspectral imaging. CARCI reduces the number of measurements by up to 85% while maintaining a high image quality, facilitating faster data acquisition and more reliable downstream analysis. Large datasets of chemical images can be obtained in a reasonable timescale, leading to improved biochemical modeling and identification of rare cells.
Article
Mathematics, Applied
Yoshito Hirata, Yuki Kitanishi, Hiroki Sugishita, Yukiko Gotoh
Summary: An algorithm is proposed to refine the reconstruction of an original time series using a recurrence plot, calculating local distances based on Jaccard coefficients with neighbors and taking their weighted average. The utility of the method is demonstrated through two examples.
Article
Chemistry, Analytical
Ilias Papastratis, Kosmas Dimitropoulos, Petros Daras
Summary: This study introduces a novel approach for context-aware continuous sign language recognition using a generative adversarial network architecture, and examines the significance of contextual information in improving recognition accuracy.
Article
Geochemistry & Geophysics
Weiwei Cai, Pengjiang Qian, Yao Ding, Meiqiao Bi, Xin Ning, Danfeng Hong, Xiao Bai
Summary: This article proposes a novel approach to address the classification of hyperspectral images by introducing multiorder statistical representation-guided graph convolution and continuous context threshold-aware network, which overcomes the limitations of traditional CNNs in modeling spatial dependencies in HSIs and feature redundancy. Extensive experiments demonstrate the competitive performance of the proposed method.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Computer Science, Information Systems
Ameni Chtourou, Pierre Merdrignac, Oyunchimeg Shagdar
Summary: Collective perception messages (CPMs) provide information about detected objects and their status to neighbors, but their performance can degrade if the wireless channel is congested or vehicles need to process excessive data. This study proposes context-aware communication schemes for controlling CPM content selection and transmission, considering aspects such as radio resource use and infrastructure availability. Simulation results demonstrate that the scheme taking into account both resource use and infrastructure availability shows the best performance in terms of CBR, packet delivery ratio, and awareness ratio.
Article
Engineering, Electrical & Electronic
Maram Bani Younes, Azzedine Boukerche
Summary: Digital maps installed in vehicles help determine relative locations and aid in decision-making. However, they are vulnerable to destruction or inaccuracy and require regular updates. To address this, a prediction protocol that analyzes traffic characteristics can be used to predict road conditions.
IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY
(2023)
Article
Computer Science, Information Systems
Gad Gad, Eyad Gad, Korhan Cengiz, Zubair Fadlullah, Bassem Mokhtar
Summary: This article introduces a framework that integrates machine learning with the Internet of Things for video captioning. The framework includes steps such as mining video-caption pairs, preprocessing the data, and proposing two deep learning models. The models are evaluated on different platforms in terms of accuracy and inference time.
Article
Computer Science, Artificial Intelligence
Jianli Zhao, Shidong Zheng, Huan Huo, Maoguo Gong, Tianheng Zhang, Lijun Qu
Summary: This study proposes a fast tensor factorization model named FWCP for context-aware recommendation. The model addresses the challenges of data sparsity and computational efficiency through unifying explicit and implicit feedback information and implementing parallel optimization. Experimental results demonstrate the superiority of FWCP over other context-aware recommendation algorithms.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Computer Science, Information Systems
Lianwei Wu, Pusheng Liu, Yuheng Yuan, Siying Liu, Yanning Zhang
Summary: Neural text transfer aims to change the style of a text sequence while preserving its original content. However, existing unsupervised learning approaches suffer from inconsistency between the transferred style and content, as well as insufficient preservation of the core semantics. To address these issues, we propose a Context-aware Style Learning and Content Recovery (CSLCR) network for neural text transfer.
INFORMATION PROCESSING & MANAGEMENT
(2023)
Proceedings Paper
Computer Science, Hardware & Architecture
Bohyung Choi, Minyoung Lee, Seung-Won Jung, Yucheng Lu
Summary: We propose a solution for correcting distortion and scattered information in panorama images, by using a weighted map in combination with original dynamic programming energy. Our method can resize images into a standard 16:9 ratio for widescreen displays without compromising aesthetic quality, and is applicable in situations with optical distortion and dispersed visual information.
2021 INTERNATIONAL CONFERENCE ON ELECTRONICS, INFORMATION, AND COMMUNICATION (ICEIC)
(2021)
Article
Computer Science, Artificial Intelligence
Meng Jian, Jingjing Guo, Ge Shi, Lifang Wu, Zhangquan Wang
Summary: This study presents a multimodal collaborative graph model for image recommendation, which uncovers users' interests by considering both visual and collaborative signals. Experimental results demonstrate the effectiveness of the model in mining users' interests and image recommendation.
APPLIED INTELLIGENCE
(2023)
Article
Computer Science, Software Engineering
Yu Song, Fan Tang, Weiming Dong, Feiyue Huang, Tong-Yee Lee, Changsheng Xu
Summary: This article focuses on the influence of visual effects on individual preferences in grid collages of small image collections. A novel balance-aware metric is proposed to evaluate the visual balance in line with human subjective perception. The metric integrates psychological achievements and incorporates a bonus mechanism for user preference. Experiments show that the metric can accurately evaluate grid collages and generate visually pleasant results comparable to manual designs.
IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS
(2023)
Article
Computer Science, Artificial Intelligence
Xiaoyu Kong, Yingying Deng, Fan Tang, Weiming Dong, Chongyang Ma, Yongyong Chen, Zhenyu He, Changsheng Xu
Summary: This article conducts a detailed analysis of the flickering effects in video stylization and proposes a new multichannel correlation network (MCCNet) to address this issue. MCCNet aligns each output frame with the input frame in the hidden feature space while preserving the desired style patterns.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Software Engineering
Cong Wang, Fan Tang, Yong Zhang, Tieru Wu, Weiming Dong
Summary: This paper addresses the inconsistency issue of different regions in facial image synthesis and proposes a method for harmonized regional style transfer using a multi-scale encoder and a multi-region style attention module. By extracting style information from multiple regions of a reference image, a harmonious result is generated. The paper also introduces style mapping networks for multi-modal style synthesis and utilizes an invertible flow model to fine-tune the style code. Experimental results demonstrate that the proposed model can reliably perform style transfer and multi-modal manipulation, producing outputs comparable to state-of-the-art methods.
COMPUTATIONAL VISUAL MEDIA
(2023)
Article
Computer Science, Software Engineering
Yuxin Zhang, Fan Tang, Weiming Dong, Haibin Huang, Chongyang Ma, Tong-Yee Lee, Changsheng Xu
Summary: This work presents a novel framework, UCAST, for unified and arbitrary style transfer. By learning style representation and conducting contrastive learning, it addresses the issue of insufficient style information utilization in existing methods. The results obtained through contrastive learning and adaptive temperature scheme outperform state-of-the-art methods.
ACM TRANSACTIONS ON GRAPHICS
(2023)
Article
Engineering, Electrical & Electronic
Lifang Wu, Xianglong Lang, Ye Xiang, Changwen Chen, Zun Li, Zhuming Wang
Summary: This paper proposes a novel end-to-end hierarchical relation inference framework based on active spatial positions for group activity recognition. The framework is able to locate active spatial positions and use them as visual tokens to infer the relations for token embeddings. Experimental results show that the proposed framework only requires individual bounding box labeling in the training stage and automatically eliminates background noise from the entire scene.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2023)
Article
Computer Science, Information Systems
Jinyang Huang, Bin Liu, Chenglin Miao, Yan Lu, Qijia Zheng, Yu Wu, Jiancun Liu, Lu Su, Chang Wen Chen
Summary: Driven by various applications, recent achievements have been made in WiFi-based Human Activity Recognition (HAR) techniques that utilize WiFi infrastructures to infer human activities. However, the ubiquitous Co-channel Interference (CCI) poses a challenge to the performance of these systems. In this paper, a novel approach called PhaseAnti is proposed to exploit the CCI-independent phase component of WiFi Channel State Information (CSI) for HAR, which achieves superior effectiveness and recognition speed compared with existing solutions.
IEEE TRANSACTIONS ON MOBILE COMPUTING
(2023)
Article
Computer Science, Artificial Intelligence
Tiancheng Lin, Zhimiao Yu, Zengchao Xu, Hongyu Hu, Yi Xu, Chang-Wen Chen
Summary: Self-supervised representation learning has limitations when applied to whole-slide pathological images due to their unique characteristics. To address this issue, we propose a novel scheme called Spatial Guided Contrastive Learning (SGCL), which leverages spatial proximity and multi-object priors for stable self-supervision. SGCL outperforms state-of-the-art methods on diverse downstream tasks across multiple datasets.
MEDICAL IMAGE ANALYSIS
(2023)
Article
Computer Science, Artificial Intelligence
Chengcheng Ma, Xingjia Pan, Qixiang Ye, Fan Tang, Weiming Dong, Changsheng Xu
Summary: Semi-supervised object detection has made significant progress, but the performance of self-labeling-based methods is limited due to the misguidance of incorrect pseudo labels predicted by the detector itself. This paper proposes a detection framework called CrossRectify, which obtains accurate pseudo labels by simultaneously training two detectors with different initial parameters. The proposed approach leverages the disagreements between detectors to discern self-errors and improves the quality of pseudo labels. Extensive experiments demonstrate that CrossRectify achieves superior performance on 2D and 3D detection benchmarks.
PATTERN RECOGNITION
(2023)
Article
Engineering, Electrical & Electronic
Chengcheng Ma, Yang Liu, Jiankang Deng, Lingxi Xie, Weiming Dong, Changsheng Xu
Summary: Pretrained vision-language models like CLIP have strong generalization capability. The Context Optimization (CoOp) method has been proposed to improve performance by learning continuous prompts. However, CoOp suffers from overfitting. This study analyzes the gradient flow and proposes Subspace Prompt Tuning (Sub PT) to eliminate overfitting. Additionally, a Novel Feature Learner (NFL) is introduced to enhance CoOp's generalization ability.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2023)
Article
Computer Science, Artificial Intelligence
Zhiyong Huang, Kekai Sheng, Ke Li, Jian Liang, Taiping Yao, Weiming Dong, Dengwen Zhou, Xing Sun
Summary: Batch normalization is widely used in deep neural networks, but it is ineffective for cross-domain tasks. This paper proposes a novel normalization method called Reciprocal Normalization (RN), which utilizes cross-domain relation to improve adaptability. Compared to batch normalization, RN is more suitable for unsupervised domain adaptation and can be easily integrated into popular domain adaptation methods.
PATTERN RECOGNITION
(2023)
Article
Art
Yuxin Zhang, Fan Tang, Weiming Dong, Thi-Ngoc-Hanh Le, Changsheng Xu, Tong-Yee Lee
Summary: The authors propose a deep neural network-based algorithm to automatically generate Portrait Map Art (PMA), a modern art form created by British portrait artist Ed Fairburn. They formulate the generation of PMA as an adaptive dual-to-single image translation problem and utilize encoder networks to analyze the appearance of the portrait and map images. The proposed model can produce new PMA images without baselines by optimizing with a cycle-consistency constraint.
Article
Computer Science, Information Systems
Pei Lv, Jianqi Fan, Xixi Nie, Weiming Dong, Xiaoheng Jiang, Bing Zhou, Mingliang Xu, Changsheng Xu
Summary: Personalized image aesthetic assessment is a hot topic with wide applications in various fields. We propose a novel user-guided framework based on deep reinforcement learning to acquire precise personalized aesthetic distribution by small amount of samples. Our approach achieves state-of-the-art results in personalized image aesthetic assessment tasks.
IEEE TRANSACTIONS ON MULTIMEDIA
(2023)
Article
Computer Science, Artificial Intelligence
Dong Chen, Xingjia Pan, Fan Tang, Weiming Dong, Changsheng Xu
Summary: In this paper, a new weakly supervised object localization framework called SPA(2)Net is proposed to accurately locate objects by preserving the structure and utilizing spatial attention in deep CNN. Unlike traditional methods, SPA(2)Net separates the localization task from the classification task by introducing a localization branch and optimizing it with a self-supervised structural-preserved localization mask. Experimental results show that SPA(2)Net achieves substantial and consistent performance gains compared to baseline approaches on two benchmark datasets.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Meishan Liu, Meng Jian, Ge Shi, Ye Xiang, Lifang Wu
Summary: In this paper, we propose a graph contrastive learning on complementary embedding (GCCE) approach to address the challenges of interaction sparsity and bias by introducing negative interests and designing perturbed graph convolution and complementary embedding generation for interest modeling. We validate the effectiveness of GCCE through experiments on real datasets and show its superiority over state-of-the-art recommendation models.
PROCEEDINGS OF THE 2023 ACM INTERNATIONAL CONFERENCE ON MULTIMEDIA RETRIEVAL, ICMR 2023
(2023)
Article
Computer Science, Artificial Intelligence
Rui Lv, Dingheng Wang, Jiangbin Zheng, Zhao-Xu Yang
Summary: In this paper, the authors investigate tensor decomposition for neural network compression. They analyze the convergence and precision of tensor mapping theory, validate the rationality of tensor mapping and its superiority over traditional tensor approximation based on the Lottery Ticket Hypothesis. They propose an efficient method called 3D-KCPNet to compress 3D convolutional neural networks using the Kronecker canonical polyadic (KCP) tensor decomposition. Experimental results show that 3D-KCPNet achieves higher accuracy compared to the original baseline model and the corresponding tensor approximation model.
Article
Computer Science, Artificial Intelligence
Xiangkun He, Zhongxu Hu, Haohan Yang, Chen Lv
Summary: In this paper, a novel constrained multi-objective reinforcement learning algorithm is proposed for personalized end-to-end robotic control with continuous actions. The approach trains a single model using constraint design and a comprehensive index to achieve optimal policies based on user-specified preferences.
Article
Computer Science, Artificial Intelligence
Zhijian Zhuo, Bilian Chen, Shenbao Yu, Langcai Cao
Summary: In this paper, a novel method called Expansion with Contraction Method for Overlapping Community Detection (ECOCD) is proposed, which utilizes non-negative matrix factorization to obtain disjoint communities and applies expansion and contraction processes to adjust the degree of overlap. ECOCD is applicable to various networks with different properties and achieves high-quality overlapping community detection.
Article
Computer Science, Artificial Intelligence
Yizhe Zhu, Chunhui Zhang, Jialin Gao, Xin Sun, Zihan Rui, Xi Zhou
Summary: In this work, the authors propose a Contrastive Spatio-Temporal Distilling (CSTD) approach to improve the detection of high-compressed deepfake videos. The approach leverages spatial-frequency cues and temporal-contrastive alignment to fully exploit spatiotemporal inconsistency information.
Review
Computer Science, Artificial Intelligence
Laijin Meng, Xinghao Jiang, Tanfeng Sun
Summary: This paper provides a review of coverless steganographic algorithms, including the development process, known contributions, and general issues in image and video algorithms. It also discusses the security of coverless steganography from theoretical analysis to actual investigation for the first time.
Article
Computer Science, Artificial Intelligence
Yajie Bao, Tianwei Xing, Xun Chen
Summary: Visual question answering requires processing multi-modal information and effective reasoning. Neural-symbolic learning is a promising method, but current approaches lack uncertainty handling and can only provide a single answer. To address this, we propose a confidence based neural-symbolic approach that evaluates NN inferences and conducts reasoning based on confidence.
Article
Computer Science, Artificial Intelligence
Anh H. Vo, Bao T. Nguyen
Summary: Interior style classification is an interesting problem with potential applications in both commercial and academic domains. This project proposes a method named ISC-DeIT, which combines data-efficient image transformer architectures and knowledge distillation, to address the interior style classification problem. Experimental results demonstrate a significant improvement in predictive accuracy compared to other state-of-the-art methods.
Article
Computer Science, Artificial Intelligence
Shashank Kotyan, Danilo Vasconcellos Vargas
Summary: This article introduces a novel augmentation technique called Dynamic Scanning Augmentation to improve the accuracy and robustness of Vision Transformer (ViT). The technique leverages dynamic input sequences to adaptively focus on different patches, resulting in significant changes in ViT's attention mechanism. Experimental results demonstrate that Dynamic Scanning Augmentation outperforms ViT in terms of both robustness to adversarial attacks and accuracy against natural images.
Article
Computer Science, Artificial Intelligence
Hiba Alqasir, Damien Muselet, Christophe Ducottet
Summary: The article proposes a solution to improve the learning process of a classification network by providing shape priors, reducing the need for annotated data. The solution is tested on cross-domain digit classification tasks and a video surveillance application.
Article
Computer Science, Artificial Intelligence
Dexiu Ma, Mei Liu, Mingsheng Shang
Summary: This paper proposes a method using neural dynamics solvers to solve infinity-norm optimization problems. Two improved solvers are constructed and their effectiveness and superiority are demonstrated through theoretical analysis and simulation experiments.
Article
Computer Science, Artificial Intelligence
Francesco Gregoretti, Giovanni Pezzulo, Domenico Maisto
Summary: Active Inference is a computational framework that uses probabilistic inference and variational free energy minimization to describe perception, planning, and action. cpp-AIF is a header-only C++ library that provides a powerful tool for implementing Active Inference for Partially Observable Markov Decision Processes through multi-core computing. It is cross-platform and improves performance, memory management, and usability compared to existing software.
Article
Computer Science, Artificial Intelligence
Zelin Ying, Dawei Cheng, Cen Chen, Xiang Li, Peng Zhu, Yifeng Luo, Yuqi Liang
Summary: This paper proposes a novel stock market trends prediction framework called SMART, which includes a self-supervised stock technical data sequence embedding model S3E. By training with multiple self-supervised auxiliary tasks, the model encodes stock technical data sequences into embeddings and uses the learned sequence embeddings for predicting stock market trends. Extensive experiments on China A-Shares market and NASDAQ market prove the high effectiveness of our model in stock market trends prediction, and its effectiveness is further validated in real-world applications in a leading financial service provider in China.
Article
Computer Science, Artificial Intelligence
Hao Li, Hao Jiang, Dongsheng Ye, Qiang Wang, Liang Du, Yuanyuan Zeng, Liu Yuan, Yingxue Wang, C. Chen
Summary: DHGAT1, a dynamic hyperbolic graph attention network, utilizes hyperbolic metric properties to embed dynamic graphs. It employs a spatiotemporal self-attention mechanism and weighted node representations, resulting in excellent performance in link prediction tasks.
Article
Computer Science, Artificial Intelligence
Jiehui Huang, Zhenchao Tang, Xuedong He, Jun Zhou, Defeng Zhou, Calvin Yu-Chian Chen
Summary: This study proposes a progressive learning multi-scale feature blending model for image deraining tasks. The model utilizes detail dilation and texture extraction to improve the restoration of rainy images. Experimental results show that the model achieves near state-of-the-art performance in rain removal tasks and exhibits better rain removal realism.
Article
Computer Science, Artificial Intelligence
Lizhi Liu, Zilin Gao, Yinhe Wang, Yongfu Li
Summary: This paper proposes a novel discrete-time interconnected model for depicting complex dynamical networks. The model consists of nodes and edges subsystems, which consider the dynamic characteristic of both nodes and edges. By designing control strategies and coupling modes, the stabilization and synchronization of the network are achieved. Simulation results demonstrate the effectiveness of the proposed methods.