Article
Computer Science, Interdisciplinary Applications
Zhaoyang Song, Xiaoqiang Zhao, Yongyong Hui, Hongmei Jiang
Summary: This study proposes a progressive back-projection network (PBPN) for COVID-CT super-resolution, which achieves better performance in terms of resolution and quality by utilizing back-projection, deep feature extraction, and upscaling in two stages.
COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE
(2021)
Article
Environmental Sciences
Bo Huang, Boyong He, Liaoni Wu, Zhiming Guo
Summary: Super-resolution (SR) reconstruction of remote sensing images is a highly active area of research, and a deep convolutional neural network-based approach named DRDAN is proposed in this paper to achieve the fusion of global and local information for improved performance. The utilization of a residual dual-attention block (RDAB) and a dual-attention mechanism (DAM) in DRDAN allows for better adaptation to regions with high-frequency information, leading to superior results compared to other DCNN-based approaches.
Article
Geochemistry & Geophysics
Mengyang Shi, Yesheng Gao, Lin Chen, Xingzhao Liu
Summary: This study proposes a deep learning-based single-image super-resolution technology that combines neural networks with traditional algorithms, leveraging the properties of Gaussian blurring kernels to improve performance. A dual-resolution local attention unfolding network is designed to estimate the image, with a row-column decoupling attention module for feature fusion, achieving superior results compared to current state-of-the-art methods on remote sensing datasets.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Automation & Control Systems
Ying Shen, Weihuang Zheng, Liqiong Chen, Feng Huang
Summary: This work proposes an efficient and concise image super-resolution model called RSHAN, which solves the problems of redundant components and insufficient ability to extract high-frequency information in the Transformer model. By fusing the local features extracted by the CNN branch and the long-range dependencies extracted by Transformers, the performance of RSHAN is improved. Extensive experiments show that the proposed method outperforms existing methods in terms of peak signal-to-noise ratio (PSNR) by up to 0.11 dB, while reducing computational complexity and inference time by 5% and 10%, respectively.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Computer Science, Information Systems
Axi Niu, Pei Wang, Yu Zhu, Jinqiu Sun, Qingsen Yan, Yanning Zhang
Summary: This paper proposes a novel Ghost Residual Attention Network (GRAN) for efficient super-resolution, which reduces computational resources by introducing Ghost Residual Attention Block (GRAB) groups. GRAB consists of the Ghost Module and Channel and Spatial Attention Module (CSAM) to overcome the drawbacks of standard convolutional operation. Experimental results demonstrate the superior performance of GRAN compared to baseline models, achieving higher performance with over ten times reduction in computational resources.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Chemistry, Multidisciplinary
Beibei Wang, Binyu Yan, Gwanggil Jeon, Xiaomin Yang, Changjun Liu, Zhuoyue Zhang
Summary: Due to hardware constraints, obtaining medical images with sufficient resolution to diagnose small lesions is difficult. This study introduces super-resolution (SR) to enhance and restore medical image details for more accurate diagnoses. A lightweight dual mutual-feedback network (DMFN) is proposed, which utilizes a feedback mechanism and a contrast-enhanced residual block (CRB) to improve image resolution. The results show that DMFN performs well without excessive computational resources, making it suitable for resource-constrained medical devices.
APPLIED SCIENCES-BASEL
(2022)
Article
Computer Science, Information Systems
Defu Qiu, Yuhu Cheng, Xuesong Wang
Summary: Atrial fibrillation (AF) is a growing medical burden globally, characterized by atrial tissue remodeling and low-pressure atrial tissue fibrosis. To address the limitations in high-resolution cardiac magnetic resonance imaging (CMRI) data acquisition, we propose a Progressive Feedback Residual Attention Network (PFRN) for CMRI super-resolution. Through feature extraction and progressive feedback modules, PFRN improves the image's detailed information and visual effect in CMRI reconstruction.
IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS
(2023)
Article
Computer Science, Artificial Intelligence
Xuanyi Li, Zhuhong Shao, Bicao Li, Yuanyuan Shang, Jiasong Wu, Yuping Duan
Summary: This paper introduces a lightweight residual shuffle attention network for image super-resolution task. It designs a residual shuffle attention block (RSAB) to extract deep features, composed of multiple enhanced residual blocks (MERB) and shuffle attention. The MERB boosts the feature representation, and the shuffle attention captures critical information extracted by grouping features. Experimental results demonstrate that the proposed network outperforms state-of-the-art methods with acceptable complexity on several benchmarks.
MACHINE VISION AND APPLICATIONS
(2023)
Article
Computer Science, Information Systems
Yichun Jiang, Yunqing Liu, Weida Zhan, Depeng Zhu
Summary: This article introduces a lightweight dual-residual network (LDRN) for single image super-resolution, proposing a new residual unit and dual-stream residual block design, as well as a new up-sampling module. Experimental results show that the network has better reconstruction performance and lightweight performance, making it suitable for real-time applications on mobile devices.
Article
Computer Science, Software Engineering
Yepeng Liu, Dezhi Yang, Fan Zhang, Qingsong Xie, Caiming Zhang
Summary: This paper proposes a deep recursive residual channel attention network (DRRCAN) model to address the problems in image super-resolution of convolutional neural network models. A channel feature fusion module is constructed to effectively fuse the feature information of different layers. Recursive blocks are adopted to reduce the number of parameters in the deep network. Residual modules and long skip connections are introduced to improve the stability and generalization ability of the model. Extensive benchmark evaluations validate the superiority of the proposed DRRCAN model compared with existing algorithms.
Article
Computer Science, Interdisciplinary Applications
Dongmei Zhu, Hongxu He, Dongbo Wang
Summary: In this study, a Feedback Attention Network (FBAN) is proposed for super-resolution reconstruction of cardiac magnetic resonance imaging (CMRI). The FBAN method improves the edge and texture information of the reconstructed image and enhances the peak signal-to-noise ratio (PSNR) and the structural similarity index (SSIM) objective evaluation indicators.
COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE
(2023)
Article
Environmental Sciences
Runrui Liu, Fei Tao, Xintao Liu, Jiaming Na, Hongjun Leng, Junjie Wu, Tong Zhou
Summary: In this paper, an improved deep learning model RAANet is proposed, which constructs a new residual ASPP by embedding attention module and residual structure into ASPP for multi-scale semantic information and improved classification accuracy of land use in remote sensing images.
Article
Computer Science, Artificial Intelligence
Saeed Anwar, Nick Barnes
Summary: The research proposes a compact and accurate super-resolution algorithm DRLN, which achieves deep supervision learning through cascading residual structures and densely concatenated residual block settings, and models inter-level and intra-level dependencies between crucial features using Laplacian attention. Comprehensive evaluations on various test datasets show that the DRLN algorithm performs significantly better in terms of visual quality and accuracy compared to other methods.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2022)
Article
Engineering, Electrical & Electronic
Huan Zhang, Yihao Cao, Jianghui Cai, Xingjuan Cai, Wensheng Zhang
Summary: This paper proposes a novel Video Super-Resolution (VSR) reconstruction technique that utilizes inter-frame and intra-frame correlations to improve the spatiotemporal resolution of low-resolution (LR) video sequences, particularly in low-light conditions. The proposed method generates hidden information for the current frame using a combination of front and rear frames, and uses a re-fusion block (RFB) to re-fuse this hidden information with the corresponding LR frame. The improved dual attention mechanism (DAM) is integrated into the network to extract more accurate intra-frame features without increasing the number of parameters. Experimental results demonstrate that the proposed model outperforms state-of-the-art methods in terms of performance.
SIGNAL PROCESSING-IMAGE COMMUNICATION
(2023)
Article
Computer Science, Interdisciplinary Applications
Dongmei Zhu, Degang Sun, Dongbo Wang
Summary: In this study, a dual attention mechanism network for single image super-resolution (SISR) is proposed. The experimental results show that DAMN can better restore the image contour features, achieve higher PSNR, SSIM, and have better visual effects.
COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE
(2022)
Article
Computer Science, Artificial Intelligence
Zheng Hui, Xinbo Gao, Xiumei Wang
Article
Computer Science, Artificial Intelligence
Yu Zheng, Qiuyu Chen, Jianping Fan, Xinbo Gao
Article
Computer Science, Information Systems
Zheng Hui, Jie Li, Xinbo Gao, Xiumei Wang
Summary: Deep neural networks have proven to significantly enhance single image super-resolution performance. While previous methods aiming at PSNR maximization often result in blurred images at high upscaling factors, the introduction of GANs has helped to mitigate this issue by generating impressive results with synthetic high-frequency textures. However, GAN-based approaches may introduce fake textures and artifacts to enhance the visual resolution of super-resolved images. This paper proposes a new perceptual image super-resolution method that progressively generates high-quality results by constructing a stage-wise network.
INFORMATION SCIENCES
(2021)
Article
Computer Science, Artificial Intelligence
Xiumei Wang, Dingning Guo, Peitao Cheng
Summary: Sequential data clustering is a challenging task in data mining, and subspace clustering is a representative tool for dealing with complex local correlation and high-dimensional structure. It is important to learn a more specific structure representation of a sequence to preserve both sequential information and efficient connections.
PATTERN RECOGNITION
(2022)
Article
Environmental Sciences
Chi Zhang, Mingjin Zhang, Yunsong Li, Xinbo Gao, Shi Qiu
Summary: The paper introduces a difference curvature multidimensional network for hyperspectral image super-resolution that leverages spectral correlation to enhance spatial resolution. This is achieved through a self-attention mechanism and bottleneck projection to reduce redundancy. Additionally, a difference curvature branch is designed as an edge indicator to preserve texture information and eliminate unwanted noise in high-dimensional space.
Article
Environmental Sciences
Jie Guo, Chengyu He, Mingjin Zhang, Yunsong Li, Xinbo Gao, Bangyu Song
Summary: Synthetic aperture radar (SAR) remote sensing is important in modern Earth observation, but interpreting SAR images is challenging. This paper proposes an edge-preserving convolutional generative adversarial network (EPCGAN) to enhance SAR-to-optical image translation by leveraging edge information and implementing content-adaptive convolution. Experiments show that EPCGAN outperforms other methods in recovering structures and yielding superior evaluation results.
Article
Engineering, Electrical & Electronic
Xi Zhang, Xiumei Wang, Peitao Cheng
Summary: This article proposes a new strategy based on contrastive learning to improve the performance of unsupervised image retrieval. By fully utilizing the structural information in semantic similarity and employing a novel framework to handle hash codes with different lengths simultaneously, better image retrieval results are achieved.
IEEE SIGNAL PROCESSING LETTERS
(2022)
Article
Computer Science, Artificial Intelligence
Xiumei Wang, Tianmeng Li, Zheng Hui, Peitao Cheng
Summary: This paper proposes a deep learning based stereo image super-resolution algorithm that utilizes additional information in stereo image pairs to enhance the quality of reconstructed images. The authors introduce an adaptive modulation alignment mechanism and the use of rectangular convolution kernel to address the challenges caused by occlusion. Experimental results demonstrate that the proposed method achieves state-of-the-art performance on multiple stereo benchmarks.
PATTERN RECOGNITION LETTERS
(2022)
Article
Environmental Sciences
Weiming Chen, Bing Han, Zheng Yang, Xinbo Gao
Summary: Ships are crucial in ocean transportation, making ship detection an essential technology for marine safety. While optical remote-sensing images are valuable for ship detection, few open datasets exist due to sensitive data issues. The proposed MSSDet framework, utilizing a joint recursive feature pyramid, shows promising results in detecting multi-scale ship objects with improved generalizability and competitive performance compared to state-of-the-art methods.
Article
Environmental Sciences
Zheng Yang, Bing Han, Weiming Chen, Xinbo Gao
Summary: Unmanned aerial vehicles (UAVs) have attracted increasing attention in recent years due to their wide range of applications. Object tracking is a critical algorithm for UAVs, but it still faces challenges such as limited textures and contours of UAV objects, and the need to constantly move the camera with the object. In this paper, we propose an end-to-end discriminative tracker called TMDiMP, which incorporates a novel memory-aware attention mechanism to generate discriminative features and overcome the object-forgetting problem. We also introduce a UAV object-tracking dataset named VIPUOTB, which differs from existing datasets in terms of object size, camera motion, and location distribution. Experimental results demonstrate the effectiveness and robustness of our proposed algorithm.
Proceedings Paper
Computer Science, Artificial Intelligence
Yanan Gu, Xu Yang, Kun Wei, Cheng Deng
Summary: The study proposed a novel and effective framework for online class-incremental continual learning, which not only considers sample selection but also fully explores semantic information in the data stream. By effectively combining gradients and mutual information, state-of-the-art performance was achieved.
2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR)
(2022)
Article
Computer Science, Artificial Intelligence
Jingyuan Yang, Jie Li, Leida Li, Xiumei Wang, Yuxuan Ding, Xinbo Gao
Summary: This study addresses the issue of subjectivity in visual emotion analysis and proposes a novel method to tackle this problem. By simulating the emotion evocation process and incorporating an attention mechanism, the proposed method is able to better predict people's emotions towards different visual stimuli.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Zheng Hui, Jie Li, Xiumei Wang, Xinbo Gao
Summary: This study introduces an adaptive modulation network (AMNet) for blind super-resolution (SR) with multiple degradations and incorporates deep reinforcement learning into the entire blind SR model to address non-differentiable issues.
2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Andrey Ignatov, Andres Romero, Heewon Kim, Radu Timofte, Chiu Man Ho, Zibo Meng, Kyoung Mu Lee, Yuxiang Chen, Yutong Wang, Zeyu Long, Chenhao Wang, Yifei Chen, Boshen Xu, Shuhang Gu, Lixin Duan, Wen Li, Wang Bofei, Zhang Diankai, Zheng Chengjian, Liu Shaoli, Gao Si, Zhang Xiaofeng, Lu Kaidi, Xu Tianyu, Zheng Hui, Xinbo Gao, Xiumei Wang, Jiaming Guo, Xueyi Zhou, Hao Jia, Youliang Yan
Summary: Video super-resolution has become increasingly important on mobile devices, but existing solutions are too resource-intensive, leading to the introduction of the first Mobile AI challenge. Participants used the REDS dataset and evaluated their models on the OPPO Find X2 smartphone to develop video super-resolution solutions that achieve real-time performance on any mobile GPU.
2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS, CVPRW 2021
(2021)
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.