Article
Computer Science, Artificial Intelligence
Yuchen Pan, Yuanyuan Shang, Tie Liu, Zhuhong Shao, Guodong Guo, Hui Ding, Qiang Hu
Summary: This paper proposes a novel Spatial-Temporal Attention Depression Recognition Network (STA-DRN) that enhances feature extraction and relevance of depression recognition by capturing global and local spatial-temporal information. The experimental results demonstrate competitive performance and visualization analysis shows significant responses in specific locations related to depression.
EXPERT SYSTEMS WITH APPLICATIONS
(2024)
Article
Computer Science, Artificial Intelligence
Jianing Teng, Dong Zhang, Wei Zou, Ming Li, Dah-Jye Lee
Summary: Typical Facial Expression Network (TFEN) proposes a new network structure to address the challenges in facial expression recognition. It uses two 2D CNNs to extract facial and expression features, and a facial feature decoupler to minimize the influence of individual facial characteristics. Experimental results show that TFEN achieves better recognition accuracy than state-of-the-art approaches on four popular dynamic FER datasets.
IEEE TRANSACTIONS ON AFFECTIVE COMPUTING
(2023)
Article
Computer Science, Artificial Intelligence
Lifeng Zhang, Xiangwei Zheng, Xuanchi Chen, Xiuxiu Ren, Cun Ji
Summary: Facial expression recognition plays an important role in various applications, but not all extracted facial features are suitable. In this paper, a spatial-temporal fusion method with attention mechanism is proposed to improve accuracy. Experimental results show competitive performance compared to state-of-the-art methods.
NEURAL PROCESSING LETTERS
(2023)
Article
Chemistry, Multidisciplinary
Kai Hu, Yiwu Ding, Junlan Jin, Liguo Weng, Min Xia
Summary: This paper proposes a novel multi-scale time sampling module and a deep spatiotemporal feature extraction module to enhance the accuracy of human motion recognition network. Comparative experiments show that the proposed method achieves performance improvement on two datasets.
APPLIED SCIENCES-BASEL
(2022)
Article
Computer Science, Artificial Intelligence
Yuanlun Xie, Wenhong Tian, Hengxin Zhang, Tingsong Ma
Summary: Recent studies have shown that deep learning has great potential in facial expression recognition tasks and has attracted more and more attention from researchers. Existing methods have achieved good results in laboratory settings, but face challenges when applied in wild environments with more complex and diverse facial expression images. This paper proposes a new method for facial expression recognition by extracting and fusing multi-level features.
Article
Computer Science, Artificial Intelligence
Yi Ding, Neethu Robinson, Su Zhang, Qiuhao Zeng, Cuntai Guan
Summary: TSception is a multi-scale convolutional neural network that can classify emotions from EEG. It learns the temporal dynamics and spatial asymmetry of EEG and achieves higher classification accuracies and F1 scores.
IEEE TRANSACTIONS ON AFFECTIVE COMPUTING
(2023)
Article
Computer Science, Artificial Intelligence
Ryo Miyoshi, Noriko Nagata, Manabu Hashimoto
Summary: An algorithm that enhances ConvLSTM by adding skip connections and temporal gates was proposed for facial expression recognition, achieving superior performance compared to state-of-the-art methods. Experiments demonstrated the effectiveness of the proposed method on eNTERFACE05 database and CK+ dataset.
NEURAL COMPUTING & APPLICATIONS
(2021)
Article
Computer Science, Artificial Intelligence
Xiaoming Peng, Abdesselam Bouzerdoum, Son Lam Phung
Summary: In this paper, a part-based method is proposed for dynamic scene recognition, which aggregates local features from video frames. Experimental results demonstrate that the proposed method is highly competitive with state-of-the-art approaches.
NEURAL COMPUTING & APPLICATIONS
(2021)
Article
Computer Science, Artificial Intelligence
Jing Liu, Yang Liu, Wei Zhu, Xiaoguang Zhu, Liang Song
Summary: Transportation activity recognition (TAR) is crucial for intelligent transportation applications. This study proposes a novel parallel model, DSTRR, which combines automatic learning of statistical, spatial, and temporal features to achieve a robust representation.
PATTERN RECOGNITION
(2023)
Article
Engineering, Biomedical
Weijun Gong, Yurong Qian, Weihang Zhou, Hongyong Leng
Summary: The recognition of dynamic facial expressions is challenging due to various factors, and obtaining discriminative expression features has been difficult. Traditional deep learning networks lack understanding of global and temporal expressions. This study proposes an enhanced spatial-temporal learning network to improve dynamic facial expression recognition.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2024)
Article
Computer Science, Interdisciplinary Applications
Bingtao Zhang, Dan Wei, Guanghui Yan, Tao Lei, Haishu Cai, Zhifei Yang
Summary: In this study, a depression recognition framework based on the feature-level fusion of spatial-temporal electroencephalography (EEG) was proposed. By analyzing EEG data and performing feature fusion, depression could be effectively recognized. The experimental results showed that the method achieved a high accuracy rate.
COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE
(2022)
Article
Automation & Control Systems
Carmen Bisogni, Aniello Castiglione, Sanoar Hossain, Fabio Narducci, Saiyed Umer
Summary: This article proposes a facial expression recognition system that can provide quick assistance to the healthcare system and exceptional services to the patients. The system utilizes multi-resolution image processing techniques and has been proven superior through experiments.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2022)
Article
Chemistry, Analytical
Taejae Jeon, Han Byeol Bae, Yongju Lee, Sungjun Jang, Sangyoun Lee
Summary: With increasing interest in stress control, many studies on stress recognition have been conducted, focusing on physiological signals and facial images. However, both methods have their limitations. To address these challenges, a deep-learning-based stress-recognition method was proposed, utilizing a large image database and temporal and spatial attention mechanisms for improved performance.
Article
Computer Science, Information Systems
Ke Zhang, Hua-Nong Ting, Yao-Mun Choo
Summary: Deep learning theory has made remarkable advancements in the field of baby cry recognition. However, existing research faces challenges of small database size and neglect of multi-domain feature integration. To address these issues, a novel approach that combines transfer learning and feature fusion is proposed. Experimental results show that the proposed method effectively mitigates model overfitting due to small datasets and the fused features are better than existing single domain feature methods.
Article
Computer Science, Information Systems
Guichen Tang, Yue Xie, Ke Li, Ruiyu Liang, Li Zhao
Summary: This paper introduces a multimodal emotion recognition method that utilizes an attention mechanism to fuse audio and video features and model time series, effectively improving recognition accuracy.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Computer Science, Hardware & Architecture
Jixin Liu, Ruxue Zhang, Guang Han, Ning Sun, Sam Kwong
Summary: This paper proposes a method for video action recognition that protects visual privacy using compressed sensing, while balancing operational efficiency with recognition accuracy. The method utilizes a convolutional 3D network model and PCA to reduce temporal complexity, and integrates a sparse representation-based classification algorithm to improve recognition performance. Experiments show the method's robustness in video action recognition tasks and its ability to adequately protect visual privacy.
JOURNAL OF SYSTEMS ARCHITECTURE
(2021)
Article
Engineering, Electrical & Electronic
Guang Han, Yuechuan Ai, Jixin Liu, Ning Sun, Guangwei Gao
Summary: The DAPD-Net proposed in this study utilizes dual attention and part drop modules to enhance person reidentification, improve network performance, and increase resilience to occlusion.
JOURNAL OF ELECTRONIC IMAGING
(2021)
Article
Computer Science, Artificial Intelligence
Ning Sun, Ling Leng, Jixin Liu, Guang Han
Summary: A SlowFast graph convolution network (SF-GCN) is proposed for improved spatial-temporal feature extraction from skeleton sequence, utilizing the architecture of SlowFast network in the GCN model. SF-GCN consists of Fast and Slow pathways to extract features of fast and slow temporal changes, respectively, which are fused and weighted using lateral connection and channel attention. This design enhances feature extraction ability while reducing computational costs significantly.
IMAGE AND VISION COMPUTING
(2021)
Article
Radiology, Nuclear Medicine & Medical Imaging
Xiujuan Liu, Jun Mao, Ning Sun, Xiangrong Yu, Lei Chai, Ye Tian, Jianming Wang, Jianchao Liang, Haiquan Tao, Lihua Yuan, Jiaming Lu, Yang Wang, Bing Zhang, Kaihua Wu, Yiding Wang, Mengjiao Chen, Zhishun Wang, Ligong Lu
Summary: This study developed a new method using deep learning to automatically detect intracranial aneurysms from CTA images. The performance of the method was evaluated, showing reliable segmentation and detection of intracranial aneurysms, with a sensitivity of 100% for large and medium-sized aneurysms.
JOURNAL OF DIGITAL IMAGING
(2023)
Article
Computer Science, Hardware & Architecture
Ning Sun, Yao Song, Jixin Liu, Lei Chai, Haian Sun
Summary: In this paper, a model called the appearance and geometry transformer (AGT) is proposed to improve the accuracy of facial expression recognition (FER) in the wild. The AGT performs feature extraction and fusion on heterogeneous data using two transformer pathways. It achieves comparable results to state-of-the-art methods on benchmark databases FERplus and RAF-DB.
COMPUTERS & ELECTRICAL ENGINEERING
(2023)
Article
Computer Science, Hardware & Architecture
Jixin Liu, Pengcheng Dai, Guang Han, Ning Sun
Summary: Rapid technological advancements have led to an increase in the number of video surveillance devices in homes. This has prompted the development of various methods for video privacy protection. This paper proposes a method for evaluating the level of privacy protection in multilayer compressed sensing videos. By using a combination of CNN and RNN convolutional networks, the proposed approach achieves better prediction and generalization performance compared to previous methods. Additionally, an association model is established between visual privacy protection score and practicability score, allowing for practical applications and evaluation of other video privacy protection methods.
COMPUTERS & ELECTRICAL ENGINEERING
(2023)
Article
Computer Science, Artificial Intelligence
Ning Sun, Jianglong Tao, Jixin Liu, Haian Sun, Guang Han
Summary: In this article, a novel end-to-end trainable 3-D face feature reconstruction and learning network (3-DF-RLN) is proposed to improve the performance of facial expression recognition (FER) in the wild. Through 3-D face reconstruction, both the missing facial information and accurate facial geometric information can be effectively obtained. The proposed 3-DF-RLN model achieves FER by fusing apparent features from 2-D face images and geometric features from 3-D facial landmarks. Experimental results on benchmark databases demonstrate the superior FER performance of the proposed method, and the face graph from the geometry pathway reveals the correlations between facial landmarks in FER.
IEEE TRANSACTIONS ON COGNITIVE AND DEVELOPMENTAL SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Ning Sun, Qingyi Lu, Wenming Zheng, Jixin Liu, Guang Han
Summary: We propose an unsupervised cross-view facial expression adaptation network (UCFEAN) that can generate and recognize cross-view facial expressions in images in an unsupervised manner. UCFEAN converts the unsupervised domain adaptation between two image spaces into semi-supervised learning in feature spaces. It uses a generative adversarial network to perform cyclic image generation and project unlabelled target images and labelled source images to the corresponding feature spaces. The proposed method achieves realistic target image generation and high precision recognition of cross-view facial expressions.
IEEE TRANSACTIONS ON AFFECTIVE COMPUTING
(2023)
Article
Radiology, Nuclear Medicine & Medical Imaging
Xiujuan Liu, Jun Mao, Ning Sun, Xiangrong Yu, Lei Chai, Ye Tian, Jianming Wang, Jianchao Liang, Haiquan Tao, Zhishun Wang, Ligong Lu
Summary: This study aimed to evaluate the ability of the stereoscopic virtual reality display system (SVRDS) in displaying the angioarchitecture of cerebral arteriovenous malformations (CAVMs) by comparing its accuracy with that of the conventional computed tomography workstation (CCTW). Retrospective analysis of computed tomography angiography images was performed on 19 patients with confirmed CAVM, and the angioarchitectural parameters were recorded and compared between SVRDS and CCTW. SVRDS showed advantages in displaying the blood vessels of CAVMs compared to CCTW, and it provided a more intuitive visualization of the overall spatial structure.
JOURNAL OF DIGITAL IMAGING
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Xiang Li, Ning Sun, Jixin Liu, Lei Chai, Haian Sun
Summary: This paper proposes an end-to-end trainable network model MSR-Trans based on the global self-attention mechanism for multi-modal scene recognition. The model utilizes two transformer-based branches to extract features from RGB image and depth data, and then uses a fusion layer to fuse these features for final scene recognition. Lateral connections are added on some layers between the two branches to explore the relationship between multi-modal information, and a dropout layer is embedded in the transformer block to prevent overfitting. Extensive experiments on SUN RGB-D and NYUD2 datasets show that the proposed method achieves recognition accuracies of 69.0% and 74.1% for multi-modal scene recognition, respectively.
ADVANCES IN NATURAL COMPUTATION, FUZZY SYSTEMS AND KNOWLEDGE DISCOVERY, ICNC-FSKD 2022
(2023)
Article
Computer Science, Hardware & Architecture
Jixin Liu, Ru Meng, Ning Sun, Guang Han, Sam Kwong
Summary: This study proposes a computer vision fall detection method to protect video privacy. By utilizing compressed sensing visual privacy protection and GAN-based feature enhancement, the method can effectively detect fall behavior with high accuracy.
Article
Computer Science, Information Systems
Jixin Liu, Rong Tan, Guang Han, Ning Sun, Sam Kwong
Summary: The study proposes a fall detection system with visual shielding to ensure the safety of elderly people at home while preserving their personal privacy. Through multilayer compressed sensing and feature extraction, accurate identification of fall behaviors is achieved.
IEEE TRANSACTIONS ON MULTIMEDIA
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Ji-xin Liu, Rong Tan, Ning Sun, Guang Han, Xiao-fei Li
2020 3RD INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE AND BIG DATA (ICAIBD 2020)
(2020)
Article
Engineering, Electrical & Electronic
Jixin Liu, Zheng Tang, Ning Sun, Guang Han, Sam Kwong
SIGNAL PROCESSING-IMAGE COMMUNICATION
(2020)
Article
Computer Science, Information Systems
Guang Han, Yan Gao, Ning Sun