Article
Computer Science, Information Systems
Marcel Grimmer, Raghavendra Ramachandra, Christoph Busch
Summary: Face Age Progression (FAP) involves synthesizing face images to simulate aging effects and predict future appearances, with recent advancements in deep generative networks improving visual fidelity and aging accuracy. However, a systematic comparison of different methods is needed to accelerate research and address open challenges in the field.
Article
Computer Science, Artificial Intelligence
Hongyu Yang, Di Huang, Yunhong Wang, Anil K. Jain
Summary: The study introduces a novel generative adversarial network approach to achieve facial age progression while ensuring aging accuracy and maintaining individual identity stability. The method employs an adversarial learning scheme to train a single generator and multiple parallel discriminators simultaneously, resulting in smooth continuous face aging sequences.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2021)
Article
Engineering, Electrical & Electronic
Lingshuang Du, Haifeng Hu
Summary: This paper introduces a novel cross-age face verification framework called the CIDA model, which effectively addresses the challenges of age variations in face recognition through two cascading networks.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2021)
Article
Engineering, Electrical & Electronic
Yongbo Wu, Lingshuang Du, Haifeng Hu
Summary: The paper proposes the Parallel Multi-path Age Distinguish Network (PMADN) model to address the challenging task of Cross Age Face Recognition (CAFR) in the field of face recognition. The model consists of two cascading networks that better extract identity features and age features, avoiding the limitations of simple linear combination of identity factor and age factor in mainstream CAFR methods. Extensive experiments on benchmark databases demonstrate the effectiveness and superiority of the proposed method.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2021)
Article
Computer Science, Artificial Intelligence
Ali Akbari, Muhammad Awais, Zhenhua Feng, Ammarah Farooq, Josef Kittler
Summary: This article focuses on subjective-exclusive cross-database age estimation, which is rarely valid in practical applications. The age estimation problem is formulated as the distribution learning framework, and a new loss function is proposed for improved performance. Experimental results show that the proposed approach outperforms state-of-the-art methods.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2022)
Article
Automation & Control Systems
Rawan Sulaiman Howyan, Emad Sami Jaha
Summary: Facial age estimation is a widely used biometric based on computer-vision, but the semantic gap between machines and humans makes it challenging to extract and process semantic features inferred by human-vision. Researchers aim to bridge this gap by integrating human-vision with traditional computer-vision features to enhance the performance of age estimation models.
INTELLIGENT AUTOMATION AND SOFT COMPUTING
(2022)
Article
Computer Science, Information Systems
Neha Sharma, Reecha Sharma, Neeru Jindal
Summary: Research on face age progression aims to accurately generate aged faces using AttentionGAN and SRGAN, addressing the challenge of unnatural modifications in existing approaches. AttentionGAN utilizes attention and content masks to produce desired results, while SRGAN generates high-resolution aged images for detailed information in images.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Article
Computer Science, Information Systems
Beichen Zhang, Yue Bao
Summary: This paper proposes a learning method called the cross-dataset training convolutional neural network (CDCNN) for age estimation from a single human face image. By simultaneously training on multiple datasets and using additional labeled data, this method achieves impressive performance in age estimation.
Article
Computer Science, Artificial Intelligence
Haoyi Wang, Victor Sanchez, Chang-Tsun Li
Summary: In this paper, a face-based age estimation framework called ADPF is proposed, which consists of two separate CNNs, AttentionNet and FusionNet. AttentionNet dynamically locates and ranks age-specific patches, while FusionNet uses these patches and facial images to predict age. Experimental results demonstrate that the proposed framework outperforms state-of-the-art methods.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2022)
Article
Engineering, Electrical & Electronic
Yangjian Huang, Haifeng Hu
Summary: Cross-age face recognition is challenging due to the significant impact of aging on facial appearance. The proposed Age Adversarial Convolutional Neural Network (AA-CNN) effectively separates aging variations from facial features for stable and person-specific recognition. Extensive experiments demonstrate the superiority and effectiveness of the AA-CNN model on various aging face datasets.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2021)
Article
Mathematics
Se Hyun Nam, Yu Hwan Kim, Jiho Choi, Seung Baek Hong, Muhammad Owais, Kang Ryoung Park
Summary: This study proposes a new method for age estimation in low-light environments, using a generative adversarial network and convolutional neural network to compensate for the brightness of facial images captured in low-light conditions, achieving more accurate age estimation performance.
Article
Computer Science, Artificial Intelligence
Jian Han, Wei Wang, Sezer Karaoglu, Wei Zeng, Theo Gevers
Summary: The proposed method for age estimation of face videos is based on facial uv texture maps reconstructed from original frames of videos, restored by a Wasserstein-based GAN. Age prediction is then made from the completed uv mappings to simultaneously capture facial uv texture map and age characteristics. With the creation of the UvAge dataset containing videos from celebrities, the method outperforms other advanced age estimation methods according to extensive experiments.
COMPUTER VISION AND IMAGE UNDERSTANDING
(2021)
Article
Chemistry, Multidisciplinary
Suli Li, Hyo Jong Lee
Summary: This paper introduces a novel attention-based feature decomposition model, called the age-invariant features extraction network, which can effectively extract age-independent features and reduce age interference. Experimental results show significant improvements compared to existing methods on multiple datasets.
APPLIED SCIENCES-BASEL
(2022)
Article
Computer Science, Information Systems
Xiaohang Bian, Jianwu Li
Summary: This paper proposes a novel face aging/rejuvenation method named CACIAE, which uses a Res-Encoder, introduces rectangular kernel, and proposes consistent identity loss to generate more natural and identity-preserving face images. Experimental results demonstrate the effectiveness of the method.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Article
Computer Science, Artificial Intelligence
Chenmou Wu, Hyo Jong Lee
Summary: The study introduces a novel method that directly generates age-aware facial representations from various age groups for cross-age face recognition. By embedding age-semantic information into latent facial representations, the proposed method achieved over 97% accuracy in extensive experiments on multiple public aging face datasets, demonstrating its effectiveness for cross-age face recognition.
NEURAL COMPUTING & APPLICATIONS
(2022)
Article
Computer Science, Information Systems
Tao Chen, Guo-Sen Xie, Yazhou Yao, Qiong Wang, Fumin Shen, Zhenmin Tang, Jian Zhang
Summary: In this paper, a one-shot semantic image segmentation method is proposed that leverages multi-class information. Episodic training strategy, pyramid feature fusion module, and self-prototype guidance branch are introduced to improve segmentation accuracy and robustness. Experimental results demonstrate the superiority of this method.
IEEE TRANSACTIONS ON MULTIMEDIA
(2022)
Article
Computer Science, Software Engineering
Jiachao Zhang, Yang Gao, Yi Xu, Yunbin Huang, Yanming Yu, Xiangbo Shu
Summary: This paper proposes an image stitching technique using computational blending zone, which utilizes dynamic programming to select optimal seam-lines and optimal regions for image blending, in order to eliminate stitching traces and ghosting.
Article
Computer Science, Artificial Intelligence
Zhenhuang Cai, Guo-Sen Xie, Xingguo Huang, Dan Huang, Yazhou Yao, Zhenmin Tang
Summary: This paper proposes a simple yet effective method named MS-DeJOR for training robust models in the presence of web noise in deep neural networks. Unlike existing methods, MS-DeJOR decouples sample selection from the training procedure, uses a negative entropy term to prevent false positives from being overemphasized, and leverages accumulated predictions to refurbish noisy labels and re-weight training images for improved performance.
PATTERN RECOGNITION
(2023)
Article
Engineering, Electrical & Electronic
Xu-Yao Zhang, Guo-Sen Xie, Xiuli Li, Tao Mei, Cheng-Lin Liu
Summary: Learning to reject is crucial for humans to become smarter because it represents a unique self-awareness that machine intelligence lacks. This article provides a comprehensive overview of the topic from the perspectives of confidence, calibration, and discrimination. It discusses the importance of calibration in ensuring confidence matches the actual likelihood of correctness and the significance of discrimination in accepting positive samples while rejecting negative ones. The article analyzes three discrimination tasks - failure rejection, unknown rejection, and fake rejection - and presents a taxonomy of methods to solve these problems. By developing a discriminative and calibrated confidence, the decision-making process can become more practical, reliable, and secure.
PROCEEDINGS OF THE IEEE
(2023)
Article
Automation & Control Systems
Shiming Chen, Peng Zhang, Guo-Sen Xie, Qinmu Peng, Zehong Cao, Wei Yuan, Xinge You
Summary: This article proposes a novel DT synthesis method based on kernel similarity embedding, which effectively learns a synthesis model for high-dimensional DT from a small number of training samples. The experiments demonstrate that the learned kernel similarity embeddings provide discriminative representations for DTs and can generate realistic DT videos with higher speed and lower computation.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2023)
Review
Engineering, Electrical & Electronic
Guo-Sen Xie, Zheng Zhang, Huan Xiong, Ling Shao, Xuelong Li
Summary: This paper discusses zero-shot learning (ZSL) which focuses on distinguishing unseen class images by training the classifier on images from seen classes. Existing works ignore the information contained in image parts, which often contain discrimination information. In this paper, an end-to-end attention-based embedding network is proposed to discover meaningful parts and ensure knowledge transfer, with an adaptive thresholding strategy to reduce noise. Experimental results show that the models achieve superior results under both ZSL and GZSL settings.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2023)
Article
Computer Science, Artificial Intelligence
Rui Yan, Lingxi Xie, Jinhui Tang, Xiangbo Shu, Qi Tian
Summary: This paper proposes a Hierarchical Graph-based Cross Inference Network (HiGCIN) for group activity recognition. It constructs, learns, and infers three levels of information, namely body-region level, person level, and group-activity level. The approach is effective in capturing spatiotemporal dependencies and inferring with multilevel visual cues.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Computer Science, Artificial Intelligence
Xiangbo Shu, Binqian Xu, Liyan Zhang, Jinhui Tang
Summary: In the semi-supervised skeleton-based action recognition task, the current mainstream approach of contrastive learning has limitations in learning global-granularity features, dealing with ambiguous pairs, and contrasting cross-granularity pairs. To address these limitations, the proposed MAC-Learning method conducts inter- and intra-granularity contrastive pretext tasks to learn multi-granularity representations and utilizes the MAC-Loss to measure the agreement/disagreement between soft-positive/negative pairs. Experimental results on NTU RGB+D and Northwestern-UCLA datasets demonstrate the superiority of MAC-Learning in semi-supervised skeleton-based action recognition tasks.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Computer Science, Artificial Intelligence
Rui Yan, Lingxi Xie, Xiangbo Shu, Liyan Zhang, Jinhui Tang
Summary: Previous methods encode multiple pieces of information independently and concatenate them for classification, ignoring the potential role of instance information in the process of visual perception. This study presents a framework to progressively extract, reason, and predict dynamic cues of moving instances from videos for compositional action recognition.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Computer Science, Information Systems
Ting Guo, Jiye Liang, Guo-Sen Xie
Summary: This paper proposes a model called GIRL, which enhances the discrimination of region features through interactive region learning among a group of images, and achieves knowledge transfer between seen and unseen classes. Experimental results demonstrate the superior performance of the GIRL model over other methods on multiple datasets.
INFORMATION SCIENCES
(2023)
Article
Computer Science, Artificial Intelligence
Guo-Sen Xie, Xu-Yao Zhang, Tian-Zhu Xiang, Fang Zhao, Zheng Zhang, Ling Shao, Xuelong Li
Summary: This study proposes a balanced semantic embedding generative network (BSeGN) to address bias issues in zero-shot learning. By designing a feature-to-semantic embedding module and a multilevel feature integration module, the study achieves balanced prediction and enhanced interdomain features.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Binqian Xu, Xiangbo Shu, Jiachao Zhang, Guangzhao Dai, Yan Song
Summary: A novel spatiotemporal decouple-and-squeeze contrastive learning (SDS-CL) framework is proposed to comprehensively learn more abundant representations of skeleton-based actions. By jointly contrasting spatial-squeezing features, temporal-squeezing features, and global features, SDS-CL achieves performance gains compared with other competitive methods, as demonstrated by extensive experimental results on four public datasets.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Binqian Xu, Xiangbo Shu, Yan Song
Summary: In this research, a novel X-invariant Contrastive Augmentation and Representation learning (X-CAR) framework is proposed to obtain X-invariant features by adaptively augmenting and representing skeleton sequences. The method achieves better accuracy compared with other competitive methods in the semi-supervised action recognition scenario.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2022)
Article
Computer Science, Information Systems
Chuanyi Zhang, Qiong Wang, Guosen Xie, Qi Wu, Fumin Shen, Zhenmin Tang
Summary: This article introduces a method for learning fine-grained tasks from web data, which purifies noisy training sets by identifying and distinguishing noisy images, and trains models to alleviate the effects of noise.
IEEE TRANSACTIONS ON MULTIMEDIA
(2022)
Article
Computer Science, Information Systems
Tao Chen, Shui-Hua Wang, Qiong Wang, Zheng Zhang, Guo-Sen Xie, Zhenmin Tang
Summary: Unsupervised domain adaptation for semantic segmentation aims to transfer knowledge from a labeled source domain to an unlabeled target domain. This paper proposes a method to enhance adversarial learning based feature alignment to address the label noise and domain mismatch problems. It introduces a classification constrained discriminator to alleviate feature distortion, uses self-training to alleviate classifier overfitting, and proposes an efficient class centroid calculation module to reduce domain discrepancy.
IEEE TRANSACTIONS ON MULTIMEDIA
(2022)
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.