MSTN: A Multi-granular Spatial-Temporal Network for video-based person re-identification

Proceedings Paper Computer Science, Artificial Intelligence

Spatio-Temporal Representation Factorization for Video-based Person Re-Identification

Abhishek Aich et al.

Summary: This study introduces a new computational unit, STRF, to address common challenges in video-based person re-identification, such as appearance similarity, occlusions, and frame misalignment. STRF features explicit pathways for learning temporal and spatial features, capturing individual appearance and motion information to improve performance of existing architectures.

2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021) (2021)

添加到收藏夹

Proceedings Paper Computer Science, Artificial Intelligence

Pyramid Spatial-Temporal Aggregation for Video-based Person Re-Identification

Yingquan Wang et al.

Summary: This paper proposes a novel Pyramid Spatial-Temporal Aggregation (PSTA) framework for tracking the same person across videos. By progressively aggregating frame-level features and fusing hierarchical temporal features into video-level representation, it effectively utilizes short-term and long-term temporal information at different hierarchies. The Spatial-Temporal Aggregation Module (STAM) enhances the aggregation capability of PSTA with Spatial Reference Attention (SRA) and Temporal Reference Attention (TRA) blocks to explore spatial and temporal correlations.

2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021) (2021)

添加到收藏夹

Proceedings Paper Computer Science, Artificial Intelligence

Video-based Person Re-identification with Spatial and Temporal Memory Networks

Chanho Eom et al.

Summary: Video-based person re-identification is challenged by spatial and temporal distractors, and this translation introduces a novel Spatial and Temporal Memory Networks (STMN) to effectively address these challenges.

2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021) (2021)

添加到收藏夹

Proceedings Paper Computer Science, Artificial Intelligence

Watching You: Global-guided Reciprocal Learning for Video-based Person Re-identification

Xuehu Liu et al.

Summary: This paper proposes a Global-guided Reciprocal Learning (GRL) framework for video-based person re-identification, utilizing feature correlation maps of local and global features and a Temporal Reciprocal Learning (TRL) mechanism to enhance the performance. Experimental results show better performance compared to state-of-the-art approaches on three public benchmarks.

2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021 (2021)

添加到收藏夹

Article Computer Science, Hardware & Architecture

Point Cloud Video Streaming: Challenges and Solutions

Zhi Liu et al.

Summary: Volumetric video, seen as the next generation of video technology and a typical use case for 5G and beyond wireless communications, requires efficient transmission for its applications. This article introduces point cloud video technology and applications, along with discussing challenges and solutions in video streaming.

IEEE NETWORK (2021)

添加到收藏夹

Article Computer Science, Artificial Intelligence