EvRepSL: Event-Stream Representation via Self-Supervised Learning for Event-Based Vision

被引:0
|
作者
Qu, Qiang [1 ]
Chen, Xiaoming [2 ]
Chung, Yuk Ying [1 ]
Shen, Yiran [3 ]
机构
[1] Univ Sydney, Sch Comp Sci, Sydney, NSW 2050, Australia
[2] Beijing Technol & Business Univ, Sch Comp & Artificial Intelligence, Beijing 102401, Peoples R China
[3] Shandong Univ, Sch Software, Jinan 250100, Peoples R China
基金
北京市自然科学基金; 中国国家自然科学基金;
关键词
Cameras; Event detection; Optical flow; Estimation; Self-supervised learning; Noise; Computer vision; Accuracy; Generators; Noise reduction; Dynamic vision sensor; neuromorphic vision; event camera; representation learning; event-based vision; SENSOR;
D O I
10.1109/TIP.2024.3497795
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Event-stream representation is the first step for many computer vision tasks using event cameras. It converts the asynchronous event-streams into a formatted structure so that conventional machine learning models can be applied easily. However, most of the state-of-the-art event-stream representations are manually designed and the quality of these representations cannot be guaranteed due to the noisy nature of event-streams. In this paper, we introduce a data-driven approach aiming at enhancing the quality of event-stream representations. Our approach commences with the introduction of a new event-stream representation based on spatial-temporal statistics, denoted as EvRep. Subsequently, we theoretically derive the intrinsic relationship between asynchronous event-streams and synchronous video frames. Building upon this theoretical relationship, we train a representation generator, RepGen, in a self-supervised learning manner accepting EvRep as input. Finally, the event-streams are converted to high-quality representations, termed as EvRepSL, by going through the learned RepGen (without the need of fine-tuning or retraining). Our methodology is rigorously validated through extensive evaluations on a variety of mainstream event-based classification and optical flow datasets (captured with various types of event cameras). The experimental results highlight not only our approach's superior performance over existing event-stream representations but also its versatility, being agnostic to different event cameras and tasks.
引用
收藏
页码:6579 / 6591
页数:13
相关论文
共 50 条
  • [41] Meta-Learning and Self-Supervised Pretraining for Storm Event Imagery Translation
    Rugina, Ileana
    Dangovski, Rumen
    Simek, Olga
    Veillette, Mark
    Khorrami, Pooya
    Soljacic, Marin
    Cheung, Brian
    2023 IEEE HIGH PERFORMANCE EXTREME COMPUTING CONFERENCE, HPEC, 2023,
  • [42] Revisit Event Generation Model: Self-supervised Learning of Event-to-Video Reconstruction with Implicit Neural Representations
    Wang, Zipeng
    Lu, Yunfan
    Wang, Lin
    COMPUTER VISION - ECCV 2024, PT XLIX, 2025, 15107 : 321 - 339
  • [43] Stereo Depth Estimation via Self-supervised Contrastive Representation Learning
    Tukra, Samyakh
    Giannarou, Stamatia
    MEDICAL IMAGE COMPUTING AND COMPUTER ASSISTED INTERVENTION, MICCAI 2022, PT VII, 2022, 13437 : 604 - 614
  • [44] Self-Supervised Facial Motion Representation Learning via Contrastive Subclips
    Sun, Zheng
    Torrie, Shad A.
    Sumsion, Andrew W.
    Lee, Dah-Jye
    ELECTRONICS, 2023, 12 (06)
  • [45] Self-Supervised Video Representation Learning via Latent Time Navigation
    Yang, Di
    Wang, Yaohui
    Kong, Quan
    Dantcheva, Antitza
    Garattoni, Lorenzo
    Francesca, Gianpiero
    Bremond, Francois
    THIRTY-SEVENTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE, VOL 37 NO 3, 2023, : 3118 - 3126
  • [46] Self-Supervised Intensity-Event Stereo Matching
    Gu, Jinjin
    Zhou, Jinan
    Chu, Ringo Sai Wo
    Chen, Yan
    Zhang, Jiawei
    Cheng, Xuanye
    Zhang, Song
    Ren, Jimmy S.
    JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY, 2022, 66 (06)
  • [47] DocMAE: Document Image Rectification via Self-supervised Representation Learning
    Liu, Shaokai
    Feng, Hao
    Zhou, Wengang
    Li, Houqiang
    Liu, Cong
    Wu, Feng
    2023 IEEE INTERNATIONAL CONFERENCE ON MULTIMEDIA AND EXPO, ICME, 2023, : 1613 - 1618
  • [48] METRICBERT: TEXT REPRESENTATION LEARNING VIA SELF-SUPERVISED TRIPLET TRAINING
    Malkiel, Itzik
    Ginzburg, Dvir
    Barkan, Oren
    Caciularu, Avi
    Weill, Yoni
    Koenigstein, Noam
    2022 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), 2022, : 8142 - 8146
  • [49] Self-Supervised Representation Learning via Neighborhood-Relational Encoding
    Sabokrou, Mohammad
    Khalooei, Mohammad
    Adeli, Ehsan
    2019 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2019), 2019, : 8009 - 8018
  • [50] Self-supervised Health Representation Decomposition based on contrast learning
    Wang, Yilin
    Shen, Lei
    Zhang, Yuxuan
    Li, Yuanxiang
    Zhang, Ruixin
    Yang, Yongshen
    RELIABILITY ENGINEERING & SYSTEM SAFETY, 2023, 239