Bootstrapped Self-Supervised Training with Monocular Video for Semantic Segmentation and Depth Estimation

被引:0
|
作者
Zhang, Yihao [1 ]
Leonard, John J. [1 ]
机构
[1] MIT, Comp Sci & Artificial Intelligence Lab, Cambridge, MA 02139 USA
关键词
D O I
10.1109/IROS51168.2021.9636330
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
For a robot deployed in the world, it is desirable to have the ability of autonomous learning to improve its initial pre-set knowledge. We formalize this as a bootstrapped self-supervised learning problem where a system is initially bootstrapped with supervised training on a labeled dataset and we look for a self-supervised training method that can subsequently improve the system over the supervised training baseline using only unlabeled data. In this work, we leverage temporal consistency between frames in monocular video to perform this bootstrapped self-supervised training. We show that a well-trained state-of-the-art semantic segmentation network can be further improved through our method. In addition, we show that the bootstrapped self-supervised training framework can help a network learn depth estimation better than pure supervised training or self-supervised training.
引用
收藏
页码:2420 / 2427
页数:8
相关论文
共 50 条
  • [1] Self-Supervised Monocular Depth Estimation Method for Joint Semantic Segmentation
    Song, Xiaogang
    Hu, Haoyue
    Ning, Jingyu
    Liang, Li
    Lu, Xiaofeng
    Hei, Xinhong
    [J]. Jisuanji Yanjiu yu Fazhan/Computer Research and Development, 2024, 61 (05): : 1336 - 1347
  • [2] Graph semantic information for self-supervised monocular depth estimation
    Zhang, Dongdong
    Wang, Chunping
    Wang, Huiying
    Fu, Qiang
    [J]. PATTERN RECOGNITION, 2024, 156
  • [3] Self-Supervised Pretraining With Monocular Height Estimation for Semantic Segmentation
    Xiong, Zhitong
    Chen, Sining
    Shi, Yilei
    Zhu, Xiao Xiang
    [J]. IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2024, 62
  • [4] RETHINKING TRAINING OBJECTIVE FOR SELF-SUPERVISED MONOCULAR DEPTH ESTIMATION: SEMANTIC CUES TO RESCUE
    Li, Keyao
    Li, Ge
    Li, Thomas
    [J]. 2021 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP), 2021, : 3308 - 3312
  • [5] MonoDVPS: A Self-Supervised Monocular Depth Estimation Approach to Depth-aware Video Panoptic Segmentation
    Petrovai, Andra
    Nedevschi, Sergiu
    [J]. 2023 IEEE/CVF WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION (WACV), 2023, : 3076 - 3085
  • [6] A LIGHTWEIGHT SELF-SUPERVISED TRAINING FRAMEWORK FOR MONOCULAR DEPTH ESTIMATION
    Heydrich, Tim
    Yang, Yimin
    Du, Shan
    [J]. 2022 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), 2022, : 2265 - 2269
  • [7] Domain Adaptive Semantic Segmentation with Self-Supervised Depth Estimation
    Wang, Qin
    Dai, Dengxin
    Hoyer, Lukas
    Van Gool, Luc
    Fink, Olga
    [J]. 2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021), 2021, : 8495 - 8505
  • [8] ThreeWays to Improve Semantic Segmentation with Self-Supervised Depth Estimation
    Hoyer, Lukas
    Dai, Dengxin
    Chen, Yuhua
    Koring, Adrian
    Saha, Suman
    Van Gool, Luc
    [J]. 2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021, 2021, : 11125 - 11135
  • [9] Plugging Self-Supervised Monocular Depth into Unsupervised Domain Adaptation for Semantic Segmentation
    Cardace, Adriano
    De Luigi, Luca
    Ramirez, Pierluigi Zama
    Salti, Samuele
    Di Stefano, Luigi
    [J]. 2022 IEEE WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION (WACV 2022), 2022, : 1999 - 2009
  • [10] Digging Into Self-Supervised Monocular Depth Estimation
    Godard, Clement
    Mac Aodha, Oisin
    Firman, Michael
    Brostow, Gabriel
    [J]. 2019 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2019), 2019, : 3827 - 3837