An Active and Contrastive Learning Framework for Fine-Grained Off-Road Semantic Segmentation

被引:8
|
作者
Gao, Biao [1 ,2 ]
Zhao, Xijun [3 ,4 ]
Zhao, Huijing [1 ,2 ]
机构
[1] Peking Univ, Key Lab Machine Percept MOE, Beijing 100871, Peoples R China
[2] Peking Univ, Sch Intelligence Sci & Technol, Beijing 100871, Peoples R China
[3] China North Artificial Intelligence & Innovat Res, Beijing 100072, Peoples R China
[4] China North Vehicle Res Inst, Beijing 100072, Peoples R China
基金
中国国家自然科学基金;
关键词
Semantics; Annotations; Adaptation models; Roads; Feature extraction; Task analysis; Labeling; Off-road; semantic segmentation; active learning; contrastive learning; VANISHING POINT DETECTION; CLASSIFICATION; TRACKING;
D O I
10.1109/TITS.2022.3218403
中图分类号
TU [建筑科学];
学科分类号
0813 ;
摘要
Off-road semantic segmentation with fine-grained labels is necessary for autonomous vehicles to understand driving scenes, as the coarse-grained road detection can not satisfy off-road vehicles with various mechanical properties. Fine-grained semantic segmentation in off-road scenes usually has no unified category definition due to ambiguous nature environments, and the cost of pixel-wise labeling is extremely high. Furthermore, semantic properties of off-road scenes can be very changeable due to various precipitations, temperature, defoliation, etc. To address these challenges, this research proposes an active and contrastive learning-based method that does not rely on pixel-wise labels, but only on patch-based weak annotations for model learning. There is no need for predefined semantic categories, the contrastive learning-based feature representation and adaptive clustering will discover the category model from scene data. In order to actively adapt to new scenes, a risk evaluation method is proposed to discover and select hard frames with high-risk predictions for supplemental labeling, so as to update the model efficiently. Experiments conducted on our self-developed off-road dataset and DeepScene dataset demonstrate that fine-grained semantic segmentation can be learned with only dozens of weakly labeled frames, and the model can efficiently adapt across scenes by weak supervision, while achieving almost the same level of performance as typical fully supervised baselines.
引用
收藏
页码:564 / 579
页数:16
相关论文
共 50 条
  • [1] Fine-Grained Off-Road Semantic Segmentation and Mapping via Contrastive Learning
    Gao, Biao
    Hu, Shaochi
    Zhao, Xijun
    Zhao, Huijing
    2021 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS (IROS), 2021, : 5950 - 5957
  • [2] OFFSEG: A Semantic Segmentation Framework For Off-Road Driving
    Viswanath, Kasi
    Singh, Kartikeya
    Jiang, Peng
    Sujit, P. B.
    Saripalli, Srikanth
    2021 IEEE 17TH INTERNATIONAL CONFERENCE ON AUTOMATION SCIENCE AND ENGINEERING (CASE), 2021, : 354 - 359
  • [3] Semantic Segmentation with Transfer Learning for Off-Road Autonomous Driving
    Sharma, Suvash
    Ball, John E.
    Tang, Bo
    Carruth, Daniel W.
    Doude, Matthew
    Islam, Muhammad Aminul
    SENSORS, 2019, 19 (11)
  • [4] OFFSED: Off-Road Semantic Segmentation Dataset
    Neigel, Peter
    Rambach, Jason
    Stricker, Didier
    VISAPP: PROCEEDINGS OF THE 16TH INTERNATIONAL JOINT CONFERENCE ON COMPUTER VISION, IMAGING AND COMPUTER GRAPHICS THEORY AND APPLICATIONS - VOL. 4: VISAPP, 2021, : 552 - 557
  • [5] ROSS: Radar Off-Road Semantic Segmentation
    Jiang, Peng
    Saripalli, Srikanth
    EXPERIMENTAL ROBOTICS, ISER 2023, 2024, 30 : 264 - 273
  • [6] GRAPH FINE-GRAINED CONTRASTIVE REPRESENTATION LEARNING
    Tang, Hui
    Liang, Xun
    Guo, Yuhui
    Zheng, Xiangping
    Wu, Bo
    2022 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), 2022, : 3478 - 3482
  • [7] TALS: A Framework For Text Analysis, Fine-Grained Annotation, Localisation and Semantic Segmentation
    Jaradat, Shatha
    Dokoohaki, Nima
    Wara, Ummul
    Goswami, Mallu
    Hammar, Kim
    Matskin, Mihhail
    2019 IEEE 43RD ANNUAL COMPUTER SOFTWARE AND APPLICATIONS CONFERENCE (COMPSAC), VOL 2, 2019, : 201 - 206
  • [8] Off-Road LiDAR Intensity Based Semantic Segmentation
    Viswanath, Kasi
    Jiang, Peng
    Sujit, P. B.
    Saripalli, Srikanth
    EXPERIMENTAL ROBOTICS, ISER 2023, 2024, 30 : 608 - 617
  • [9] Strengthen contrastive semantic consistency for fine-grained image classification
    Wang, Yupeng
    Wang, Yongli
    Ye, Qiaolin
    Lang, Wenxi
    Xu, Can
    PATTERN ANALYSIS AND APPLICATIONS, 2025, 28 (02)
  • [10] Enhanced Deep Learning Framework for Fine-Grained Segmentation of Fashion and Apparel
    Usmani, Usman Ahmad
    Happonen, Ari
    Watada, Junzo
    INTELLIGENT COMPUTING, VOL 2, 2022, 507 : 29 - 44