Multi-Source Soft Pseudo-Label Learning with Domain Similarity-based Weighting for Semantic Segmentation

被引:2
|
作者
Matsuzaki, Shigemichi [1 ]
Masuzawa, Hiroaki [1 ]
Miura, Jun [1 ]
机构
[1] Toyohashi Univ Technol, Dept Comp Sci & Engn, Hibarigaoka 1-1,Tenpaku Cho, Toyohashi, Aichi, Japan
关键词
D O I
10.1109/IROS55552.2023.10342159
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
This paper describes a method of domain adaptive training for semantic segmentation using multiple source datasets that are not necessarily relevant to the target dataset. We propose a soft pseudo-label generation method by integrating predicted object probabilities from multiple source models. The prediction of each source model is weighted based on the estimated domain similarity between the source and the target datasets to emphasize contribution of a model trained on a source that is more similar to the target and generate reasonable pseudo-labels. We also propose a training method using the soft pseudo-labels considering their entropy to fully exploit information from the source datasets while suppressing the influence of possibly misclassified pixels. The experiments show comparative or better performance than our previous work and another existing multi-source domain adaptation method, and applicability to a variety of target environments.
引用
收藏
页码:5852 / 5857
页数:6
相关论文
共 50 条
  • [1] Multi-source pseudo-label learning of semantic segmentation for the scene recognition of agricultural mobile robots
    Matsuzaki, Shigemichi
    Miura, Jun
    Masuzawa, Hiroaki
    ADVANCED ROBOTICS, 2022, 36 (19) : 1011 - 1029
  • [2] Multi-Source Domain Adaptation with Collaborative Learning for Semantic Segmentation
    He, Jianzhong
    Jia, Xu
    Chen, Shuaijun
    Liu, Jianzhuang
    2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021, 2021, : 11003 - 11012
  • [3] Multi-source Domain Adaptation for Semantic Segmentation
    Zhao, Sicheng
    Li, Bo
    Yue, Xiangyu
    Gu, Yang
    Xu, Pengfei
    Hu, Runbo
    Chai, Hua
    Keutzer, Kurt
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 32 (NIPS 2019), 2019, 32
  • [4] Multi source-free domain adaptation based on pseudo-label knowledge mining
    Zhou, Fang
    Xu, Zun
    Wei, Wei
    Zhang, Lei
    PATTERN RECOGNITION LETTERS, 2025, 187 : 80 - 85
  • [5] Similarity-based Multi-label Learning
    Rossi, Ryan A.
    Ahmed, Nesreen K.
    Eldardiry, Hoda
    Zhou, Rong
    2018 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS (IJCNN), 2018,
  • [6] Continuous Pseudo-Label Rectified Domain Adaptive Semantic Segmentation with Implicit Neural Representations
    Gong, Rui
    Wang, Qin
    Danelljan, Martin
    Dai, Dengxin
    Van Gool, Luc
    2023 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR, 2023, : 7225 - 7235
  • [7] Multi-source domain adaptation for panoramic semantic segmentation
    Jiang, Jing
    Zhao, Sicheng
    Zhu, Jiankun
    Tang, Wenbo
    Xu, Zhaopan
    Yang, Jidong
    Liu, Guoping
    Xing, Tengfei
    Xu, Pengfei
    Yao, Hongxun
    INFORMATION FUSION, 2025, 117
  • [8] Similarity-based Multi-source Transfer Learning Approach for Time Series Classification
    Senanayaka, Ayantha
    Al Mamun, Abdullah
    Bond, Glenn
    Tian, Wenmeng
    Wang, Haifeng
    Fuller, Sara
    Falls, T. C.
    Rahimi, Shahram
    Bian, Linkan
    INTERNATIONAL JOURNAL OF PROGNOSTICS AND HEALTH MANAGEMENT, 2022, 13 (02) : 1 - 25
  • [9] Unsupervised multi-source domain adaptation for person re-identification via feature fusion and pseudo-label refinement
    Tian, Qing
    Cheng, Yao
    He, Sizhen
    Sun, Jixin
    COMPUTERS & ELECTRICAL ENGINEERING, 2024, 113
  • [10] Dual consistent pseudo label generation for multi-source domain adaptation without source data for medical image segmentation
    Cai, Binke
    Ma, Liyan
    Sun, Yan
    FRONTIERS IN NEUROSCIENCE, 2023, 17