Building and road detection from remote sensing images based on weights adaptive multi-teacher collaborative distillation using a fused knowledge

被引:4
|
作者
Chen, Ziyi [1 ]
Deng, Liai [1 ]
Gou, Jing [1 ]
Wang, Cheng [2 ]
Li, Jonathan [3 ,4 ]
Li, Dilong [1 ]
机构
[1] Huaqiao Univ, Dept Comp Sci & Technol, Fujian Key Lab Big Data Intelligence & Secur, Xiamen Key Lab Data Secur & Blockchain Technol, 668 Jimei Rd, Xiamen 361021, Peoples R China
[2] Xiamen Univ, Sch Informat, Xiamen 361005, Peoples R China
[3] Univ Waterloo, Dept Geog & Environm Management, Waterloo, ON N2L 3G1, Canada
[4] Univ Waterloo, Dept Syst Design Engn, Waterloo, ON N2L 3G1, Canada
基金
中国国家自然科学基金;
关键词
Knowledge distillation; Remote sensing; Building extraction; Road extraction; CLASSIFICATION; NETWORK;
D O I
10.1016/j.jag.2023.103522
中图分类号
TP7 [遥感技术];
学科分类号
081102 ; 0816 ; 081602 ; 083002 ; 1404 ;
摘要
Knowledge distillation is one effective approach to compress deep learning models. However, the current distillation methods are relatively monotonous. There are still rare studies about the combination of distillation strategies using multiple types of knowledge and employing multiple teacher models. Besides, how to optimize the weights among different teacher models is still an open problem. To address these issues, this paper proposes a novel approach for knowledge distillation, which effectively enhances the robustness of the distilled student model by a weights adaptive multi-teacher collaborative distillation. Moreover, the proposed method utilizes feature knowledge exchange guidance between teacher networks to transfer more comprehensive feature knowledge to the student model, which further improves the learning capability of hidden layers' details. The extensive experimental results demonstrate that the proposed method achieves state-of-the-art performance on Massachusetts Roads Dataset, LRSNY Roads Dataset, and WHU Building Dataset. Specifically, under the guidance of the first ensemble of teacher networks, we obtained IoU scores of 47.33%, 78.15%, and 80.71%, respectively. Under the guidance of the second ensemble of teacher networks, we obtained IoU scores of 48.56%, 79.51%, and 81.35%, respectively.
引用
收藏
页数:13
相关论文
共 50 条
  • [21] Optimized Lithological Mapping from Multispectral and Hyperspectral Remote Sensing Images Using Fused Multi-Classifiers
    Pal, Mahendra
    Rasmussen, Thorkild
    Porwal, Alok
    [J]. REMOTE SENSING, 2020, 12 (01)
  • [22] U-Net Based Road Area Guidance for Crosswalks Detection from Remote Sensing Images
    Chen, Ziyi
    Luo, Ruixiang
    Li, Jonathan
    Du, Jixiang
    Wang, Cheng
    [J]. CANADIAN JOURNAL OF REMOTE SENSING, 2021, 47 (01) : 83 - 99
  • [23] Building-Road Collaborative Extraction From Remote Sensing Images via Cross-Task and Cross-Scale Interaction
    Guo, Haonan
    Su, Xin
    Wu, Chen
    Du, Bo
    Zhang, Liangpei
    [J]. IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2024, 62 : 1 - 16
  • [24] AN EFFECTIVE ROAD EXTRACTION METHOD from REMOTE SENSING IMAGES BASED on SELF-ADAPTIVE THRESHOLD FUNCTION
    Wang, Zhuozheng
    Zhang, Meng
    Liu, Wei
    [J]. 2019 ASIA-PACIFIC SIGNAL AND INFORMATION PROCESSING ASSOCIATION ANNUAL SUMMIT AND CONFERENCE (APSIPA ASC), 2019, : 455 - 460
  • [25] Multi-Scale DenseNets-Based Aircraft Detection from Remote Sensing Images
    Wang, Yantian
    Li, Haifeng
    Jia, Peng
    Zhang, Guo
    Wang, Taoyang
    Hao, Xiaoyun
    [J]. SENSORS, 2019, 19 (23)
  • [26] Road extraction from remote sensing images based on a multi-scale asymmetric dual attention mechanism
    Qu, Shenming
    Liu, Suchen
    Han, Fengyu
    Xie, Yuan
    [J]. REMOTE SENSING LETTERS, 2024, 15 (08) : 751 - 761
  • [27] Building Footprint Extraction From VHR Remote Sensing Images Combined With Normalized DSMs Using Fused Fully Convolutional Networks
    Bittner, Ksenia
    Adam, Fathalrahman
    Cui, Shiyong
    Koerner, Marco
    Reinartz, Peter
    [J]. IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, 2018, 11 (08) : 2615 - 2629
  • [28] A KNOWLEDGE-BASED METHOD FOR ROAD DAMAGE DETECTION USING HIGH-RESOLUTION REMOTE SENSING IMAGE
    Wang, Jianhua
    Qin, Qiming
    Zhao, Jianghua
    Ye, Xin
    Qin, Xuebin
    Yang, Xiucheng
    Wang, Jun
    Zheng, Xiaopo
    Sun, Yuejun
    [J]. 2015 IEEE INTERNATIONAL GEOSCIENCE AND REMOTE SENSING SYMPOSIUM (IGARSS), 2015, : 3564 - 3567
  • [29] Cropland encroachment detection via dual attention and multi-loss based building extraction in remote sensing images
    Wang, Junshu
    Cai, Mingrui
    Gu, Yifan
    Liu, Zhen
    Li, Xiaoxin
    Han, Yuxing
    [J]. FRONTIERS IN PLANT SCIENCE, 2022, 13
  • [30] Automatic building segmentation from remote sensing images using multi-layer level set framework
    Guo, Jing
    Jiang, Jie
    Cao, Shixiang
    [J]. Hongwai yu Jiguang Gongcheng/Infrared and Laser Engineering, 2014, 43 (04): : 1332 - 1337