Optics and Precision Engineering, Volume. 31, Issue 13, 2000(2023)
Semi-supervised instance object detection method based on SVD co-training
Figures & Tables(9)
Fig. 3. Comparison of Gram matrix before and after convergence， decomposition and finetuning
Fig. 5. Detection accuracy of pre-training SVD-RCNN on GMU dataset
Table 1. Results of ablation experiment
View tableView in ArticleTable 1. Results of ablation experiment
数据集 关键样本数量 收敛-分解-微调 自适应的自标注 mAP/% GMU 199 72.1 199 √ 77.6 199 √ 75.8 199 √ √ 79.3
Table 2. Comparison of detection accuracy of training methods on different datasets
View tableView in ArticleTable 2. Comparison of detection accuracy of training methods on different datasets
数据集 关键样本数量 方 法 mAP/% GMU 199 半监督SVD-RCNN 79.3 3 851 全监督Faster RCNN 81.3 AVD 255 半监督SVD-RCNN 75.6 5 083 全监督Faster RCNN 78.5 BHID 55 半监督SVD-RCNN 76.1 1 110 全监督Faster RCNN 78.4
Table 3. Comparison of semi-supervised methods on GMU
View tableView in ArticleTable 3. Comparison of semi-supervised methods on GMU
数据集
（标注样本数量）
基础网络 方 法 mAP/% GMU
（199）
Faster RCNN Unbiased Teacher ［ 7 ］76.5 Soft Teacher ［ 19 ］78.3 本文算法 79.3
Tools
Get Citation
Copy Citation Text
Rui WANG, Siyang FAN, Jingwen XU, Zhiqing WEN. Semi-supervised instance object detection method based on SVD co-training[J]. Optics and Precision Engineering, 2023, 31(13): 2000
Paper Information
Category: Information Sciences
Received: Jul. 27, 2022
Accepted: --
Published Online: Jul. 26, 2023
The Author Email: WANG Rui (wangr@buaa.edu.cn)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20