曲阜师范大学学报（自然科学版）2025，Vol.51Issue(3)：74-80,7.DOI:10.3969/j.issn.1001-5337.2025.3.074
基于YOLO-DA的商品识别算法
The commodity recognition algorithm based on YOLO-DA
摘要
Abstract
In this paper,a single-stage domain adaptive commodity recognition algorithm YOLO-DA(YOLO-domain adaptive)is proposed.Firstly,adaptation adjustments are made to the YOLO algorithm for cross-domain tasks and the RPC dataset.Secondly,the neck network structure is redesigned,incorpora-ting the BiFPN concept to re-fuse features at multiple scales.Finally,a Gradient Reversal Layer is added behind the backbone network for adversarial training on the training and testing sets,further approaching the goal of do-main adaptation.The training results of the improved network model on the RPC dataset show that the mean aver-age precision(mAP)reaches 65.25％.Compared with the baseline network,the detection accuracy is signifi-cantly improved,and the cases of missed detection and false detection are notably reduced.关键词
YOLOv7/VariFocal Loss/BiFPN/领域自适应/梯度反转层Key words
YOLOv7/VariFocal Loss/BiFPN/domain adaptive/gradient reversal layer分类
计算机与自动化引用本文复制引用
田海峰,邱茂顺,张维健..基于YOLO-DA的商品识别算法[J].曲阜师范大学学报（自然科学版）,2025,51(3):74-80,7.基金项目
曲阜师范大学科技项目(kj2021hx054). （kj2021hx054）
