Towards automatic visual inspection: A weakly supervised learning method for industrial applicable object detection

Ce Ge, Jing Wang, Jingyu Wang, Qi Qi, Haifeng Sun, Jianxin Liao

Abstract: Highlights•Introduced weakly supervised deep learning to industrial automation for automatic visual inspection.•A new proposal-guided segmentation-aggregation framework for weakly supervised object detection.•Graph-based image segmentation and dropscore strategy to address the common failures.•Evaluation and analysis on collected insulator images and public benchmarks.