We explore a one-stage method for surface anomaly detection in industrial scenarios. On one side, encoder-decoder segmentation network is constructed to capture small targets as much as possible, and then dual background suppression mechanisms are designed to reduce noise patterns in coarse and fine manners. On the other hand, a classification module without learning parameters is built to reduce information loss in small targets due to the inexistence of successive down-sampling processes. Experimental results demonstrate that our one-stage detector achieves state-of-the-art performance in terms of precision, recall and f-score.

Textured surface anomaly detection is a significant task in industrial scenarios. In order to further improve the detection performance, we proposed a novel two-stage approach with an attention mechanism. Firstly, in the segmentation network, the feature extraction and anomaly attention modules are designed to capture the detail information as much as possible and focus on the anomalies, respectively. To strike dynamic balances between these two parts, an adaptive scheme where learnable parameters are gradually optimized is introduced. Subsequently, the weights of the segmentation network are frozen, and the outputs are fed into the classification network, which is trained independently in this stage. Finally, we evaluate the proposed approach on DAGM 2007 dataset which consists of diverse textured surfaces with weakly-labeled anomalies, and the experiments demonstrate that our method can achieve 100% detection rates in terms of TPR (True Positive Rate) and TNR (True Negative Rate).