A cotton fabric defect small sample semantic segmentation method, device and storage medium

Abstract

The application provides a cotton fabric flaw small sample semantic segmentation method, device and storage medium, and relates to the field of textile quality monitoring.The method comprises the following steps: acquiring an initial data set of cotton fabric flaws, preprocessing cotton fabric flaw images in the initial data set to obtain an augmented data set, improving based on a PATNet model, constructing a cotton fabric flaw small sample semantic segmentation model, training by using the augmented data set, and acquiring a semantic segmentation result of a to-be-segmented image by using the trained cotton fabric flaw small sample semantic segmentation model.The device and the storage medium are used for implementing the method.The application has the beneficial effects that the small sample semantic segmentation model is applied to cotton fabric flaw segmentation, has good segmentation performance and cross-domain effect, and is practical;the double-branch decoder structure is proposed in the model and the self-attention mechanism is fused, the number of samples required for model training is reduced, the segmentation accuracy of the flaw target area is improved, and the applicability of the model is improved.

Images