Fabric defect detection method based on EFC feature extraction and mixed loss function

The invention discloses a fabric defect detection method based on EFC feature extraction and a mixed loss function. The fabric defect detection method comprises the steps that fabric pictures containing defects are collected and preprocessed, and a data set is randomly divided into a training set and an evaluation set; labeling defect areas on the fabric pictures containing the defects by using LabelImg, generating labeling files, and storing the labeling files into a corresponding training set and an evaluation set; the YOLOv8 model is improved, and a fabric defect detection network is constructed; training the fabric defect detection network by using the training set, and evaluating the performance of the trained fabric defect detection network by using the evaluation set to obtain a fabric defect detection model; and inputting the to-be-detected picture into a fabric defect detection model for detection, wherein a detection result comprises defect types and confidence coefficients. The invention discloses