Training method and device of small sample semantic segmentation model, equipment and medium

• Main Authors: YAN RUGEN, JIANG BIN, WANG CHENYANG, HU CHANGLONG, XING XING, WANG XUEYAN
• Format: Patent
• Language: Chinese; English
• Published: 28.05.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS

The invention provides a small sample semantic segmentation model training method and device, equipment and a medium, and the method comprises the steps: obtaining an image data set which comprises a plurality of support images and a plurality of query images; constructing an initial small sample semantic segmentation model based on an autocorrelation prior guidance and self-enhancement feature fusion network structure; performing feature extraction processing on the image data set to generate image feature information; and training the initial small sample semantic segmentation model based on the image feature information to generate a target small sample semantic segmentation model. According to the method, the target information in each layer of image features generated by the query image can be fully explored, and in addition, the invention also provides a new feature fusion method, so that the features containing different semantic information are fully fused, and the segmentation performance and general