Xiamen University, Intel, and DJI Jointly Release GIM: Learning Zero-Shot Image Matching from Internet Videos

Image matching is a fundamental task in computer vision. Recently, matching models based on deep learning have gradually gained popularity. Xiamen University, Intel, and DJI have introduced GIM: Learning Generalizable Image Matcher from Internet Videos. GIM enables matching models to learn strong generalization capabilities from internet videos. The GIM framework is suitable for all matching models, including DKM, LoFTR, and SuperGlue. The authors have proposed the first ...