Academics | The Hong Kong University of Science and Technology

CLAM-based Image-Caption Generation
P roject Background: The high cost of annotation in the medical field has led to a lack of annotated data related to whole slide imaging (WSI), thereby limiting the performance of many downstream tasks, such as training and application of pathology CLIP.

Project Objective: Utilize the existing capabilities of CLAM to perform weakly supervised - few-shot annotations on WSI images. For all WSI tasks, given <10 WSI data with only WSI-level labels, provide patch-level labels for all such WSI task images.

Research Progress: The process of generating patch-level labels from WSI-level labels has been completed, with an overall progress of 70%.