View article

Creating large-scale, well-annotated datasets is vital for training AI algorithms in tumor detection. However, with limited resources, it is challenging to determine the best type of annotations when annotating massive amounts of unlabeled data. To address this issue, we focus on polyps in colonoscopy videos and pancreatic tumors in abdominal CT scans, both requiring extensive pixel-wise annotation due to the high dimensional nature of the data. In this paper, we develop a new annotation strategy, termed Drag&Drop, which simplifies the annotation process to drag and drop, proving more efficient for temporal and volumetric imaging. Furthermore, we introduce a novel weakly supervised learning method based on the watershed algorithm to leverage Drag&Drop annotations. Experimental results show that, with limited resources, allocating weak annotations from diverse patients enhances model robustness more effectively than per-pixel ones on a limited set of images. In summary, this research proposes an efficient annotation strategy that is useful for creating large-scale datasets for screening tumors in various medical modalities.