When SAM Meets Shadow Detection
This is an incremental evaluation of an existing method on a new task, highlighting limitations for researchers in computer vision.
The study tested the Segment Anything Model (SAM) on shadow detection across four benchmarks, finding its performance unsatisfactory compared to specialized models, with no concrete numbers provided.
As a promptable generic object segmentation model, segment anything model (SAM) has recently attracted significant attention, and also demonstrates its powerful performance. Nevertheless, it still meets its Waterloo when encountering several tasks, e.g., medical image segmentation, camouflaged object detection, etc. In this report, we try SAM on an unexplored popular task: shadow detection. Specifically, four benchmarks were chosen and evaluated with widely used metrics. The experimental results show that the performance for shadow detection using SAM is not satisfactory, especially when comparing with the elaborate models. Code is available at https://github.com/LeipingJie/SAMSh.