Evaluating SAM2's Role in Camouflaged Object Detection: From SAM to SAM2
This incremental finding highlights a specific limitation in SAM2 for researchers working on segmentation models, potentially guiding improvements in the SAM family.
The paper investigates a performance decline in SAM2's auto mode for camouflaged object detection compared to SAM, revealing that SAM2 struggles with perceiving objects without prompts in this challenging task.
The Segment Anything Model (SAM), introduced by Meta AI Research as a generic object segmentation model, quickly garnered widespread attention and significantly influenced the academic community. To extend its application to video, Meta further develops Segment Anything Model 2 (SAM2), a unified model capable of both video and image segmentation. SAM2 shows notable improvements over its predecessor in terms of applicable domains, promptable segmentation accuracy, and running speed. However, this report reveals a decline in SAM2's ability to perceive different objects in images without prompts in its auto mode, compared to SAM. Specifically, we employ the challenging task of camouflaged object detection to assess this performance decrease, hoping to inspire further exploration of the SAM model family by researchers. The results of this paper are provided in \url{https://github.com/luckybird1994/SAMCOD}.