Grounded SAM: Assembling Open-World Models for Diverse Visual Tasks
This work provides a versatile pipeline for connecting various vision models to solve diverse visual tasks, though it is incremental as it combines existing components.
The authors tackled the problem of enabling detection and segmentation of any region based on arbitrary text inputs by integrating Grounding DINO with SAM, achieving 48.7 mean AP on the SegInW zero-shot benchmark.
We introduce Grounded SAM, which uses Grounding DINO as an open-set object detector to combine with the segment anything model (SAM). This integration enables the detection and segmentation of any regions based on arbitrary text inputs and opens a door to connecting various vision models. As shown in Fig.1, a wide range of vision tasks can be achieved by using the versatile Grounded SAM pipeline. For example, an automatic annotation pipeline based solely on input images can be realized by incorporating models such as BLIP and Recognize Anything. Additionally, incorporating Stable-Diffusion allows for controllable image editing, while the integration of OSX facilitates promptable 3D human motion analysis. Grounded SAM also shows superior performance on open-vocabulary benchmarks, achieving 48.7 mean AP on SegInW (Segmentation in the wild) zero-shot benchmark with the combination of Grounding DINO-Base and SAM-Huge models.