#SAHI

SAHI provides an efficient vision library for large-scale object detection and segmentation. It tackles the challenges of small object detection and large image inference, enhancing accuracy in practical applications. Supporting frameworks like YOLOv5, MMDetection, and Detectron2, it seamlessly integrates into current workflows. Features include sliced and standard prediction, COCO dataset functionalities, error analysis, and interactive visualization. Widely adopted with over 200 citations and competitive successes, SAHI advances computer vision technology.

Grounded SAM 2 utilizes Grounding DINO and Florence-2 to efficiently ground, segment, and track objects in video content. It offers features such as SAHI, SAM 2.1 compatibility, and custom video input support, aiming to simplify open-set detection and video tracking tasks.