Computer Vision Toolbox Model for Grounding DINO Object Detection

Grounding DINO is a zero-shot pre-trained Vision Language Model (VLM) that enables open vocabulary, text-prompted object detection.

Grounding DINO enables zero-shot object detection from textual inputs, without requiring dedicated class training on the input term. It can therefore detect objects outside of its training set. It combines a Transformer-based DINO object detector with grounded pre-training.