groundingdino

X-AnyLabeling: A lightweight, efficient, and unified cross-platform desktop application for annotating text, image, video, point cloud and multimodal data, combining versatile built-in tools with state-of-the-art AI models and flexible multi-format export.

Based on GroundingDino and SAM, use semantic strings to segment any element in an image. The comfyui version of sd-webui-segment-anything.

使用onnxruntime部署GroundingDINO开放世界目标检测，包含C++和Python两个版本的程序

juxtapose: Multi-Person Pose Tracking Inference SDK with RTMDet, YOLOv8, GDino, RTMPose (ONNX) & Trackers (ByteTrack & BotSORT) & Tapnet with custom ROIs + FastAPI GPU exe

Extract objects from images or video via text prompt

基于 React、C++20、SQLite、RAM 与 GroundingDINO 的自动图像标注平台，支持同步/异步任务和 COCO/YOLO 导出。

This project represents a GroundingDINO Inference (zero-shot object detection) procedure with both methods (CLI and Script). This implementation will help the reader to know the sequence of commands and exemplifying commands for running a quick zero-shot object detection. Additionally, the reader may get insight into code (script) execution.

GrabTag cuts one reference image into named regions and gives every phrase its own box: tags are shape/position contracts, not material locks (obey them or re-materialise the same shape). No LoRA, no style drift. Reference at native size inherits patterns, not just silhouettes. Pure Python, CPU-only (autoregion: GroundingDINO + SAM2).

Support batch inference of Grounding DINO. "Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection"

Prompt based automatic annotation

Language-guided semantic robot navigation in NVIDIA Isaac Sim (ROS 2): SLAM + Nav2 + GroundingDINO scene graph + a Qwen2.5-VL robot brain that turns natural-language commands into navigation goals.

Text-free breast ultrasound video segmentation | ACM Multimedia 2026

language‑guided object detection that outputs a single pick target for robotic pick‑and‑place tasks

Turns plain english commands into robot navigation via open-vocabulary object detection and LLM-driven replanning on failure.

Run GroundingDINO as a local REST API using Docker. Easily send images and text prompts for object grounding.

Open vocabulary object detection using GroundingDINO with natural language prompts.

VLM-assisted object detection, COCO dataset generation, and YOLO fine-tuning tools.

Official implementation of KARINA (IEEE CAI 2026): LMM-generated ingredient knowledge for RGB-D nutrition estimation.

Real-robot language-guided pick/place on Sagittarius: GroundingDINO, calibrated planar mapping, and rule-based task sequencing over ROS1/MoveIt.