Grounding Dino Base
Grounding DINO is an open-set object detection model that achieves zero-shot object detection capabilities by combining the DINO detector with a text encoder.
Downloads 1.1M
Release Time : 9/25/2023
Model Overview
This model extends closed-set object detection to open-set detection, enabling the detection of objects not annotated in the training set through text descriptions. It achieves an outstanding performance of 52.5 AP on COCO zero-shot detection.
Model Features
Open-set Detection Capability
Detects new objects without specific category training, breaking the closed-set limitations of traditional detectors.
Text-guided Detection
Guides the detection process through natural language descriptions, achieving semantic-level object localization.
Zero-shot Transfer
Performs excellently on unseen object categories, such as achieving 52.5 AP on COCO zero-shot detection.
Model Capabilities
Zero-shot Object Detection
Multi-category Joint Detection
Text-guided Image Analysis
Use Cases
Intelligent Surveillance
Anomalous Object Detection
Detects abnormal items in surveillance footage through text descriptions
Can identify new types of hazardous items not included in the training data.
Content Moderation
Prohibited Content Identification
Detects image violations based on dynamically defined text rules
Adapts to new moderation rules without requiring retraining.
Featured Recommended AI Models