DETR — Object Detection AI Model
DETR detects across the 91 COCO categories — people, cars, dogs, chairs — and returns both an annotated image and a text list of what it found. More accurate than YOLOS-Small and a little slower. CPU-capable.
What it does
DETR finds the objects in a photograph and returns each one with a box and a confidence score, across the same 91-category COCO label space as YOLOS: people, cars, dogs, chairs, cups.
Rather than scanning the image in pieces the way older detectors do, it reads the whole frame at once with a transformer and predicts every object directly. More accurate than YOLOS-Small and a little slower, and it runs on CPU.
When to use it
- General detection across the COCO categories, with no training
- Accuracy first, where YOLOS-Small trades some away for speed
- Several objects found and located in one picture at once
- Running on CPU rather than taking up a GPU to do it
- An annotated image plus a text list, for a later step to read
Details
- CategoryObject Detection
- Year2020
- LicenseApache-2.0
Compliance & Provenance
- ProviderOpen-source · Specialized
- EU AI Act RiskMinimal Risk
- Art. 50 TransparencyNot applicable
Inputs & Outputs
- ImageInputimage
Source image for object detection
- ImageOutputimage
Detected objects are highlighted with bounding boxes
- TextOutputstring
Detected types of objects with confidence scores
Tags
Alternatives in Object Detection
- DETR (Face Detector)
Face detection (DETR ResNet-50 fine-tuned). Fast.
- DETR (License Plate Detector)
License-plate detection (DETR ResNet-50). Fast.
- RF-DETR Medium
Real-time object detection on 80 COCO classes (RF-DETR Medium).
- YOLO-S
Fast object detection, 91 COCO classes (YOLOS-Small).
- YOLO-S (Fashion Item Detection)
Fashion item detection — 46 apparel classes (YOLOS).