DEIM and DEIMv2

These two tools are **ecosystem siblings**, specifically an original research project and its direct successor, with DEIMv2 building upon and improving DEIM by integrating more recent object detection advancements like DINOv3 for real-time performance.

DEIM
63
Established
DEIMv2
61
Established
Maintenance 16/25
Adoption 10/25
Maturity 16/25
Community 21/25
Maintenance 16/25
Adoption 10/25
Maturity 15/25
Community 20/25
Stars: 1,464
Forks: 189
Downloads:
Commits (30d): 1
Language: Python
License:
Stars: 1,590
Forks: 172
Downloads:
Commits (30d): 4
Language: Jupyter Notebook
License:
No Package No Dependents
No Package No Dependents

About DEIM

Intellindust-AI-Lab/DEIM

[CVPR 2025] DEIM: DETR with Improved Matching for Fast Convergence

DEIM is a training framework that significantly improves object detection models by making them faster and more accurate. It takes raw image data and processes it to quickly identify and pinpoint objects within those images. This is ideal for researchers and engineers working on computer vision applications who need to develop highly efficient and precise object detection systems.

real-time object detection computer vision engineering image analysis AI model training edge AI deployment

About DEIMv2

Intellindust-AI-Lab/DEIMv2

[DEIMv2] Real Time Object Detection Meets DINOv3

This project offers highly accurate real-time object detection, instance segmentation, and human pose estimation. It takes live video feeds or image streams as input and identifies specific objects, separates them from their background, and detects human body positions, outputting precise bounding boxes, segmentation masks, and skeletal keypoints. This is ideal for professionals in manufacturing, security, or sports analytics who need to monitor or analyze visual data in real-time.

real-time-video-analytics manufacturing-quality-control security-surveillance human-activity-recognition sports-performance-analysis

Scores updated daily from GitHub, PyPI, and npm data. How scores work