A python library to build applications and systems with self-contained Computer Vision capabilities.
note: Install all dependencies before you start
https://github.com/OlafenwaMoses/ImageAI/#dependencies
With state-of-the-art deep learning algorithms like RetinaNet, YOLOv3, and TinyYOLOv3, ImageAI can accomplish all of these. You may use ImageAI to do picture analysis and detecting tasks.
Pre-trained dataset model
RetinaNet Model - resnet50_coco_best_v2.1.0.h5
YOLOv3 Model - yolo.h5
TinyYOLOv3 Model - yolo-tiny.h5
| Original image | Result |
|---|---|
 |
 |
Sample result :
car : 69.07421350479126
motorcycle : 99.39049482345581
motorcycle : 83.08526873588562
person : 67.89465546607971
car : 67.60438084602356
car : 58.61821174621582
person : 55.3065299987793
motorcycle : 95.38490176200867
motorcycle : 51.138150691986084