Object Detection

This is exercise for object detection to help us building a dataset containing vision & trajectory data.

ssd_mobilenet and tiny-yolo(coco) for realtime video input

YOLOV3 for image input

Usage

1. tiny-yolo

• download the tiny-yolo file and put it to model_data file

$ python3 test_tiny_yolo.py 

2. ssd-mobilenet

$ python3 test_ssd_mobilenet.py 

3. YOLOV3

• download the yolov3 file and put it to model_data file

$ python3 test_yolov3.py 

YOLOV2 vs YOLOV3

• YOLOV2

• YOLOV3

An Instance

• at the porch of WHU

Acknowledgments

• Thanks to keras-yolo3 for yolov3-keras part