yolo v2 is an object detection algorithm using deep learning.
This repository contains learning and testing in keras and tensorflow.
- refactor trainning code
- tracking function(SORT and Deep SORT) based object detection of yolov2
- demo gif
- python3.6
- opencv for python2 involving movie function
- dataset(image and annotation using XML format)
- keras(tensorflow)
If you want to learn with your dataset you need to edit config.json.
Please prepare the annotation in XML format with reference to yolov 2 formula.
python train.py -c config.json
python predict.py -c config.json -w path/to/pretrained_weights -i path/to/image -o path/to/output_dir
todo
Tracking code performed by SORT algorithm using Kalman filter.
todo