The results is showing the frame where bat hit the ball.
git clone https://github.com/MotiBaadror/cricket_video_analysis
cd cricket_video_analysis
pip install - r requirements.txt
download the pretrained model from here
Keep the weights in the checkpoints folder
Keep the video in data/ folder
Use inference_tf2_colab.ipynb to get the results from models of tensorflow model.
You can try several different configs available here
cd src
python image_ai_inference.py
image_ai_inference.py to get the json for the cooridinate
results will be saved in the results dir
after that to get the frame where bat hits ball
python generate_results.py
To get the final output I am calculating euclidian distence between bet and ball.
dst = math.sqrt((x1-x2)**2+(y1-y2)**2)
The notebook contain all the helper code which is not organised.