• Deep learning models (Classification and Object Detection using CNN) for automatic generation of text/ audio-based ball by ball commentary of a cricket match.
• Do ⭐ the repository if it helped you in anyway.
• Website URL: https://sites.google.com/view/cric-talk
• This is not an open source project so I have not updated the code in this repo. You can visit the above given website to know more about the project. In case of any other details you can reach me by email or over LinkedIn.
final_3.mp4
The use of artificial intelligence in sports isn’t a major surprise, given the advancement in technology. However, automated commentary is still a fantasy and the way AI and ML are developing, it could soon be a reality. The development of such a system will change the way of sports broadcasting. Hence, in this work we have tried to generate a part of automatic ball-by-ball commentary of a cricket match.
We have made a dataset of more than 2.3 lakh images from videos. The dataset was made by converting video to frames and cropping the required objects from the frames.
The dataset was divided into 80:20 ratio for training and testing of individual models.
The sample dataset is uploaded in this repository.
The output of our project is uploaded on our website. The URL is mentioned above.
• Krishna Dalal
• Debashis Das https://github.com/d4deva