Our task will be to separately identify and outline various players on a field
- Ability to identify players in a field
- Ability to separate different classes of players
- Separate player classes via bounding boxes
- Separate and Classify individual frames as bowlers' or batsmen's
- Separate relevant frames of the video
- Separate video into frames which are desired and those which aren't
- Planned Classes of frames
- Run up begin
- Run up end
- Shot begin
- Shot end
- Other
- Planned Classes of frames
- Drop your desired video file into
res/media
folder (create if directory not present). Rename the video to vid.mp4 - Start a http-server in the current directory (i.e. the directory where this readme is located)
- Navigate to the localhost adress pointed by the http server from your web browser, (firefox recommended)
- Load Model using the Load Model button
- Once the model is loaded start the video and click on make predictions, the web browser will start outlining people.
- Note: the detection is done entirely on the client side and depends completely on the machine used.