A summary of the object detection model
I stepped away from the first iteration of this project last year when I realized I was going to have to annotate thousands of images to create a robust player/ball detection model. In the next few months, I'll design and build such a modelling framework. This post outlines the methods and services which will be used for the framework.
YOLOv5
The current object detection model is YOLOv5. Recall some of the results from the latest model iteration were: overall: 91.0% P, 85.5% R, 85.8 mAP50 front-player: 99.5% P, 99.6% R, 99.5 mAP50 back-player: 99.5% P, 99.4% R, 99.3 mAP50 tennis-ball: 74.1% P, 57.0% R, 58.7 mAP50
Video Example
For a short example of the inference, click here.
When does the model tend to not classify an object?
For players, xxx.
For tennis balls, xxx.
When does the model tend to mis-classify an object?
For players, xxx.
For tennis balls, xxx.
Post-Processing of Tennis Players
The model performs strongly, so a simple imputation strategy using interpolation has worked so far. Interpolation is done through Python's xxx module. Here are the specs: - window: - function: mean - na length: