Hoop or ball? customized object detection
In this project I use the tensorflow object detection API to build a model to detect hoop and ball from pictures. I give lots of credit to this tutorial, which gives step-by-step instructions to build a model to detect mac-n-cheese. In my case, I only used 9 images for training and 4 images for testing while still getting some reasonable results (such as the following image). One can easily get a more accurate model by training with more samples.
