How to apply this to object detection in a video dataset

[spaul13]

First of all really great project. I can print the training accuracy with the training steps. Can @zonghua94 plz tell me how to measure the detection/inference accuracy?

is it possible to use this project for object detection in a video dataset likewise Lena and UCF101? If it's possible, can you please guide me on how and what to modify to train the network for object detection then finally detect it?