Real-Time Grab Detection in Videos

This project proposes to make an object detection network using deep learning techniques that predicts in real-time if an RGB video contains a hand grasping an object

This is a fork of the implementation of AlexeyAB's implementation of YoloV4 on darknet. linked here: https://github.com/AlexeyAB/darknet

DEMO VIDEOS

Results on a video from the Grab Dataset Evaluation Set

https://drive.google.com/open?id=1Hs_dKiOXMXJupfJTYxankmKEhLa_U2Q0

Results on a video from the UTGrasp Dataset

https://drive.google.com/open?id=1L9LAARDvmwcIoDtLduz9YnWYOrOeSDeK

Documentation on how to use or further train the object detector will follow