Home

Welcome to the drv_package wiki!

The deep robot vision package aims to utilize state-of-the-art deep learning methods in the field of computer vision to enhance the capability of robot vision system as well as promote the performance of robot vision in practical household usage. The project is a part of Northeastern University's household companion robot project and serves as the main vision solution.

With this package, the robot can locate and fetch objects in real scenes, recognize people and interact with people. The object detection module integrated the Faster-RCNN https://github.com/rbgirshick/py-faster-rcnn and the object tracking function was based on GOTURN https://github.com/davheld/GOTURN, which leads to excellent robustness and fast process speed.