A convolutional Neural Network object detection model implemented using DarkNet framework and trained on MS-COCO dataset of labeled images. Runs at 15-20 FPS on a mid-level GPU. An Android App gives the location of detected objects as audio feedback to the visually challenged.
Darknet is an open source neural network framework written in C and CUDA. It is fast, easy to install, and supports CPU and GPU computation.
Just execute shell script yolo_detect_sh