Saliency Guided Faster R-CNN for object detection and Recognition

Abstract

The major purpose of our project to dectect object using faster R-CNN to develop a technique for programmed object finding and counting on throughways. Our method does not rely on foundation; rather, it employs a channel through which we identify and count the cars, record a video or a photograph, and then make a decision to give the total number of object. The purpose of this approach is to create a superior object detector structure inside the project for smoother detection of object flow and to increase a overall efficiency.