Deep Learning techniques for Object Detection in Images: Exploratory Study

Abstract

The aim of computer vision is to visualize the data available in images or videos. Visualizing such objects by machine is known as object detection in images. It is one of the most recent research areas. It is used to detect and recognize various objects present in an image. Object detection can be used in detecting tumors in the medical field, animal detection in agriculture, detecting whether a person is carrying any weapon or not, in e-commerce sites, image-based search engines, self-driving cars, and many more. Convolutional Neural Network (ConvNet or CNN) is a dominant technology that is used in detecting objects because it makes the process fast. The area is quite popular among researchers as well as industry person. There are various techniques used for object detection like CNN, R-CNN, Fast RCNN, and Faster RCNN. In this paper, a comparative study of all these techniques is done to show which one is better than other.