Intelligent Object detection with classification and localization using Deep Learning

Abstract

 Human eye is blessed to have an advantage of being enabled to differentiae and recognize the objects visually. Computer vision is enabling the same for the computer machines. Applications of human interaction with computer & object detection is enormous. The era of past few years has been of scientific achievements in the area of computer vision for the purpose of object detection. One of the primary goals of exisiting AI technology is the intelligent human computer interaction. Many Object Detection techniques were proposed in the earlier years, was presented as a summary in [1] [2]. Many a researchers have been shifting their trend to utilize diferent multifaceted classification and feature extraction methods just to enhance the correctness of system that does the object detectio. It has always been a challenging task to implement such system in real time applications. The earlier researches have reached to their threshold in the accuracy in the problems in the computer vision. Ever since the Deep learning technique took its way, the improvement in the accuracy of these problems is worth noticing. Image classification to predict the class of the images is among the major problems. Another one, a bit tougher problem is that of the Image Localization. Here in this problem, the images may contain the single object and the expected predicton is the class of the location of the object in the image (it expectes a box surrounding the object known as bounding box). In this article the problem of localization and classification both is taken into picture in the object detection. A bounding box for all the objects in the given image and the class of the object prediction technique has been proposed in the current article.