In this paper, Recognition and tracking of the object is one of the critical challenges that many of the existing works are focusing on. Mainly the detection of object and tracking is very much popular because of the growing demands in the video surveillance system applications like traffic controlling, medical image processing and satellite image processing applications. This is also of the most powerful algorithms in the field of computer vision, machine learning, Artificial Intelligence based applications. The ultimate objective of these underlying object recognition-based systems is to comprehend the type of images, characteristics, location of the each images in the space and the tracking the movements of each objects while moving. Many objects detection applications used for detection of objects are mainly concentrating on identification of human as it has got lot of attention in the existing research works. This section describes a new approach to recognizing objects using the CNN approach for the detection of the non-living objects as well as human objects. The main objective of this section is providing a framework that can be able to identify both the type of objects like living and non-living. Once they are identified they can be classified and using support vector machine technique, this will be helpful for the theft identification using the surveillance systems.