Abstract: Object recognition dependent on deep learning is a significant application in deep learning innovation, which is recognized by its solid capacity of element learning and component portrayal contrasted with the conventional object recognition techniques is recognized. The paper presents the traditional strategies for object recognition and clarifies the connection and distinction between the old-style techniques and the deep learning strategies in object detection. It presents the development of object recognition techniques dependent on deep learning and explains the most commonplace object detection techniques. The paper centres around the framework plan and the functioning standard of the models and looks at the model exhibition in the constant and the accuracy of detection.