Survey of The Problem of Object Detection In Real Images

Object detection and recognition are important problems in computer vision. Since these problems are meta-heuristic, despite a lot of research, practically usable, intelligent, real-time, and dynamic object detection/recognition methods are still unavailable. The accuracy level of any algorithm or even Google glass project is below 16% for over 22,000 object categories. With this accuracy, it’s practically unusable. This paper reviews the various aspects of object detection and the challenges involved. The aspects addressed are feature types, learning model, object templates, matching schemes, and boosting methods. Most current research works are highlighted and discussed. Decision making tips are included with extensive discussion of the merits and demerits of each scheme. The survey presented in this paper can be useful as a quick overview and a beginner’s guide for the object detection field. Based on the study presented here, researchers can choose a framework suitable for their own specific object detection problems and further optimize the chosen framework for better accuracy in object detection.