An Inter-Comparative Survey on State-of-the-Art Detectors—R-CNN, YOLO, and SSD

In recent years, breakthrough enhancements in computer hardware and supercomputers made object detection a significant topic of research. Accurate object detection models are computationally expensive and are inefficient on simpler and limited configuration settings while faster models achieve real-time speed, work well on simpler configurations but fail to be accurate. There is always a trade-off between speed and accuracy. There is no clear-cut answer on which detector performs the best. The user will have to make a choice based on the requirement. This paper aims at analyzing numerous CNN-based object detection algorithms—R-CNN, Fast R-CNN, Faster R-CNN, You Only Look Once (YOLO), and Single Shot MutliBox Detector (SSD)—and make comparisons concerning performance, precision and speed and state as to which algorithm performs better under certain constraints. This enables the user to pick an object detector of his/her choice that better addresses the demands of an application.