Object Classification
Maheshkumar H. Kolekar · 2018
Object classification detects moving objects in a video sequence and classifies them into categories such as humans, vehicles, birds, clouds, or animals. Although the problem of classifying moving objects is very trivial for human beings, it is very challenging for a computer algorithm to do the same with human-level accuracy. Shape-based features for classifying the object is one of the major challenges, because of the changes in illumination, scale, pose, and camera position in the scene. Motion-based object classification is a very important step in visual surveillance systems. In 2001, Paul Viola and Michael Jones proposed an algorithm for real-time object detection. In particular, convolutional neural networks (CNN) have shown very good results in image classification, image segmentation, and computer vision problems. A regional convolutional neural network (RCNN) is a state-of-the-art visual object detection system that combines bottom-up region proposals with rich features computed by a convolutional neural network.