Analysis of Neural Networks for Object Detection using Image Processing Techniques

Abstract

Moving real-world object detection is still a difficult task. Whilst recent research data sets increase the number of training sets and test examples to get closer to real world problems, it is another important question apart from accuracy that detectors can process large data sets in reasonable time. Not only the education instances, but the number of classes is significant. Moving object detection requires finding items in a video sequence frame. An object detection mechanism in either frame is needed in - form of monitoring, or when the object first appears in the film. Different history strategies used in the literature have been simulated during moving object detection. In this study we implement a Gaussian mixture analysis and backward propagation using neural network for object detection.