Scene Understanding Using Back Propagation by Neural Network

In this paper, we proposed an efficient method to address the problem of scene understanding that is based on neural network (NN) and image segmentation. We utilized a multilayer perceptron (MLP) to train the network and features are extracted using pixels in the RGB color space. In this work, object samples in images with varying lighting conditions are used to obtain a wide object color distribution. The training data is generated from positive and negative training patterns in the color planes. Subsequently, training set is fed to an MLP, trained by the back propagation algorithm using these object samples. We apply the above mentioned NN-based object classifier to the test image which is applied to image segmentation and corresponding to the pixel level of the object in test image particular object is determined.