Focal Point of Learning

This article presents an automatic way to discover the divisions of an image. These will surely have a positive influence on the learning outcome. Our main contribution is to provide a practical method to improve classification performance by selecting the best divisions of interest. After splitting the image into several rectangles, our method exhaustively searches for the rectangles that have the best precision among all possible combinations using MLP (Multi Layer Perceptron) [10]. Then we merge these rectangles into a single rectangle which is called the LFP learning focal point. Experimental studies on images of sardine cans have shown that the proposed methodology improves classification results.