A new system for segmentation and recognition of scenery images

This paper proposes a new system for segmentation and recognition of scenery images. Most of the conventional methods for image recognition are based on rules or pattern matching. It is true that they have high accuracy. However, the generalization ability is not sufficient and they require heuristic knowledge. In addition, these methods are used on the assumption that images are segmented appropriately. The proposed system alleviates these shortcomings by effective combination of the K-mean-type algorithm, backpropagation algorithm and fuzzy inference neural network