Neural Network Based Context Dependent Recognition Approach for Low Indexed Images

摘要

An integral approach to explore the resolution of the image variation is presented in this paper. The natural images are observed to be variant with respect to the resolution variation in all directions. The resolutional variations are used as feature information to neural network for the recognition of images with image content variation. The paper presents the texture variation with shape as reference feature for the image recognition with neural network approach. The approach of textural information lists neural network approach is found to be more accurate in estimation accuracy as compared to the conventional content based neural approach