Advance Image Classification System

Advance image classification system focuses on synthetic (e.g non-photographic) & Natural (e.g photographic) images. The classification of images based on semantic description is a challenging and important problem in automatic image identification. An algorithm for natural and synthetic image classification has been developed. Some features are extracted from the raw images data in order to exploit the difference of color pattern and spatial correlation of pixels in natural and synthetic images. These features have poor accuracy if used alone but when combined together forming a more complex and accurate global classifier, their precision can be boosted. Our image classification algorithm will use low-level image features such as Color map, edge map, energy level, Threshold ratio & nearest neighborhood classifier for classifying the image into synthetic and natural.