Image Categorization with PCA-SICEF

Image category recognition is important to access visual information on the level of objects and scene types. This paper presents an automatic recognition system of scene and object with PCA-SICEF feature for digital color images. SICEF (scale-invariant color and edge feature) is an extension of the conventional local SIFT (scale-invariant feature transform) feature, which only include edge invariance of local image region but not any color information. So the SIFT feature is not enough for distinguish image categorization especially for scene types, where the color information plays an important role for recognition. Therefore, we improve SIFT by including color feature for local image region, and name it as SICEF feature. However, the dimension of the extracted SICEF feature is so high that we use PCA (principle component analysis) to reduce the dimension, and then, use the PCA-domain SICEF (PCA-SICEF) for image classification. Experimental results show that it is much more efficient by our proposed PCA-SICEF feature than conventional SIFT feature.