Image retrieval based on spatail feature
Tingting Shi, Yan Li · Journal of Computer Applications · 2009
A novel image feature descriptor SCH based on spatial feature was proposed.It used MCVAE algorithm based on the color vector angle and Euclid distance to detect original color image edge,and a new greatest and smallest component color invariant model to realize the original image quantization.Furthermore,for the edge pixels,the distribution of the color pairs at an edge was extracted by edge correlation matrix.Meanwhile,for the non-edge pixels,the local color distribution of pixels was extracted by color histogram.Then,a new sin similarity metric was used to measure the similarity between the image features.Finally a content-based image retrieval prototype system called SttImageRetrieval was developed with VC,with establishment of a comprehensive database of images called IMAGEDB using Oracle 9i database.Experimental results show that our algorithm outperforms the methods of color statistical feature for image indexing and retrieval.