Content-Based Image Retrieval for Medical Image

In this paper, the SIMPLIcity (Semantics-sensitive Integrate Matching for Picture Libraries), an image retrieval system is introduced. The feature extraction is based on Histogram, color layout and coefficients of wavelet transform. This retrieving system adopts feature database for matching so as to reduce the search space which is especially useful in a larger image database. Retrieval images are selected according to the closest similar measures computed by distance. Under experiments of 600 medical images, this method shows a higher performance of matching. And the similarity defined by l1 for searching is more suitable for gray images.