Image Retrieval Technique Using Rearranged Freeman Chain Code

An Image retrieval method based on the statistical representation of the freeman chain code is proposed. Freeman chain code is used for representing the contour of an image object. We do the feature extraction by smoothing the object for noise reduction, convert to binary image by thresholding, then the boundary of the object is detected and the first difference of the chain code is computed instead of the chain code itself. The histogram count of the chain code is computed and we rearrange the histogram count in ascending order. After computing the rearrange histogram count we compute the mean and variance of the rearrange chain code. Based on these feature we retrieved image from the database. We used the database that contains approximately one thousand images. The proposed method usually shows better results.