A novel image retrieval algorithm combining global statistical feature and local bitmap feature was proposed. Firstly

compute the mean value and the stand deviation of every image

capturing the global statistical feature. Then divided every image into 4×4 image blocks without overlapping

and compute the mean value of every image block according to block truncation code

if the mean value is larger than the one of the whole image

then the image block is set as “1”

otherwise

the image block is set as “0”

so capturing the image bitmap features. Finally

do retrieval by integrating the features normalized to get the best similar matching function. Experiment result indicates that: the retrieval performance of integrating two global and local features is more prior than using single feature; Comparing with other similar schemes

the algorithm is robust

and retrieval efficiency is higher.