Shape is a very important visual and semantic feature used to depict image

and it can be revealed by image pixels’ regional distribution. This paper proposes a region-based shape representation

a new “density distribution feature (DDF)”. After shape center orientation and region partition

two M dimensional feature vectors are got. The first feature vector represents the relatively density of object pixels within each sub-image. And the second represents the difference of relatively density in the direction of radial coordinates. When matching the similarity

we first used the Gaussian model to normalize the two dimensional feature vectors. Then we integrated them to calculate similarity distance. The experiments results showed that this shape feature can depict the image well and is invariant to translation

scale and rotation. The paper also evaluated the effectiveness of the proposed descriptor with respect to Moment Invariants.