Shape recognition using a moment algorithm

At first we introduce Hu moment invariant. According to its higher computational complexity and higher time consuming, we introduce an improved Hu moment as we called boundary moment. This paper proves boundary moment invariants computed using the shape boundary only, which tremendously reduce computations compared to traditional Hu moment. Then we do experiments to recognize shape using boundary moment invariants and the result proves that the boundary moment invariants can be used as effective features in shape recognition.