An Improved Image Signature Based on ICM Combined with SRIC

An improved image signature is proposed and extracted. Firstly, a color information based ICM(Intersecting Cortical Model) image signature is presented. Then a new object feature named SRIC(Silhouette Rotate Interceptive Curve) is described. After these researches, a more distinguishable image signature algorithm based on ICM combined with SRIC is proposed and realized. The Contrast experimentation results show that the distinguishability of the proposed image signature feature is better than the original image signature, and has higher process speed, lower computation complexity. It can be used as a characteristic of video object for identification in practical real-time applications.