Invariance analysis and application of the zoomed-and-shrunk image singular value vector

The image singular value vector has been applied extensively to image processing and recognition for its stability and invariance in the transformation of translation,rotation, transposition and mirroration. This paper is to try to analyse how the image singular value vector to change with image size changing by performing matrix operation and experiment; then apply the conclusions to template-matching. Experiments are performed to show that the singular value has robust performance and the strategy based on the singular value is practical and efficient in image processing.