Gray-scale skeletonization of near-infrared vein patterns using the improved watershed algorithm in vein pattern biometrics

In order to analyze the shape of the vein pattern rapidly and accurately and to reduce the computing errors of the subsequent processing stages, a novel approach is proposed to extract the skeletons of vein patterns directly from gray-scale samples, which combines with the watershed algorithm and region merging. The rule of the region merging is composed of measurements of the pixel´s value and distribution. This approach can handle the over segmentation problem resulted from classical watershed algorithm. Besides by doing the region merging processing while marking watershed, this new algorithm can speed up the process. This method is just 0.212 times of the traditional methods which combine watershed transformation and region merging simply. The experiment result getting from this new method has also compared with several other methods and shown its outstanding performance.