中文摘要：

针对灰度共生矩阵对指纹图像分割过程中人工选取阈值不精确、繁琐等缺点,提出了一种采用自适应阈值分割的灰度共生矩阵的指纹图像分割算法。首先,用整幅指纹图像的对比度方差值的均值Mv作阈值对图像进行初分割;然后,不断调整Mv,通过试验验证当指纹区域对比度方差值的均值Pv与Mv的比值在一个特定的区间（即Pv/Mv∈[1.5,2]）时,才能获得最好的分割效果,由此获得灰度共生矩阵的自适应分割阈值Mv,从而精确地分割出指纹图像的有效区域。试验结果表明,相比于已有的分割算法,该算法在分割错误率和耗时方面均较优,并且分割更准确。

英文摘要：

In the process of fingerprint image segmentation,matrix artificial selection of the threshold is not accurate,tedious.A fingerprint image segmentation algorithm is proposed by improved gray level co-occurrence matrix.Firstly,Mvis the mean value of the contrast variance for fingerprint images,Pvis the mean value of the contrast variance between the fingerprint region.Mvis adopted as the image segmentation threshold.Then,Mvis adjusted continuously,when the Pv/Mvis between 1.5and 2the best segmentation results is obtained through experiment.The adaptive segmentation threshold of gray level co-occurrence matrix is obtained.Experimental results show that this algorithm is superior to the segmentation error rate,time consuming and more accurate compared with the existing segmentation algorithm.