基于标记分水岭的蛋白点检测算法

Protein Spots Detection Algorithm Based on Marker-controlled Watershed

为解决经典分水岭算法的过分割问题,提出了一种基于标记分水岭的蛋白点检测算法。该算法首先对凝胶电泳图像进行预处理去除噪声、背景和条纹,然后在形态梯度图像上利用峰值法找出分水岭的内部标记,最后采用经典的分水岭方法检测出蛋白点,避免了蛋白点过分割现象。通过凝胶电泳图像的验证实验,结果表明本文算法的有效性。

To solve the over-segmentation problem of classical watershed algorithm,this paper proposes a marker-controlled watershed algorithm for protein spots detection. The proposed algorithm first applies image processing technology to eliminate noise,remove streak and correct background,then utilizes peak-value method to find inner mark of watershed,at last adopts classical watershed algorithm to detect spots. Accordingly it can avoid over-segmentation. The experimental results of 2D-gel electrophoresis image show that this new algorithm is effective.