高噪声显微图像下的微管提取算法研究

Study on Microtubule Extraction Algorithm in High Noise Microimage

为了提取通过冷冻电镜获得的微管图像中的微管目标,本文提出了一种改进的区域生长法。首先根据对微管图像的分析,提出了一种针对高噪声微管图像的复合增强方案,有效的增强了图像对比度。然后使用Otsu算法对微管图像进行粗分割。最后,再用区域生长法提取微管目标。实验结果表明,该方法有较强的抗噪性,同时能够准确提取微管目标。

In order to extract microtubule targets from microtubule images obtained by cryoelectron microscopy,an improved regional growth method was proposed. Firstly, according to the analysis of the microtubule image, a composite enhancement scheme for the high-noise microtubule image was proposed, which effectively enhanced the image contrast. Then the microtubule image is roughly segmented by Otsu algorithm. Finally, regional growth method was used to detect microtubule targets. Experimental results show that this method has strong noise resistance and can accurately extract microtubule targets.