【摘要】 目的:针对采用普通的图像分割方法对照度不均的皮肤图像分割效果差的情况,提出了一种改进的图像分割方法。方法:首先将图像去背景,将图像划分为若干子块,接着采用OTSU和迭代阈值算法确定每个子图像块的最优阈值,利用此阈值完成对每个子块的分割,最后将子块合并,同时消除图像中噪声的影响。结果:采用改进的方法很好地实现了对照度不均皮肤图像的分割,其分割成功率可达95.5%。结论:在MATLAB平台上,改进方法能有效地实现对照度不均的皮肤表面图像的分割,并且在定性和定量评价中取得了良好的实验效果。更多还原

【Abstract】 Objective: Because of unequal illumination,noise and other factors in the skin image,this paper develops an improved segmentation algorithm to improve the poor effects of skin image by ordinary segmentation algorithm.Methods: Firstly,subtract the background from the image,secondly,divide the preprocessing image into several blocks,and then segment each sub-image block by the optimal threshold from the help of Otsu method and iterative threshold,at last eliminate the noise in the image.Results: The new method works well to segment the uneven illumination skin images,and the success rate of segmentation is 95.5%.Conclusion: With the help of MATLAB platform,the improved method could effectively realize the uneven illumination of the skin surface image segmentation,and has obtained good experimental results no matter in qualitative or quantitative evaluation.更多还原