摘要
活体细胞图像中细胞与背景对比度低,背景中存在密集分布的颗粒状噪声,要有效分割图像并提取细胞特征,预处理就显得尤为重要。针对这一问题将一种改进的均值偏移滤波算法应用到对活体细胞图像背景去除中,在去除图像背景的同时保留了细胞的形貌,并且探讨了核函数带宽选择对均值偏移滤波算法处理效果的影响。
In living cells images, the contrast ratio between the cells and the background is low and there are dense granular noises in the background. In order to segmentate the images and extract the features of the cells effectively,it is important to preprocess the images before the segmentation. To the questions above,a improved mean shift filetering is applied to elimitate the background noises on the promise of keeping the morphology of cells unchange. It laso has a discussion about the effects of kernel bandwidth on the output of meanshift filtering.
出处
《长春理工大学学报（自然科学版）》
2013年第5期150-153,共4页
Journal of Changchun University of Science and Technology(Natural Science Edition)
基金
吉林省自然科学基金资助项目(201215136)
