液基细胞学检查是目前国际上比较先进的一种宫颈癌细胞学检查技术,对宫颈癌的诊断具有较高的准确率。然而在自然沉降式的液基细胞制片过程中,白细胞有时会随液基细胞沉降到基片上,镜检时不利于医生观察液基细胞,不仅会影响制片的质量、液基细胞计数,甚至会影响诊断,因此提出一种基于色彩色度改善液基细胞图像质量的方法。将液基细胞图像从RGB色彩空间转换到LUV色彩空间,在有效避开亮度L对液基细胞特征提取影响的前提下,充分利用U和V通道色度信息,根据液基细胞色度特征,对液基细胞进行定位提取,采用阈值分割以及形态学中的腐蚀和膨胀方法有效滤除白细胞等干扰信息,从而有效改善液基细胞的图像质量。实验验证了该方法的有效性和实用性。

Liquid based cytology detection is currently an advanced technology for cervical cancer screening in the world and it has high accuracy in diagnosis of cervical cancer. However, during the process of liquid based cell preparation by natural sedimentation, the leukocyte will sometimes be found on the basal piece together with the liquid based cells, which is adversely impact for doctors to observe the liquid based cells during microscopic examination. The quality of basal piece, the counting of the liquid based cells and even the diagnosis can also be failed. So a method for improving the imaging quality of the liquid based cells based on color chromaticity is proposed. The image of liquid-based cell is transferred from RGB color space to LUV color space, and under the premise of effectively avoiding the effect of lightness L on the feature extraction of liquid-based cells, the liquid-based cell is located and extracted according to the chromaticity characteristics of the liquid-based cell through the full utilization of the U and V channels chromaticity information. The method of threshold segmentation and the corrosion and expansion methods in morphology are adopted to effectively filter out the interference information such as white blood cells, so as to improve the image quality of liquid-based cell effectively. Experimental results show the effectiveness and practicability of the proposed method.