论文部分内容阅读
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 is currently a more advanced international cervical cytology, cervical cancer diagnosis of high accuracy. However, during the process of naturally sedimentation liquid-based cell preparation, leukocytes sometimes settle on the substrate with the liquid-based cells, and the microscopic examination is not conducive to the observation of the liquid-based cells by doctors. This not only affects the quality of the preparation, liquid-based cell counts, And even affect the diagnosis. Therefore, a method of improving liquid-based cell image quality based on colorimetry is proposed. The liquid-based cell images are converted from RGB color space to LUV color space. Under the premise of effectively avoiding the influence of the lightness L on the liquid-based cell feature extraction, the chrominance information of the U and V channels is fully utilized. According to the chroma characteristics of liquid- The localization and extraction of liquid-based cells, the use of threshold segmentation and morphological corrosion and expansion method effectively filter out interference information such as leukocytes, so as to effectively improve liquid-based cell image quality. Experiments show that the method is effective and practical.