中文摘要：

针对胸水细胞显微图像的特点提出一种改进的随机Hough变换（MRHT，modified randomized hough transform）检测圆和椭圆的算法。该算法分为两步：利用椭圆的几何性质，确定可能的椭圆中心的位置；在限定区域内，通过多次3点随机抽样，计算椭圆除中心坐标外的其他3个参数。研究表明，该算法可以同时检测多个圆和椭圆，可以从胸水细胞显微图像的复杂背景中较为准确地提取圆形和椭圆形细胞。实验结果表明，该算法有较高的检测效率，检测精度和较强的鲁棒性。

英文摘要：

According to the properties of pleural effusions cellular microscopic images, this paper presents an algorithm for the detection of ellipses（including circles） in images with modified randomized hough transform（MRHT）. The algorithm consists of two steps：first,locate the centers of ellipses using the ellipse＇s geometrical properties;second,calculate ellipse＇s three parameters besides the center coordinates using multiple three-points random samples in restricted areas. This algorithm may be used in multiple circles and ellipses real-time detection and circular and elliptic cells extraction from pleural effusions cellular microscopic images with complicated backgrounds. Experimental results show that this algorithm is efficient, accurate and robust.