计算机图象识别技术测量血小板及其聚集体面积

Area Measurment of Platelets and Aggregate by Computer Pattern Recognition

用计算机图象识别技术测定血小板及其聚集体面积的速度及精度优于目前采用的其它方法。首先从显徽图象中将血小板从背景中分割出来,若是聚集体,则将其分裂,以估算其中的血小板数;最后采用聚合算法得到血小板的分布情况,列表显示结果。本文给出了100例健康人富含血小板血浆中悬浮血小板的大小。平均血小板面积在70例男性和30例女性中分别为3.64±0.87μm^2和3.54±0.66μm^2.血小板轴率为0.84±0.05。

By using computer pattern recog-nition technique, the area measure-ment of platelets is superior to thecurrent other methods in the measure-ment speed and accuracy. The first-step is to extract the platelets andaggregates from background ofthe image, and then to estimate thenumber of platelets in each aggre-gate and measure their areas. Thefinal step is to obtain the member ofplatelets in each aggregate with arecursive clustering algorithm. In thepaper, the platelet sizes of 100 hea-lthy volunteers are given. Mean pla-telet area is 3.64± 0.87μm^2 in 70males and 3.54±10.66μm^2 in 30 fem-ales, respectively. The platelet axisrate is 0.84±10.05.