提出一种快速全自动超声子宫图像节育环物体识别算法,算法包括全自动超声子宫图像分割和特定模式识别框架部分.基于719幅超声子宫图像开展算法性能验证.实验结果表明,该算法无需人工干预,平均耗时527ms/幅,有环图像识别准确率达81.1%,无环图像识别准确率达94.7%.
A fast automatic IUD (intra-uterine device) recognition algorithm for the ultrasound uterus images was proposed. Components of the algorithm inclucled an automatic ultrasound uterus image segmentation algorithm and a specific pattern recognition framework. The performance of the algorithm was demonstrated by using 719 ultrasound uterus images, and the experimental results show that the average computation time is 527 milliseconds per frame. The recognition accuracy of the uterus image with or without IUD is 81.1% or 94. 7% respectively.