摘要
针对医学超声图像的特点及医疗诊断实时性的要求,提出了一种基于整数DCT变换的超声膀胱图像边缘检测算法;该算法首先对膀胱超声图像进行预处理,利用整数DCT变换基在压缩域内求取超声图像的梯度,再结合对预处理后的图像进行动态确定阈值的方式进行边缘检测,最终得到膀胱的边界,简化了运算,节省了时间;实验表明,该算法在满足临床医疗诊断实时性的同时获得了较好的质量和边缘完整的膀胱边界。
According to the characteristics of medical ultrasound image and the demand that mediclal diagnosis must real time,this paper preseats a method that edge detection of ultrasound bladder image based on integer DCT transform.First,the algorithm carries on to the bladder ultrasound image preprocessing,and get the gradient from ultrasound image in the compressed domain using integer DCT transform radix.Then,it combines with the image preprocessed dynamic threshold determination method for edge detection.Finally,get the bladder boundary,simplifies the operation and save time at the same time.Experiment show that,the algorithm has better quality edge and complete bladder boundary while meeting the clinical medical diagnosis in real time.
出处
《计算机测量与控制》
2015年第2期624-626,629,共4页
Computer Measurement &Control
基金
国家自然科学基金资助项目(61202044)
四川省科技创新苗子工程项目(20132022)
关键词
超声图像
边缘检测
整数DCT变换
梯度
动态阈值
ultrasound image
edge detection
integer DCT transform
gradient
dynamic threshold
