基于局部特征的体育运动损伤超声医学图像分割  

Segmentation of Ultrasonic Medical Images Based on Local Features of Sports Injury

作　　者：杨继宏[1] YANG Ji-hong(Anhui University of Science and Technology,Huainan,Anhui 232001,China)

机构地区：[1]安徽理工大学,安徽淮南232001

出　　处：《河北北方学院学报（自然科学版）》2021年第11期52-56,共5页Journal of Hebei North University:Natural Science Edition

摘　　要：目的为了提高体育运动损伤超声医学的检测识别能力,提出基于局部特征的体育运动损伤超声医学图像分割方法,展开对体育运动损伤超声医学检测识别能力的研究。方法建立体育运动损伤超声医学图像分割模型,利用运动骨骼肌分块匹配技术对体育运动损伤超声医学图像的信息进行增强处理,提取图像特征,通过CT亮点特征透射分析体育运动损伤超声医学图像的特征细节,结合深度卷积神经网络训练方法对提取的体育运动损伤点进行体育运动损伤超声医学图像自动探测,实现体育运动损伤超声医学图像分割。结果仿真结果表明,采用该方法进行体育运动损伤超声医学图像分割的精度较高,误差值可以控制在0.103以下,最终达到零误差的效果。结论证实本文方法对体育运动损伤点检测的分辨力和准确性较高,实际应用能力较强。Objective To improve the detection and recognition ability of sports injuries in ultrasonic medicine,a segmentation method of ultrasonic medical images of sports injuries based on local features was proposed,and the research on the detection and recognition capabilities of sports injuries in ultrasonic medicine was carried out.Methods A segmentation model of ultrasonic medical images of sports injuries was established.The skeletal muscle segmentation matching technology was used to enhance the information of ultrasonic medical images of sports injury.After extraction of image features,the feature details of sports injury ultrasonic medical images were analyzed through CT highlight feature transmission.Combined with the training method of deep convolution neural network,the extracted sports injury points were automatically detected in the ultrasonic medical image of sports injury,and the ultrasonic medical image segmentation of sports injury was realized.Results The simulation results showed that the accuracy of the ultrasonic medical image segmentation of sports injuries using this method was high,the error value was controlled below 0.103,and the effect of zero error was finally achieved.Conclusion It is proved that this method has high resolution and accuracy in sports injury point detection,and has strong practical application ability.

关 键 词：局部特征 体育运动 损伤检测 超声医学图像分割 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]