运动员脊椎损伤图像部位准确识别仿真

摘要

The traditional method usually obtains the recognition area of spinal cord injury image and defines the feature vector of image edge,but ignores the optimization for recognition result of spinal cord injury image,which leads to the low accuracy of recognition result.This article puts forward a method for recognizing the image of position of athletes' spinal cord injury based on particle swarm optimization algorithm.The obtained image of spinal cord injury is processed,and the gradient edge feature of image of spinal cord injury is extracted,then the gradient approximative coefficient of normalized edge in spinal cord injury image is multiplied by the mutual information,thus the result is taken as the objective function of image recognition of spinal cord injury.The particle swarm algorithm is introduced to optimize the objective function.Finally,the accurate recognition of spinal cord injury image is realized.Simulation result shows that the proposed method has high accuracy for recognizing the location of athletes' spinal cord injury.%对运动员脊椎损伤图像的部位进行识别,对运动员脊椎损伤的准确诊断起到重要作用.脊椎损伤图像常混有噪声,导致图像局部信息失真,严重影响医生对运动员脊椎损伤诊断结果的准确性.传统方法通常获得脊椎损伤图像的识别区域,定义出图像边缘特征向量,忽略了对损伤图像识别结果的寻优,导致识别结果准确度较低.提出基于粒子群算法的运动员脊椎损伤图像部位识别方法.该方法对获取的脊椎损伤图像进行初步处理,提取脊椎损伤图像梯度边缘特征,将脊椎损伤图像归一化边缘的梯度近似性系数与互信息相乘,作为脊椎损伤图像识别的目标函数,引入粒子群算法对目标函数进行寻优,最终对脊椎损伤图像实现准确识别.实验结果表明,所提方法对运动员脊椎损伤图像部位识别准确度较高,具有一定的实用性.