杜兴氏肌营养不良(DMD)是一种严重的儿童腿部神经肌肉罕见病。传统的诊断和检测方案一般为有创手段,会带给患儿极大的痛苦。基于受试者的磁共振图像(MRI),采用计算机辅助检测手段探索了有效的无创检测方法。实验分别选用sym4和db4两种小波基函数,对患儿组和健康对照组的MRI进行三种尺度的小波分解,从所得的分解图像中提取12个纹理特征参数,并利用人工神经网络(ANN)算法对图像参数进行分类识别。结果显示:在受试者的两类MRI加权图像(T1和T2)中,T1图像能更好地区分患儿与健康儿童;利用db4函数对图像进行小波分解,其效果略优于sym4函数,且在三种小波分解尺度中,以二层分解最优;利用ANN算法对图像进行分类识别,其灵敏度、特异度和准确率分别高达98.5%、97.3%和97.9%。该处理方法有望为临床提供客观有效的辅助诊断手段,可作为DMD疾病无创检测的尝试探索。 Duchenne muscular dystrophy (DMD) is a severe disorder of leg neuromuscular disease in children. Traditional diagnostic and testing programs are generally invasive means that will bring great pain to children. Based on the subject’s magnetic resonance imaging (MRI), an effective noninvasive detection method was explored using computer-assisted detection. Three kinds of wavelet decompositions were performed on MRI in children and healthy controls by using two kinds of wavelet basis functions, sym4 and db4, respectively. Twelve texture parameters were extracted from the decomposed images. The artificial neural network (ANN) ) Algorithm to classify and identify image parameters. The results show that T1 images can better distinguish children from healthy children in two types of MRI-weighted images (T1 and T2) of subjects; using db4 function to decompose the images is better than sym4, In the three wavelet decomposition scales, the decomposition is the best in the second layer. The classification, recognition and classification of the images by ANN algorithm are as high as 98.5%, 97.3% and 97.9% respectively. The treatment method is expected to provide an objective and effective auxiliary diagnostic tool for clinical trials, which can be used as a non-invasive test for exploring DMD diseases.