针对超声心动图像质量差、噪声多，传统卷积神经网络架构对超声心动图像的学习能力有限、表达不充分的缺点，提出了一种基于标准切面识别的房间隔缺损（Atrial Septal Defect）智能辅助诊断模型。该模型通过对超声心动图像切面识别，充分融合其不同切面的语义特征，使得诊断的准确率得到明显提升。此外，还对其进行双边滤波保边去噪，并基于此模型搭建房间隔缺损智能辅助诊断系统（简称ASD辅助诊断系统）。结果表明，该ASD辅助诊断系统的准确率高达97.8%，且与传统卷积神经网络相比大大降低了假阴性率。
For echocardiography being noisy and fuzzy, the traditional convolutional neural network architecture has limited learning ability and feature expression, an Atrial Septal Defect intelligent auxiliary diagnostic model architecture based on feature view classification is proposed. The model architecture integrates the semantic characteristics of several views, which makes the accuracy of diagnosis significantly improved. In addition, aimed to de-noise and preserve edges, bilateral filtering algorithm is performed. Furthermore, an atrial septal defect intelligent auxiliary diagnostic system (ASD auxiliary diagnostic system) is built based on proposed model. The results show that the accuracy of the ASD auxiliary diagnostic system reach to 97.8%, and the false negative rate is greatly reduced compared with the traditional convolutional neural network architecture.