End-to-end non-contact atrial fibrillation automatic detection system and method based on vPPG signal

Abstract

The invention discloses an end-to-end non-contact atrial fibrillation automatic detection system based on a vPPG signal. The system comprises a data preprocessing module which is used for recording aface video of a to-be-detected user, removing the positions, with large beginning and ending noise interference, of the recorded video, carrying out downsampling, and cutting the video to be with theunified length and size; a pulse wave extraction module which is used for extracting the vPPG signal of the face video through a P3D convolutional neural network; a data denoising module which is usedfor denoising vPPG data based on a neural network of FCN-DN; and an atrial fibrillation detection module which is firstly used for training a model to enable the model to learn to divide an atrial fibrillation segment and a non-atrial fibrillation segment, and then inputting a vPPG signal segment to be detected into the trained atrial fibrillation detection model so as to judge whether the vPPG signal to be detected contains the atrial fibrillation segment or not. A parallel network combining a long-term and short-term memory network containing an attention mechanism and a convolutional network provided by the invention enables the detection of the model to be comprehensive, high in precision and good in effect.

Description

technical field [0001] The invention belongs to the technical field of medical data analysis, and more specifically relates to an automatic detection system and method for atrial fibrillation based on P3D convolution and attention mechanism long-short-term memory network for facial videos. Background technique [0002] Atrial fibrillation is an irregular heartbeat caused by abnormal activity of the heart. In elderly patients over 80 years old, the incidence of atrial fibrillation is usually between 10% and 17%. In 2020, the number of patients with atrial fibrillation worldwide has reached 33 million. The late stage of atrial fibrillation is usually accompanied by cardiovascular diseases such as thrombosis and heart failure. Therefore, early detection and early prevention of atrial fibrillation signals are very important. Atrial fibrillation is diagnosed clinically by doctors analyzing 12-lead ECG (electrocardiogram, electrocardiogram) signals. The diagnostic criteria of e...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the above defects or improvement needs of existing methods, the present invention provides an end-to-end non-contact atrial fibrillation automatic detection system and method based on vPGG signals, which solves the problem of remote non-contact extraction of pulse from face video in the prior art. The occurrence of atrial fibrillation is monitored by waves, the time and space correlation of pulse signals are not considered at the same time, the irregular fragments are not considered to account for a greater weight in model training, resulting in low detection accuracy, and the use of 12-lead ECG detection is costly and difficult to operate. cumbersome, not suitable for daily monitoring, etc.

Images