CPC A61B 5/7267 (2013.01) [A61B 5/4803 (2013.01); A61B 5/7257 (2013.01); G10L 25/15 (2013.01); G10L 25/18 (2013.01); G10L 25/21 (2013.01); G10L 25/24 (2013.01); G10L 25/30 (2013.01); G10L 25/45 (2013.01); G10L 25/66 (2013.01); G10L 25/90 (2013.01); G10L 25/93 (2013.01)] | 7 Claims |
1. A difficult airway evaluation method based on a machine learning voice technology, comprising:
step (1), acquiring voice data of a patient;
step (2), carrying out feature extraction on the voice data, obtaining a pitch period of pronunciations, and acquiring a voiced sound feature and unvoiced sound features based on the pitch period of pronunciations; wherein the voiced sound feature is a formant, and the unvoiced sound features are short-time energy and a short-time average zero crossing ratio; and
step (3), constructing a difficult airway evaluation classifier based on the machine learning voice technology, analyzing the received voiced sound feature and unvoiced sound features by the trained difficult airway evaluation classifier, and carrying out scoring on the severity of a difficult airway to obtain an evaluation result of the difficult airway;
wherein the difficult airway evaluation classifier based on the machine learning voice technology in the step (3) is a fully connected neural network, and the fully connected neural network comprises one input layer, three hidden layers, and one output layer; and the fully connected neural network has an initial learning rate of 1 and an attenuation rate of 0.0001, a Rectified Linear Unit (ReLU) activation function is adopted, and an optimization function is Stochastic Gradient Descent (SGD).
|