Early Alzheimer's disease identification system and device

Abstract

The invention discloses an early Alzheimer's disease recognition system and device, and relates to a disease recognition technology. The scheme is provided for the problem that in the prior art, the disease recognition operation requirement is huge, fundamental frequency information of a voice signal is extracted, and the moment when speaking starts and the moment when speaking ends in the voice signal are determined; the method comprises the following steps of: generating a corresponding voice segment signal Seg from a voice signal of a time length before and after each speaking starting moment and each speaking ending moment, and acquiring a spectrogram of each voice segment signal Seg; obtaining a final feature of each voice segment signal; carrying out feature vector extraction on the whole section of voice signal; and identifying the feature vector of the whole section of voice signal, and outputting an identification result. The method has the advantages of reducing the parameter quantity and calculation quantity of the whole system, improving the recognition speed, and being especially suitable for general investigation and exploration work of base-level personnel in community activities.

Description

Technical field [0001] The present invention relates to condition recognition techniques, and more particularly to an early Alzheimer's identification system and device. Background technique [0002] Alzheimer's disease is a nervous system disease, which is hidden in the age of 70, so it is also known as old dementia. Alzheimer's patients often appear in life: memory barriers, language barriers, visual spatial disorders, calculation disorders, etc. At present, about 50 million people in the world have an Alzheimer's syndrome, with an average of less than 3 seconds. China's Alzheimer's patients are close to 15 million, and the number is the world's first. With the aging of population aging, Alzheimer is increasingly attracted by people. Although Alzheimer is confirmed once the diagnosis is confirmed, if it can be found in the early stage, it is still discovered as soon as possible and as early as possible, and it is possible to effectively delay the development of the disease. The...

AI Technical Summary

Problems solved by technology

[0003] At present, the diagnosis of Alzheimer's disease through medical means requires professional medical testing equipment, and the process is complicated and the price is high, so it is not suitable for large-scale screening in elderly communities

Currently using traditional machine learning methods such as support vector machines and naive Bayesian algorithms, the recognition effect is not good

At present, deep learning methods are used to directly use the general features of speech without optimizing the features for the special task of Alzheimer's disease recognition. At the same time, the entire audio is directly modeled, which not only leads to a large number of parameters in the model and computational complexity The volume is large, the recognition speed is slow, and the key parts of the entire speech cannot be highlighted

Images