Patient activity analysis system and method for behavior recognition based on wearable device

Abstract

The embodiment of the invention discloses a patient activity analysis system and method for behavior recognition based on a wearable device. The system comprises an acceleration sensor, a patient client, a data collection unit, a database server, a computing server and a doctor client, wherein the acceleration sensor is used for acquiring patient behavior information data; the patient client is used for acquiring patient personal information and performing feature selection on the obtained patient behavior information; the data collection unit is used for collecting the acquired patient behavior information and personal information and performing extraction and computation processing on the patient behavior information; the database server is used for storing a patient personal informationand daily behavior information record; the computing server is used for performing computational recognition processing on patient behavior information data; the doctor client is used for querying the patient behavior record according to a patient visit number. In the embodiment of the invention, the problem that a doctor knows patient daily activities difficultly can be solved, and the doctor can make evaluation according to the patient daily activities; a patient can conveniently view personal history activity data and get advices from their doctors.

Description

technical field [0001] The invention relates to the technical fields of the Internet and pattern recognition, mobile medical care, and intelligent judgment, and in particular to a patient activity analysis system and method thereof based on wearable devices for behavior recognition. Background technique [0002] There has always been a problem in the hospital. Doctors and nurses often cannot accurately understand the activities of each patient before and after surgery. They can only get a general understanding of the daily activity time of the patient by asking the patient. There is no quantitative way to show the specific daily behavior of the patient. and how long. On the other hand, when patients are hospitalized, they dare not increase the amount of exercise without permission from the doctor, which is not very beneficial for postoperative recovery. Therefore, there is a need for a software that can accurately display the patient's daily exercise time to solve this prob...

AI Technical Summary

Problems solved by technology

First, the daily behavior characteristics of patients with different diseases are very different, and it is difficult to accurately identify them through a general model

Second, for patients, it is more difficult to collect training sets, and it is often difficult to have sufficient training data, which will lead to poor generalization ability of the model

[0005] Among the above-mentioned problems, any of them will lead to low accuracy of patient behavior recognition

In order to provide patients and doctors with accurate exercise duration information, and to give accurate evaluations based on the patient's daily exercise duration to help them adjust their daily activity time, we need to train a personalized behavior recognition model for each patient, but there is currently no A complete solution to the problem

Images