A video analysis device and a video analysis method for elderly people abnormity

Abstract

The invention discloses a video analysis device and a video analysis method for elderly people abnormity. The method comprises the steps of collecting picture materials containing faces of old peopleand marking facial expressions of the old people in pictures, and forming a training data set; Positioning the face in the picture; Calibrating face key points in the positioning area; Constructing aconvolutional neural network; Training the convolutional neural network by utilizing the positioned face part and the key points in the picture, and deploying and training to obtain a model of the convolutional neural network; Performing framing processing on the to-be-tested video, and predicting the expression of the old person in each frame of picture by using the trained neural network model;If the expressions of the elderly people in the continuous frames are predicted to be abnormal (such as anger or impairment) all the time, carrying out some external emotion interventions on the elderly people, such as playing of soothing music, efficient videos or extraction of recorded videos by family members, so that the elderly people are in a positive emotion state.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to a video analysis device and a video analysis method for elderly abnormalities. Background technique [0002] Since the 20th century, my country has entered an aging society. With the increase of age, the mental health of the elderly has become a part worthy of attention. At present, the most commonly used monitoring method is to obtain the information of the elderly through cameras or other sensing devices. real-time situation. For video surveillance, we can directly observe the situation of the elderly through the camera, but we cannot always pay attention to the situation of the camera in real time, so we can further process the obtained video information and report the abnormal situation of the elderly to the guardian. Human facial expressions are one of the main communication methods to express emotions and information. Studies have shown that human facia...

AI Technical Summary

Problems solved by technology

[0004] To sum up, the problems existing in the existing technology are: most of the face recognition is for static images, and the accuracy rate of a few expression recognition for video is relatively low, and the network structure parameters and calculation amount of most expression recognition are very high. Large, the training is time-consuming, so a more accurate and fast training expression recognition method is needed

Images