Method and device for detecting eye state in video

Abstract

The invention discloses an eye state detection method in a video. The method includes the steps of: performing face detection on a current video frame of a current user to obtain the first eye area feature of the current video frame; calling a pre-trained SVM classifier; The first eye region feature is sent to the SVM classifier for classification, and the current user's eye state is detected according to the classification result. There is no need to add a special eye neural network to obtain eye region features for recognition as in related technologies, and only a small number of neural networks can be used to detect and acquire the above eye region features and train them. The detection process is no longer complicated, and the eye The feature acquisition efficiency of the inner region is greatly improved, which solves the cumbersome operation of using two neural networks in the current technology, and can more accurately judge the current user's eye state.

Description

technical field [0001] The present disclosure relates to the technical field of face recognition, and in particular to a method and device for detecting eye states in a video. Background technique [0002] The eye is the most important part of the human face, and the most direct sensory organ in contact with foreign objects. The detection method of the eye state is also widely used. For example, motor vehicle drivers will experience fatigue after driving for a long time, and the eye status detection can be used to reflect the fatigue of the driver's eyes and prevent traffic accidents. Parents can detect whether the child is obedient and sleeping through eye state detection. For example, an online learning platform can also be used to remind users to concentrate and study hard through state detection. [0003] At present, most eye state detection methods are eye state detection methods in videos based on feature analysis. The steps are: intercept video face pictures and use ...

AI Technical Summary

Problems solved by technology

This method requires the use of many neural networks to obtain the above-mentioned eye region features, which is cumbersome and time-consuming, and the judgment of the eye state of the video character based on a single frame is not stable enough and the efficiency is low.

Images