Mobile terminal face attribute recognition method based on deep learning in home environment

Abstract

The invention discloses a mobile terminal face attribute recognition method based on deep learning in a home environment. The method is characterized in that a data acquisition end, a transmission device taking a router and an internet cloud end as cores, and an identification terminal taking an Android mobile phone and a face attribute identification APP as cores form an identification system; the Android mobile phone APP acquires remote video data acquired by the data acquisition terminal; an LBP face detector in an OpenCV 3.4.1 library is used for carrying out face detection on the remote video; and performing face attribute recognition on the detected face by using a lightweight mini _ Xaction deep learning recognition model provided by the face attribute recognition terminal, and giving an alarm prompt in time for abnormal expressions and strangers. According to the invention, the lightweight identification model is transplanted to the mobile terminal equipment, and the mobile terminal equipment completes the identification of the face attribute, thereby avoiding the channel congestion caused by the access of a plurality of users to the server, and reducing the operation costof the server. The method is high in recognition accuracy and high in speed, and can meet the requirements of the development of smart home for face attribute recognition.

Description

technical field [0001] The invention relates to a face attribute recognition method at a mobile terminal, belonging to the technical field of pattern recognition. Background technique [0002] With the advent of the information age, smart home has come into life and has a considerable scale of development. At present, home appliance control, lighting control, telephone remote control, indoor and outdoor remote control, environmental monitoring, HVAC control and infrared forwarding, etc. This technology has been applied to the smart home system, which makes the smart home system more and more perfect. In the smart home system, there is relatively little analysis of face attributes. The actual home environment has certain requirements for the recognition of face attributes such as face recognition, expression recognition, gender recognition, and age recognition. For example, among family members, it is necessary to identify the elderly and children's facial expressions, so th...

AI Technical Summary

Problems solved by technology

However, due to the limited processing power of the mobile terminal, it is necessary to choose a model with high recognition rate and lightweight

Research on lightweight recognition models has also achieved some results, such as the MobileNet network launched by Google. According to different network hyperparameters, the trained model is about 2MB to 17MB, and it can be deployed on Android phones to achieve garbage classification. , urban management case target recognition and other functions, although the MobileNet network reduces the amount of parameters to a certain extent, and the smallest model is only 2MB, but it is not very ideal for facial expression recognition in facial attribute recognition, because facial expressions are constantly changing, And the changes are subtle, and it is necessary to accurately judge the expression of each frame of the video, so it is also necessary to find a model that is lighter and has a higher recognition rate

Images