Face attribute recognition method and electronic equipment
Through the face attribute recognition method based on multitasking classification, the problem of low facial attribute recognition accuracy in the prior art is solved. By optimizing model parameters and classification process, the recognition accuracy and feasibility of model deployment are improved.
CN119672772APending Publication Date: 2025-03-21QINGDAO HI-IMAGE TECH CO LTD
0 Cites 0 Cited by
Patent Information
- Application Number
- CN202311221788.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-09-21
- Publication Date
- 2025-03-21
AI Technical Summary
Technical Problem
In the prior art, the face attribute recognition task has a low recognition accuracy due to the shared network part of the shared network model.
Method used
The face attribute recognition method based on multi-task classification is adopted. By obtaining the face image to be processed and the pre-set face attribute set, preprocessing and classification are performed, training samples are determined, and the target face attribute recognition model is obtained by adjusting the model parameters to optimize the loss value.
Benefits of technology
It improves the accuracy of model training and the accuracy of facial attribute recognition, reduces the computing power and bandwidth of the model, and facilitates model deployment in practical applications.
✦ Generated by Eureka AI based on patent content.
Abstract
The invention discloses a face attribute recognition method and electronic equipment, and the method comprises the steps: carrying out the preprocessing of a to-be-processed face image, and obtaining a feature data set of each to-be-processed face image under a preset face attribute set; classifying the face attribute set according to the type of each face attribute included in the face attribute set and the feature data set; determining a classified face attribute set as a training sample; inputting the training samples and pre-labeled attribute tags corresponding to the training samples to the initial face attribute recognition model to obtain a prediction result; determining a loss value according to each prediction result, a pre-labeled attribute tag and a set loss function; adjusting model parameters of the initial face attribute recognition model according to the loss value until the determined loss value is within a preset range, and obtaining a target face attribute recognition model; and performing face attribute recognition by using the target face attribute recognition model. And the identification accuracy is improved.
Need to check novelty before this filing date? Find Prior Art