Similarity Threshold Determination Method, Device, Equipment, and Storage Medium
By calculating the similarity of face pictures and determining the target similarity threshold, the problems of large classification error and low accuracy caused by artificial setting of face verification threshold in the prior art are solved, and a higher classification accuracy and face verification accuracy are achieved.
CN114120383BActive Publication Date: 2025-06-24SF TECH CO LTD
2 Cites 0 Cited by
Patent Information
- Application Number
- CN202010882569.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2020-08-28
- Publication Date
- 2025-06-24
- Estimated Expiration
- 2040-08-28
AI Technical Summary
Technical Problem
In the prior art, the threshold used for classification in face verification is usually set by technicians based on past experience, resulting in large classification errors and low classification accuracy.
Method used
By obtaining multiple face pictures, the similarity of each pair of pictures is calculated, and these similarities are compared with the preset candidate threshold, the evaluation index of each candidate threshold is calculated, and the candidate similarity threshold that reaches the index threshold is determined as the target similarity threshold.
Benefits of technology
This method can be better applied to the verification of all face pictures, reduce classification errors, improve classification accuracy, and thus improve the accuracy of face verification.
✦ Generated by Eureka AI based on patent content.
Abstract
The present application discloses a method, apparatus, device, and storage medium for determining a similarity threshold. The method includes: obtaining multiple face images; combining any two images among the multiple face images to obtain multiple image sets, and calculating the similarity between the two images in each image set; comparing each similarity with each candidate similarity threshold in a preset candidate threshold set to obtain corresponding comparison results; calculating an evaluation index for each candidate similarity threshold according to the comparison results; and determining the candidate similarity threshold whose evaluation index reaches the index threshold as the target similarity threshold. The target similarity threshold calculated in the present application can be better applied to the verification of all face images, reduce the classification error, improve the classification accuracy rate, and thus improve the face verification accuracy rate, compared with the threshold set according to experience in the prior art.
Need to check novelty before this filing date? Find Prior Art