Single-sample face recognition method based on sparse representation of hybrid extension block dictionary
A hybrid expansion and sparse representation technology, which is applied in character and pattern recognition, unstructured text data retrieval, image data processing, etc., can solve the problems of lack of discrimination, lack of consistency, poor face recognition effect, etc., to achieve Eliminate redundant information between pixels, reduce the atomic dimension of the dictionary, and facilitate accurate recognition
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0089] Example 1: Experiment on CAS-PEAL database:
[0090] First as figure 1 shown building blocks, figure 1 Take B=4 in it:
[0091] The CAS-PEAL face database contains 1040 types of people with a total of 99,594 face images (including 595 males and 445 females). All images were collected in a special collection environment, covering four main changing conditions of posture, expression, ornaments and lighting, and some face images had changes in background, distance and time span. The present invention selects 9031 images for experiments, and some sample images such as image 3 As shown, the image size is 120×100 pixels.
[0092] The standard datasets of target objects on the CAS-PEAL database, general datasets of non-target objects, occlusion datasets, and intra-class variation datasets are designed as follows:
[0093] (1) The general dataset of non-target objects includes 180 categories of people with changes in illumination and 80 categories of people with changes i...
Embodiment 2
[0107] Example 2: Experiment on AR database:
[0108] Same as Example 1 first as figure 1 shown building blocks, figure 1 Take B=4 in it:
[0109] The AR face database contains 126 types of people (56 females, 70 males), with a total of more than 4000 front-aligned faces. Each type of person is shot in two stages, each stage contains 13 images, including 4 images of illumination changes, 3 images of expression changes, 3 images of glasses occlusion, and 3 images of scarf occlusion. In the present invention, 100 types of people are selected for experiments, and the images are cropped and normalized, and the size after cropping is 120×100 pixels.
[0110] The standard datasets for target objects on the AR database, general datasets for non-target objects, occlusion datasets, and intra-class variation datasets are designed as follows:
[0111] (1) The standard sample set of the target object consists of any 30 types of people in the AR database, and the first non-interference...
Embodiment 3
[0121] Example 3: Experiment on LFW database:
[0122] Same as Example 1 first as figure 1 shown building blocks, figure 1 Take B=4; LFW (Labeled Faces in the Wild database) database is a real face database collected from the Internet, with a total of 13,233 face images of 5,749 categories, including illumination, expression, posture, occlusion, age, race, etc. This kind of mixed interference is more challenging for accurate face recognition. The present invention selects persons containing more than 10 images for identification, and obtains 158 types of persons. In order to facilitate the experiment, 10 images of each type of person are selected here for 1580 samples for experiments. Some samples such as Image 6 As shown, the image size is 120×100 pixels.
[0123] Since the samples in the LFW database are collected from real environments with various mixed disturbances, typical occlusion datasets and intra-class variation datasets cannot be constructed on the LFW databa...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com