Unlock instant, AI-driven research and patent intelligence for your innovation.
A face recognition method and device based on fusion of key point and region features
What is Al technical title?
Al technical title is built by PatSnap Al team. It summarizes the technical point description of the patent document.
A technology of regional features and face recognition, which is applied in character and pattern recognition, instruments, computing, etc., can solve problems such as low classification ability of classifiers, insufficient accuracy and robustness of face recognition, and incompleteness
Active Publication Date: 2021-05-07
北京云和互动信息技术有限公司
View PDF10 Cites 0 Cited by
Summary
Abstract
Description
Claims
Application Information
AI Technical Summary
This helps you quickly interpret patents by identifying the three key elements:
Problems solved by technology
Method used
Benefits of technology
Problems solved by technology
[0004] To this end, the embodiment of the present invention provides a face recognition method and device based on key point region feature fusion to solve the problem of low recognition accuracy of existing face recognition methods and the lack of features of face recognition methods based on convolutional neural networks. The extraction is incomplete and incomplete, which leads to the problem of low classification ability of the classifier and insufficient accuracy and robustness of face recognition
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more
Image
Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
Click on the blue label to locate the original text in one second.
Reading with bidirectional positioning of images and text.
Smart Image
Examples
Experimental program
Comparison scheme
Effect test
Embodiment 1
[0037] Such as figure 1 As shown, the embodiment of the present invention proposes a face recognition method based on key point region feature fusion, the method comprising:
[0038] S100. Perform data preprocessing on a training set or a test set. In this embodiment, the training set uses the MS-Celeb-1M data set.
[0039]MSR IRC is currently one of the largest and highest-level image recognition competitions in the world. MS-Celeb-1M is the public data set of this competition. From 1M celebrities, according to their popularity, select 100K individuals, and then use The search engine, according to the selected 100K individuals, each person searches about 100 pictures, a total of 100K×100=10M pictures. The training set includes 1,000 celebrities. These 1,000 celebrities are randomly selected from 1M celebrities, and after labeling, each celebrity has about 20 pictures, which cannot be found on the Internet.
[0040] Specifically, data preprocessing on the training set or te...
Embodiment 2
[0054] Corresponding to Embodiment 1 above, this embodiment proposes a face recognition device based on feature fusion of key point regions, which includes:
[0055] The preprocessing module is used to perform data preprocessing on the training set or test set.
[0056] The network building module is used to build a convolutional neural network model.
[0057] The training module is used to use the preprocessed training set to train the convolutional neural network model, extract the key point region features of the human face and perform fusion.
[0058] The test module is used to test the convolutional neural network model by using the preprocessed test set.
[0059] Further, the preprocessing module includes:
[0060] Data cleaning unit, used for data cleaning of training set or test set;
[0061] The face detection unit is used to detect faces in the face pictures in the training set or test set, and extract the face images;
[0062] The face alignment unit is configur...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More
PUM
Login to View More
Abstract
The embodiment of the present invention discloses a face recognition method and device based on fusion of key point region features. The method includes: performing data preprocessing on a training set or a test set; constructing a convolutional neural network model; The training set trains the convolutional neural network model, extracts and fuses features of the key points of the face; uses the preprocessed test set to test the convolutional neural network model. Using the convolutional neural network to fuse the features of the extracted face key point area, it overcomes the shortcomings of traditional face recognition methods that are sensitive to various uncertain interferences such as illumination differences, facial expressions, and whether there is occlusion. The global features and invariant features of the face, the recognition model is more robust and accurate, strengthens the connection between different features, makes the extracted features more complete and complete, can describe the face information more accurately, and improves the classification classification ability of the device.
Description
technical field [0001] Embodiments of the present invention relate to the technical field of computer visionprocessing, and in particular to a face recognition method and device based on feature fusion of key point regions. Background technique [0002] Computer vision is currently the hottest research direction in the field of deep learning. In particular, more and more computer vision technologies have been applied. One of the most widely used technologies is face recognition technology, which is also used in biological A variety of technologies in the field of feature recognition are constantly updating and iterating the technologies that survive after competition and elimination. At present, face recognition methods are mainly divided into two categories, one is representation-based methods, and the other is feature-based methods. The basic idea of the appearance-based method is to convert the two-dimensional face image to another space, and then analyze the face pat...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More
Application Information
Patent Timeline
Application Date:The date an application was filed.
Publication Date:The date a patent or application was officially published.
First Publication Date:The earliest publication date of a patent with the same application number.
Issue Date:Publication date of the patent grant document.
PCT Entry Date:The Entry date of PCT National Phase.
Estimated Expiry Date:The statutory expiry date of a patent right according to the Patent Law, and it is the longest term of protection that the patent right can achieve without the termination of the patent right due to other reasons(Term extension factor has been taken into account ).
Invalid Date:Actual expiry date is based on effective date or publication date of legal transaction data of invalid patent.
Login to View More 
Login to View More