CN111832499A - Simple face recognition classification system - Google Patents
Simple face recognition classification system Download PDFInfo
- Publication number
- CN111832499A CN111832499A CN202010695682.8A CN202010695682A CN111832499A CN 111832499 A CN111832499 A CN 111832499A CN 202010695682 A CN202010695682 A CN 202010695682A CN 111832499 A CN111832499 A CN 111832499A
- Authority
- CN
- China
- Prior art keywords
- face
- module
- recognition
- neural network
- face recognition
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
- G06V40/166—Detection; Localisation; Normalisation using acquisition arrangements
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- General Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a simple face recognition and classification system, which comprises a face capturing module, a data processing module, a neural network module and a recognition testing module; the face capturing module utilizes an OpenCV or a Dlib library to mark and extract a face in a captured face of a camera to obtain a face data set; the data processing module processes the face data set obtained by the face capturing module so that the face data set can be trained and learned by a neural network; the neural network module trains and tests the processed data obtained by the data processing module to obtain a weight model file for network training; and the recognition testing module is used for demonstrating face recognition. According to the invention, a simple face recognition classification system is built by quickly utilizing a TensorFlow deep learning framework, so that more reference possibilities are provided for the application of deep learning in face recognition; meanwhile, the system also provides more possibilities in the fields of application of face recognition, such as security verification, entrance guard recognition, face payment and the like.
Description
Technical Field
The invention relates to a classification system, in particular to a simple face recognition classification system which provides more reference possibilities for the application of deep learning in face recognition.
Background
At present, the face recognition technology applying deep learning frames at home and abroad is relatively mature, main research focuses on faster recognition speed and better recognition effect, related technical backgrounds mostly select a certain deep learning frame, such as TensorFlow, pitorch, paddlepaddle, and other frames, and then the selected frame is used for building a network. In addition to the final recognition effect influenced by the network design, the setting of the data set also has an important influence on the face recognition effect, and most of the current practices of finishing the frame application face recognition are to process the data set through the existing known face data set, so that the self-defined operation is not much.
Disclosure of Invention
In order to solve the defects of the technology, the invention provides a simple face recognition and classification system.
In order to solve the technical problems, the invention adopts the technical scheme that: a simple face recognition classification system comprises a face capture module, a data processing module, a neural network module and a recognition test module;
the face capturing module utilizes an OpenCV or a Dlib library to mark and extract a face in a captured face of a camera to obtain a face data set; or crawling a star picture on a microblog by using a given code, manually screening in a local folder for the first time to remove non-conforming pictures, and then extracting a face by using an OpenCV or a Dlib library; manually carrying out secondary manual inspection on all files which are cut and extracted with a python library to delete pictures with unobvious human face features, thereby obtaining a human face data set required by the human face recognition;
the data processing module processes the face data set obtained by the face capturing module so that the face data set can be trained and learned by a neural network;
the neural network module trains and tests the processed data obtained by the data processing module to obtain a weight model file for network training;
the recognition testing module is used for demonstrating face recognition, recognition and classification are carried out in two modes of capturing faces through a camera and typing in a face picture file path, feature extraction is carried out on the captured faces through a Dlib library, each feature value is confirmed through a model file, and finally corresponding recognition accuracy is given.
Furthermore, the data processing module marks the face by using the one-hot vector, further determines the size of the face picture, and re-cuts the face picture which does not meet the requirement to ensure that the sizes of the face picture are unified to 64mm multiplied by 64 mm.
Furthermore, the neural network module is realized by using a highly modularized Keras deep learning library and using a TensorFlow-GPU deep learning framework as a rear end in a network building part.
Further, the neural network module extracts picture features by means of two-layer convolution and one-layer pooling, then completes a classifier task by using one layer of full connection, divides processed data of the data processing module into a training set and a testing set according to a ratio of 7:3 after random processing, learns on a built network, and performs 120 batches of training tests to obtain a corresponding training weight model which is used as an important basis for face recognition.
Compared with the existing face recognition project, the invention can obtain the following effects:
(1) through a small amount of codes, the independent python files can locally realize a simple face recognition system.
(2) The method only needs to be deployed in anaconda (open source Python packet manager), can be rapidly transplanted to each platform, and is not strong in dependency.
(3) The face data set adopted by the system is convenient to obtain and has higher self definition, more possibility is generated for the acquisition of the face data, and the existing known face data set is not simply utilized for identification.
(4) The GPU version of the Tensorflow deep learning framework is adopted, and calculation acceleration during training is completed by using CUDA10.0 and cuDNN7.6.5.
(5) The system has the advantages that the probability accuracy of the face in each data set is high in face recognition and classification, and a large amount of face data are not needed for learning.
According to the invention, a simple face recognition classification system is built by quickly utilizing a TensorFlow deep learning framework, so that more reference possibilities are provided for the application of deep learning in face recognition; meanwhile, the system also provides more possibilities in the fields of application of face recognition, such as security verification, entrance guard recognition, face payment and the like.
Drawings
Fig. 1 is an overall frame diagram of the present invention.
Fig. 2 is a schematic view of the connection structure of the modules of the present invention.
Figure 3 is a flow diagram of the face capture module.
FIG. 4 is a flow diagram of a data processing module.
Fig. 5 is a diagram of the structure of the convolutional neural network of the present invention.
Fig. 6 is a system architecture diagram of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and specific embodiments.
As shown in fig. 1-6, a simple face recognition and classification system includes a face capture module, a data processing module, a neural network module, and a recognition test module;
the face capturing module utilizes an OpenCV or a Dlib library to mark and extract a face in a captured face of a camera to obtain a face data set; or crawling a star picture on a microblog by using a given code, manually screening in a local folder for the first time to remove non-conforming pictures, and then extracting a face by using an OpenCV or a Dlib library; manually carrying out secondary manual inspection on all files which are cut and extracted with a python library to delete pictures with unobvious human face features, thereby obtaining a human face data set required by the human face recognition; the face capture module is mainly used for preparing a data set of the project, the used data is not needed to be too much, the acquisition mode is simple, and the data set is representative.
The data processing module processes the face data set obtained by the face capturing module so that the face data set can be trained and learned by a neural network; and the data processing module marks the face by using the one-hot vector, further determines the size of the face picture, and re-cuts the face pictures which do not meet the requirements to ensure that the sizes of the face pictures are unified to be 64mm multiplied by 64 mm.
The neural network module trains and tests the processed data obtained by the data processing module to obtain a weight model file for network training; the neural network module is realized by using a highly modularized Keras deep learning library and using a TensorFlow-GPU deep learning framework as a rear end in a network building part.
The neural network module extracts picture features by means of two-layer convolution and one-layer pooling, then completes a classifier task by using one layer of full connection, divides processed data of the data processing module into a training set and a testing set according to a ratio of 7:3 after random processing, learns on a built network, and goes to a corresponding training weight model after 120 batches of training tests, wherein the training weight model is used as an important basis for face recognition.
The recognition testing module is used for demonstrating face recognition, recognition and classification are carried out in two modes of capturing faces through a camera and typing in a face picture file path, feature extraction is carried out on the captured faces through a Dlib library, each feature value is confirmed through a model file, and finally corresponding recognition accuracy is given.
The specific implementation mode of the invention mainly comprises the following operations:
the first step is as follows: the installation of anaconda is completed, after the installation is completed, a python3.5 version environment is newly built by using anaconda powershell prompt, which can be named TensorFlow, and Spyder is installed by clicking on the main page of the anaconda.
The second step is that: TensorFlowGPU-1.15, Keras and related OpenCV, Dlib libraries were installed. All required libraries can use pip commands to install related whl files in anaconda powershell project, TensorFlow1.15gpu versions are required to be installed, Keras latest versions are required, OpenCV and dlib are required to be installed, and versions of python3.5 and plot libraries are required.
The third step: CUDA10.0 and cuDNN7.6.5 are installed.
The fourth step: and directly opening the Spyder and then running the corresponding python file. Firstly, a code file for capturing a human face is operated, when the code file is operated, a program calls a camera to intercept the human face of a video stream frame by frame, and simultaneously, under the action of a human face recognition classifier, the program only stores pictures of the human face, and after the operation is finished, a folder of 2000 pictures of the human face is formed. And then selecting a microblog ID of a certain blogger to crawl microblog pictures, operating a crawling code, storing all pictures under all bloggers of the blogger, manually processing after storing, replacing the microblog ID, and crawling a new group of face pictures. And moreover, a code for processing the picture is operated, and the crawled picture is cut and the face of the person is extracted. And after a uniform human face data set is obtained, executing a code file for image reprocessing to finish image marking, and simultaneously making graying and size uniformity. Then, running the training code, and obtaining a corresponding test loss and test accuracy curve graph and a model file after the running is finished; and finally, running a face recognition code to perform recognition and classification.
The present invention will be described in further detail with reference to examples.
The first embodiment is as follows:
the face capturing module crawls a star picture on a microblog by using a given code, performs manual screening for the first time in a local folder to remove non-conforming pictures, and then extracts a face by using an OpenCV or a Dlib library; manually carrying out secondary manual inspection on all files which are cut and extracted with a python library to delete pictures with unobvious human face features, thereby obtaining a human face data set required by the human face recognition; the data set used in this face recognition project is composed of 2000 pictures of oneself captured by computer camera, crawled on microblog, and then processed by hand 324 pictures of Liu Chang of man and 1155 pictures of Zheng Chang Xing Zheng Shuang, finally processed by network 321 pictures of Liu De Hua and 342 wu Zu.
The above embodiments are not intended to limit the present invention, and the present invention is not limited to the above examples, and those skilled in the art may make variations, modifications, additions or substitutions within the technical scope of the present invention.
Claims (4)
1. A simple face recognition classification system is characterized in that: the system comprises a face capturing module, a data processing module, a neural network module and an identification testing module;
the face capturing module utilizes an OpenCV or a Dlib library to mark and extract a face in a captured face of a camera to obtain a face data set; or crawling a star picture on a microblog by using a given code, manually screening in a local folder for the first time to remove non-conforming pictures, and then extracting a face by using an OpenCV or a Dlib library; manually carrying out secondary manual inspection on all files which are cut and extracted with a python library to delete pictures with unobvious human face features, thereby obtaining a human face data set required by the human face recognition;
the data processing module processes the face data set obtained by the face capturing module so that the face data set can be trained and learned by a neural network;
the neural network module trains and tests the processed data obtained by the data processing module to obtain a weight model file for network training;
the recognition testing module is used for demonstrating face recognition, recognizing and classifying two modes of capturing a face and typing a face picture file path through a camera, extracting features of the captured face by using a Dlib library, confirming each feature value through a model file, and finally giving out corresponding recognition accuracy.
2. The simplified face recognition classification system of claim 1, wherein: and the data processing module marks the face by using the one-hot vector, further determines the size of the face picture, and re-cuts the face pictures which do not meet the requirements to ensure that the sizes of the face pictures are unified to be 64mm multiplied by 64 mm.
3. The simplified face recognition classification system of claim 1, wherein: the neural network module is realized by using a highly modularized Keras deep learning library and using a TensorFlow-GPU deep learning framework as a rear end in a network building part.
4. The simplified face recognition classification system of claim 3, wherein: the neural network module extracts picture features by means of two-layer convolution and one-layer pooling, then completes a classifier task by using one layer of full connection, divides processed data of the data processing module into a training set and a testing set according to a ratio of 7:3 after random processing, learns on a built network, and goes to a corresponding training weight model after 120 batches of training tests, wherein the training weight model is used as an important basis for face recognition.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010695682.8A CN111832499A (en) | 2020-07-17 | 2020-07-17 | Simple face recognition classification system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010695682.8A CN111832499A (en) | 2020-07-17 | 2020-07-17 | Simple face recognition classification system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111832499A true CN111832499A (en) | 2020-10-27 |
Family
ID=72922927
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010695682.8A Pending CN111832499A (en) | 2020-07-17 | 2020-07-17 | Simple face recognition classification system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111832499A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113343773A (en) * | 2021-05-12 | 2021-09-03 | 上海大学 | Facial expression recognition system based on shallow convolutional neural network |
| CN114627499A (en) * | 2022-03-07 | 2022-06-14 | 上海应用技术大学 | Online safety helmet face recognition method based on convolutional neural network |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107729872A (en) * | 2017-11-02 | 2018-02-23 | 北方工业大学 | Facial expression recognition method and device based on deep learning |
| CN107742117A (en) * | 2017-11-15 | 2018-02-27 | 北京工业大学 | A kind of facial expression recognizing method based on end to end model |
| CN109522861A (en) * | 2018-11-28 | 2019-03-26 | 西南石油大学 | A kind of micro- expression recognition method of face multiclass |
| CN110472460A (en) * | 2018-05-11 | 2019-11-19 | 北京京东尚科信息技术有限公司 | Face image processing process and device |
| CN111126347A (en) * | 2020-01-06 | 2020-05-08 | 腾讯科技（深圳）有限公司 | Human eye state recognition method and device, terminal and readable storage medium |
| CN111295669A (en) * | 2017-06-16 | 2020-06-16 | 马克波尔公司 | Image processing system |
| CN111401107A (en) * | 2019-01-02 | 2020-07-10 | 上海大学 | Multi-mode face recognition method based on feature fusion neural network |
-
2020
- 2020-07-17 CN CN202010695682.8A patent/CN111832499A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111295669A (en) * | 2017-06-16 | 2020-06-16 | 马克波尔公司 | Image processing system |
| CN107729872A (en) * | 2017-11-02 | 2018-02-23 | 北方工业大学 | Facial expression recognition method and device based on deep learning |
| CN107742117A (en) * | 2017-11-15 | 2018-02-27 | 北京工业大学 | A kind of facial expression recognizing method based on end to end model |
| CN110472460A (en) * | 2018-05-11 | 2019-11-19 | 北京京东尚科信息技术有限公司 | Face image processing process and device |
| CN109522861A (en) * | 2018-11-28 | 2019-03-26 | 西南石油大学 | A kind of micro- expression recognition method of face multiclass |
| CN111401107A (en) * | 2019-01-02 | 2020-07-10 | 上海大学 | Multi-mode face recognition method based on feature fusion neural network |
| CN111126347A (en) * | 2020-01-06 | 2020-05-08 | 腾讯科技（深圳）有限公司 | Human eye state recognition method and device, terminal and readable storage medium |
Non-Patent Citations (4)
| Title |
|---|
| 张光旭: "非理想条件下视频序列人脸识别算法研究与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 李红蕾: "构建图形图像数据集的方法概述", 《计算机产品与流通》 * |
| 汪雅丹: "基于卷积神经网络的人脸识别研究与实现", 《电声技术》 * |
| 王荣等: "多变环境下基于多尺度卷积网络的猪个体识别", 《江西农业大学学报》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113343773A (en) * | 2021-05-12 | 2021-09-03 | 上海大学 | Facial expression recognition system based on shallow convolutional neural network |
| CN114627499A (en) * | 2022-03-07 | 2022-06-14 | 上海应用技术大学 | Online safety helmet face recognition method based on convolutional neural network |
| CN114627499B (en) * | 2022-03-07 | 2024-04-09 | 上海应用技术大学 | On-line safety helmet face recognition method based on convolutional neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108898137B (en) | Natural image character recognition method and system based on deep neural network | |
| CN108537136B (en) | Pedestrian re-identification method based on attitude normalization image generation | |
| CN107316001A (en) | Small and intensive method for traffic sign detection in a kind of automatic Pilot scene | |
| CN112381075B (en) | Method and system for carrying out face recognition under specific scene of machine room | |
| CN110599445A (en) | Target robust detection and defect identification method and device for power grid nut and pin | |
| CN113807276B (en) | Smoking behavior identification method based on optimized YOLOv4 model | |
| CN110705357A (en) | Face recognition method and face recognition device | |
| CN112487848B (en) | Character recognition method and terminal equipment | |
| CN111832499A (en) | Simple face recognition classification system | |
| US11605210B2 (en) | Method for optical character recognition in document subject to shadows, and device employing method | |
| CN111046858A (en) | Image-based animal species fine classification method, system and medium | |
| CN111008576A (en) | Pedestrian detection and model training and updating method, device and readable storage medium thereof | |
| CN112464925A (en) | Mobile terminal account opening data bank information automatic extraction method based on machine learning | |
| CN111476232A (en) | Water washing label detection method, equipment and storage medium | |
| CN111126112B (en) | Candidate region determination method and device | |
| CN110826364B (en) | Library position identification method and device | |
| CN110826534A (en) | Face key point detection method and system based on local principal component analysis | |
| CN110618129A (en) | Automatic power grid wire clamp detection and defect identification method and device | |
| CN111310751A (en) | License plate recognition method and device, electronic equipment and storage medium | |
| CN110969173A (en) | Target classification method and device | |
| CN116189063B (en) | Key frame optimization method and device for intelligent video monitoring | |
| CN110689066B (en) | Training method combining face recognition data equalization and enhancement | |
| CN116580232A (en) | Automatic image labeling method and system and electronic equipment | |
| CN114639013A (en) | Remote sensing image airplane target detection and identification method based on improved Orient RCNN model | |
| CN111127327B (en) | Picture inclination detection method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |