Real-time check-in system based on face recognition
Facedata [Folders]---- source of dataset(Data set files cannot be disclosed)
Facedata-2 [Folders]---- Dataset after data enhancement
saved_model [Folder]---- source of loading models
project [ipynb] ---- Main code for data enhancement, PCA, MLP and real-time recognition programs
attendanc [xlsx] ---- the attendance assign list form
bg1 [qrc] & bg1_rc [py] ---- the source of backgound in GUI
FACE [py] ---- Part of the main project code that will be used for the check-in system is converted into a python file
faceRecg [py] ---- source code of GUI and please run it as main programme
Data set files can be replaced. Put it in the Facedata folder.
Start by running the project.ipynb file. Perform tasks such as data processing, data enhancement, and then model training. Then run the faceRecg.py file for real-time check-in detection, replacing the file's model path file with the result you got. If you do not want to train the model, you can directly use the model in saved_model. best_model_final_2024-02-28_15-27-36.keras is the optimal model
Facial recognition stands as a pivotal biometric technology, exerting significant influence in contemporary society. With the ongoing advancements in machine learning, facial recognition systems find extensive applications across security, identity verification, intelligent surveillance, and various other domains. This project involves the design and implementation of a facial recognition system utilizing Principal Component Analysis (PCA) methodology. Leveraging machine learning techniques such as Multi-Layer Perceptron (MLP) algorithms, the focus is on creating an advanced facial recognition system. The primary objective is to automate facial recognition and classification tasks with high precision and efficiency.
Our objective is to design a sign-in system, which can effectively solve these problems. Through face recognition technology, students can automatically identify and sign in through the camera without manual operation, saving time and labor costs. At the same time, the face recognition system has high accuracy and reliability, effectively preventing fraud and errors in sign-in data. In addition, the system can record students' sign-in status in real time and provide real-time data analysis and management functions to meet the timeliness requirements for sign-in data during the teaching process.
During the development of our system, the entire project includes, ● Data collecting ● Data preprocessing ● Appropriate algorithm ● Coding ● Real-time detection
Applied PCA(Principal Components Analysis)
In face image data, each pixel can be regarded as a feature, and the dimensionality of the image is usually very high, which leads to problems of high computational complexity and poor model generalization ability. PCA maps the original high-dimensional feature space to a low-dimensional subspace through linear transformation, thereby achieving dimensionality reduction of the data while retaining the most important feature information.
Optimize the PCA process by using SVD to speed up the operational efficiency
