Comparative Analysis to Improve the Image Accuracy In Face Recognition System Using Hybrid LDA Compared With PCA
V.R. Thushitha, Ms. Priya · 2022 International Conference on Business Analytics for Technology and Security (ICBATS) · 2022
Aim-This study depicts the improvisation and comparison of accuracy of two different face recognition algorithms to improvise the novel face detection rate from the stored database under various lighting conditions. Materials and methods-From Kaggle, a total of 13000 face samples are being collected from LFW people face recognition dataset. Linear Discriminant Analysis (LDA) and Principal Component Analysis (PCA) algorithms are improvised and compared to recognize the faces and to increase the accuracy rate. Results-Based on the MATLAB simulation and verification, LDA shows the recognition rate of 85.2% and PCA shows 73%. From the statistical analysis, the significant accuracy ratio is <0.05. Conclusion-It is concluded that the LDA algorithm shows enhanced features in edge detection, image equalization and image normalization than the PCA algorithm in the face recognition system for the dataset considered.