A Combination between Deep learning for feature extraction and Machine Learning for Recognition
Mahmoud Ali, Dinesh Kumar · 2021
Feature extraction is a technique that analyzes faces in pictures or videos which gives the features of the face, and then it determines the specific target. Deep learning has made a real leap in the field of feature extraction. The developers have developed facial feature techniques by relying on deep learning technology to improve the accuracy and efficiency of face analysis and recognition. As a result of the above, developers and programmers began to pay great attention to face recognition technology and started to use it in so many applications. Therefore, this paper provides a comparison between the results of classification methods such as kNN, SVM, Random Forest, and Logistic Regression, based on the deep learning methods for feature extraction such as Squeeze Net and Inception model on the Pins Face Recognition dataset. Logistic Regression with Ridge regularization based on the results of the Inception model was the fastest model and it achieved the best accuracy 94.87% in 68 minutes.