Please download the AT&T face recognition dataset fromhttps://www.kaggle.com/kasikrit/att-database-of-faces(Links to an external site.)(40 subjects with 10 different mugshots for each).
Follow MATLAB’s transfer learning examplehttps://www.mathworks.com/help/deeplearning/examples/transfer-learning-using-alexnet.html(Links to an external site.). Create a network that identifies each subject as a class (instead of the aforesaid example’s objects and their labels, here we have face images assigned to their unique identities). You can use the same parameters as MATLAB’s example but also feel free to tune them for better results. Report on resulting validation accuracies similar to the aforesaid MATLAB example. Don’t forget to read MATLAB help files on “imageDatastore” and its various options.
I need validation accuracy and screenshots of MATLAB output window , plus all your MATLAB code and network objects.
Already registered? Login
Not Account? Sign up
Enter your email address to reset your password
Back to Login? Click here