University of Tehran
In the era of Covid-19, we have become more reliant on virtual interactions such as Zoom meetings, Skype. These live stream webcam videos have become a rich data source to explore. This project will explore the use case of age, gender, and emotion prediction which could facilitate salespeople to understand their customers better, for example.
In this project, our goal was to predict the Age, Gender, and Facial Expressions of humans from their images. After feature extraction from images, we had a vector of features and we have to predict the age, gender, and Facial Expressions of humans from their images.
In the data pre-processing phase, we performed some tasks such as data cleaning, data editing, data reduction. After that, we implemented different algorithms of classification(SVM, decision tree, ..) and regression to find the best performance.
Programming Language: Python
Course: Machine Learning Course
Results:
Data downsampling in the pre-processing phase to balance the size of each class in the training phase:
Accuracy of different classifications for classifying gender labels:
Accuracy of different classifications for classifying facial expressions labels:
Accuracy of different regression algorithm for age labels: