Angela Brown
For our project for EECS 351, our goal was to try to implement image and video filters similar to the filters found in Snapchat. We wanted to try to implement filters such as hats, glasses, and mustaches, meaning that we had to detect noses, eyes, and faces as a whole. After we implemented some of the filters using Matlab's built in functions we also tried to improve our filtering algorithms.
Throughout this project, we applied techniques we learned in class and ones we learned through research. The class tools we utilized are system diagrams, 2D convolution, and machine learning. Outside tools used in this project include facial recognition algorithms, types of feature detection, and video processing.
​
After working on this project, we learned how to use facial recognition in Matlab using the Computer Vision Toolbox and the theory behind how this works. We learned how to extend our image processing to video processing. We learned how to incorporate machine learning techniques to gain accuracy. We have greatly increased our working knowledge of Matlab.
​
By the end of the semester, we have successfully created and implemented filters on image and video files using facial recognition algorithms, and improved these algorithms using training techniques. If we had more time to work on this project, we would want to implement our process on real-time videos and put filters on multiple faces in one image or video file. We would spend additional time on training to include more samples in the training process and to optimize the parameters we set. We would love to explore different algorithms and systems such as neural networking to try to improve our accuracy.
​
TEAM
ABOUT
