Facial Recognition

The goal of this project was to use OpenCV to detect faces and build a model to perform facial recognition.

I collected images for the following three different people: myself, Simu Liu, and Andrew Ng. I then used those photos to train the model. I decided to use other Asian males in the dataset to see how well the model will perform. While this project was made on a smaller scale, the program has the capability to be used for a larger and more diverse set of people which will help avoid bias.

Of the 15 test images, the model only labeled one of the images wrong.

Surprisingly, the model only labeled one image wrong even though I used images in the testing set that were a bit different than those I used in the training set. For example, I used an image of a much younger Andrew with glasses, one with me wearing a mask and another one of me making a funny face. These were all things that were not included in the training set. Even with such a small training set of 12 images per person, the model was able to correctly label all but one image. However, I could probably improve the model by training on more data.

The IPYNB file can be found on my Github which is linked on the Home page.