COGS181(Neural Networks and Deep Learning)_FinalProject

Abstract: In recent years deep learning, especially Convolutional Neural Networks (CNNs,) has shown exponential progress in face recognition tasks and achieved state of the art results. Face recognition involves identifying an individual given their pictures/videos. It is widely used in many industries, from law enforcement to social media to entertainment. This task can be accomplished by machine learning models such as CNNs, Transfer Learning Models, or Recurrent Neural Networks. This paper proposes two CNN architectures and evaluates their performance on the Labeled Faces in the Wild dataset, additionally, parameter-efficient fine-tuning and various techniques will be used to improve the performance of the models and make them more computationally efficient and accurate. Labeled Faces in the Wild is a public benchmark dataset for such a task it contains over 13,000 images of labeled faces.