Benchmarking deep learning techniques for face recognition

Abstract

Recent progresses in Convolutional Neural Networks (CNNs) and GPUs have greatly advanced the state-of-the-art performance for face recognition. However, training CNNs for face recognition is complex and time-consuming. Multiple factors need to be considered: deep learning frameworks, GPU platforms, deep network models, training datasets and test datasets. The deep models under different frameworks may perform differently. Based on this concern, we compare three deep learning frameworks and benchmark the performance of different CNN models on five GPU platforms. The scalability issue is also explored. Our findings can help researchers select appropriate face recognition models, deep learning frameworks, GPU platforms, and training datasets for their face recognition tasks.