Sebastian Bunda, Luuk Spreeuwers and Chris ZeinstraBrömme, ArslanDamer, NaserGomez-Barrero, MartaRaja, KiranRathgeb, ChristianSequeira Ana F.Todisco, MassimilianoUhl, Andreas2022-10-272022-10-272022978-3-88579-723-4https://dl.gi.de/handle/20.500.12116/39700Converting convolutional neural networks such as MobileNets to a full integer representation is already quite a popular method to reduce the size and computational footprint of classification networks but its effect on face recognition networks is relatively unexplored. This work presents a method to reduce the size of MobileFaceNet using sub-byte quantization of the weights and activations. It was found that 8-bit and 4-bit versions of MobileFaceNet can be obtained with 98.68% and 98.63% accuracy on the LFW dataset which reduces the footprint to 25% and 12.5% of the original weights respectively. Using mixed-precision, an accuracy of 98.17% can be achieved whilst requiring only 10% of the original weight footprint. It is expected that with a larger training dataset, higher accuracies can be achieved.enResource Limited Face RecognitionDeep Neural NetworksQKerasSub-byte QuantizationText/Conference Paper10.1109/BIOSIG55365.2022.98970251617-5490