Slim-CNN: A Light-Weight CNN for Face Attribute Prediction

  • 2019-07-03 23:02:44
  • Ankit Sharma, Hassan Foroosh
  • 16


We introduce a computationally-efficient CNN micro-architecture Slim Moduleto design a lightweight deep neural network Slim-Net for face attributeprediction. Slim Modules are constructed by assembling depthwise separableconvolutions with pointwise convolution to produce a computationally efficientmodule. The problem of facial attribute prediction is challenging because ofthe large variations in pose, background, illumination, and dataset imbalance.We stack these Slim Modules to devise a compact CNN which still maintains veryhigh accuracy. Additionally, the neural network has a very low memory footprintwhich makes it suitable for mobile and embedded applications. Experiments onthe CelebA dataset show that Slim-Net achieves an accuracy of 91.24% with atleast 25 times fewer parameters than comparably performing methods, whichreduces the memory storage requirement of Slim-net by at least 87%.


Quick Read (beta)

This feature is not avaialbe for this paper.