Average Biased ReLU Based CNN Descriptor for Improved Face Retrieval
The original preprint is available at: https://arxiv.org/abs/1804.02051
cnn_vgg_faces_abrelu.m returns the Average Biased ReLU Based CNN Face Descriptor for an input image.
MatconvNet is used in this research.
The original trained VGGFace model is used and updated to the AB-ReLU based Face VGGFace descriptor by Dr. Shiv Ram Dubey, IIIT Sri City (https://sites.google.com/site/shivram1987/).
The original trained VGGFace model is available at: http://www.vlfeat.org/matconvnet/pretrained/
This code can be used only for the academic and research purposes and can not be used for any commercial purposes.
Cite the paper as (in case you are using this code): 'S.R. Dubey, S. Chakraborty, Average Biased ReLU Based CNN Descriptor for Improved Face Retrieval. arXiv Preprint, 1804.02051, 2018.'
This Research is funded by IIIT Sri City, INDIA through Institute Research Grant. We gratefully acknowledge the support of NVIDIA Corporation with the donation of the GeForce Titan X Pascal GPU for our research.