.. vim: set fileencoding=utf-8 :
.. Tiago de Freitas Pereira <tiago.pereira@idiap.ch>


===================================
Resnet V1 trained by David Sandberg
===================================

Very recently, David Sandberg has made publicly available his implementation as well as trained models for a new FR-CNN named `**FaceNet** <https://github.com/davidsandberg/facenet>`_.
This is the closest open-source implementation of the FaceNet CNN proposed by [Schroff_2015], for which neither a pre-trained model nor the training-set is publicly available.
Sandberg's FaceNet implements an Inception-ResNet V1 DNN architecture [Szegedy2017]_.
In these tests, we have used the 20170512-110547 model, trained on the MS-Celeb-1M dataset.
Using this model, FaceNet achieves a FR performance of 99.2\% on the LFW dataset.
We use this representation to construct enrollment and probe templates, which are compared to each other using the Cosine measure.


More information `check here <https://www.idiap.ch/software/bob/docs/bob/bob.ip.tensorflow_extractor/master/index.html>`_.