Frankenstein: Learning Deep Face Representations using Small Data - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Journal Articles IEEE Transactions on Image Processing Year : 2018

Frankenstein: Learning Deep Face Representations using Small Data


Deep convolutional neural networks have recently proven extremely effective for difficult face recognition problems in uncontrolled settings. To train such networks, very large training sets are needed with millions of labeled images. For some applications, such as near-infrared (NIR) face recognition, such large training datasets are not publicly available and difficult to collect. In this work, we propose a method to generate very large training datasets of synthetic images by compositing real face images in a given dataset. We show that this method enables to learn models from as few as 10,000 training images, which perform on par with models trained from 500,000 images. Using our approach we also obtain state-of-the-art results on the CASIA NIR-VIS2.0 heterogeneous face recognition dataset.
Fichier principal
Vignette du fichier
bare_jrnl.pdf (1.67 Mo) Télécharger le fichier
Vignette du fichier
hal.jpg (27.59 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Format : Figure, Image

Dates and versions

hal-01306168 , version 1 (22-04-2016)
hal-01306168 , version 2 (25-09-2017)



Guosheng Hu, Xiaojiang Peng, Yongxin Yang, Timothy Hospedales, Jakob Verbeek. Frankenstein: Learning Deep Face Representations using Small Data. IEEE Transactions on Image Processing, 2018, 27 (1), pp.293-303. ⟨10.1109/TIP.2017.2756450⟩. ⟨hal-01306168v2⟩
1595 View
1583 Download



Gmail Facebook Twitter LinkedIn More