Frankenstein: Learning Deep Face Representations using Small Data

Guosheng Hu, Xiaojiang Peng, Yongxin Yang, Timothy M. Hospedales, Jakob Verbeek

Research output: Contribution to journalArticlepeer-review


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 data sets are not publicly available and difficult to collect. In this paper, we propose a method to generate very large training data sets of synthetic images by compositing real face images in a given data set. 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 data set.
Original languageEnglish
Pages (from-to)293 - 303
Number of pages11
JournalIEEE Transactions on Image Processing
Issue number1
Early online date26 Sep 2017
Publication statusPublished - 1 Jan 2018


Dive into the research topics of 'Frankenstein: Learning Deep Face Representations using Small Data'. Together they form a unique fingerprint.

Cite this