Frankenstein: Learning Deep Face Representations Using Small Data
By Guosheng Hu, Xiaojiang Peng, Yongxin Yang, Timothy M. Hospedales, Jakob Verbeek. 2016
127 citations recorded by the source corpus.
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.
Subfields: Computer Vision and Pattern Recognition
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Computer Vision and Pattern Recognition
- Björn Barz
- Chandra Thapa
- Joachim Denzler
- Sergey I. Nikolenko
- Seyit Ahmet Camtepe
- M. Mahdi Roozbahani
- Umme Aiman
- Huili Huang
Citations in the corpus
- ScaleNet: An Unsupervised Representation Learning Method for Limited Information (2023, background): “Multiple studies employed data augmentation techniques such as scaling, rotating, cropping, and generating synthetic samples to produce more training samples [22,45,23].”
- Synthetic Data for Deep Learning (2019, background): “3 we have discussed the “Frankenstein” pipeline for compositing human faces [256].”
- SplitFed: When Federated Learning Meets Split Learning (2020, background): “, GPUs) as well as large datasets to produce high accuracy [1,2].”
- Deep Learning on Small Datasets without Pre-Training using Cosine Loss (2019, background): “[16], for instance, composite face parts from different images to create new face images and Shrivastava et al.”
- Face recognition using modified deep learning neural network (2017, background): “In such cases the use of deep learning is put off as the network will overfit severely with smaller training data [7].”