DeepBE: Learning Deep Binary Encoding for Multi-Label Classification
Institute of Automation, CAS
The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, pp. 39-46, 2016
@InProceedings{Li_2016_CVPR_Workshops,
author={Li, Chenghua and Kang, Qi and Ge, Guojing and Song, Qiang and Lu, Hanqing and Cheng, Jian},
title={DeepBE: Learning Deep Binary Encoding for Multi-Label Classification},
booktitle={The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops},
month={June},
year={2016}
}
The track 2 and track 3 of ChaLearn 2016 can be considered as Multi-Label Classification problems. We present a framework of learning deep binary encoding (DeepBE) to deal with multi-label problems by transforming multi-labels to single labels. The transformation of DeepBE is in a hidden pattern, which can be well addressed by deep convolutions neural networks (CNNs). Furthermore, we adopt an ensemble strategy to enhance the learning robustness. This strategy is inspired by its effectiveness in fine-grained image recognition (FGIR) problem, while most of face related tasks such as track 2 and track 3 are also FGIR problems. By DeepBE, we got 5.45% and 10.84% mean square error for track 2 and track 3 respectively. Additionally, we proposed an algorithm adaption method to treat the multiple labels of track 2 directly and got 6.84% mean square error.
June 30, 2016 by hgpu
