Face Hallucination via Split-Attention in Split-Attention Network
Authors:
Open Access
- 17 October 2021
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 5501-5509
Abstract
Recently, convolutional neural networks (CNNs) have been widely employed to promote the face hallucination due to the ability to predict high-frequency details from a large number of samples. However, most of them fail to take into account the overall facial profile and fine texture details simultaneously, resulting in reduced naturalness and fidelity of the reconstructed face, and further impairing the performance of downstream tasks (e.g., face detection, facial recognition). To tackle this issue, we propose a novel external-internal split attention group (ESAG), which encompasses two paths responsible for facial structure information and facial texture details, respectively. By fusing the features from these two paths, the consistency of facial structure and the fidelity of facial details are strengthened at the same time. Then, we propose a split-attention in split-attention network (SISN) to reconstruct photorealistic high-resolution facial images by cascading several ESAGs. Experimental results on face hallucination and face recognition unveil that the proposed method not only significantly improves the clarity of hallucinated faces, but also encourages the subsequent face recognition performance substantially. Codes have been released at https://github.com/mdswyz/SISN-Face-Hallucination.Keywords
All Related Versions
Funding Information
- Hubei Technology Innovation Project (2019AAA045)
- 2020 Hubei Province High-value Intellectual Property Cultivation Project
- Wuhan Enterprise Technology Innovation Project (202001602011971)
- Central Government Guides Local Science and Technology Development Special Projects (2018ZYYD059)
- Graduate Innovative Fund of Wuhan Institute of Technology (CX2020223)
- National Natural Science Foundation of China (62072350, U1903214, 61971165, 62001334)
This publication has 32 references indexed in Scilit:
- Robust and efficient face recognition via low-rank supported extreme learning machineMultimedia Tools and Applications, 2017
- Joint Face Detection and Alignment Using Multitask Cascaded Convolutional NetworksIEEE Signal Processing Letters, 2016
- Global face reconstruction for face hallucination using orthogonal canonical correlation analysis2015
- Deep Learning Face Attributes in the Wild2015
- Learning a Deep Convolutional Network for Image Super-Resolution2014
- Locality-constraint iterative neighbor embedding for face hallucination2013
- SCface – surveillance cameras face databaseMultimedia Tools and Applications, 2009
- Hallucinating Face by EigentransformationIEEE Transactions on Systems, Man and Cybernetics, Part C (Applications and Reviews), 2005
- Super-resolution through neighbor embedding2004
- Image quality assessment: from error visibility to structural similarityIEEE Transactions on Image Processing, 2004