Shuowen (Sean) Hu

Dr. Shuowen (Sean) Hu


Research Focus & Metrics

Academic Career Length

12Years (2015-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
12 February 2015 in Journal of the Optical Society of America A
Verified
Journal Article
Open access
PDF file
SHShuowen (Sean) Hu
Shuowen (Sean) Hu
JCJonghyun Choi
Jonghyun Choi
ACAlex L. Chan
Alex L. Chan
WSWilliam Schwartz
William Schwartz
Although visible face recognition has been an active area of research for several decades, cross-modal face recognition has only been explored by the biometrics community relatively recently. Thermal-to-visible face recognition is one of the most difficult cross-modal face recognition challenges, because of the difference in phenomenology between the thermal and visible imaging modalities. We address the cross-modal recognition problem using a partial least squares (PLS) regression-based approach consisting of preprocessing, feature extraction, and PLS model building. The preprocessing and feature extraction stages are designed to reduce the modality gap between the thermal and visible facial signatures, and facilitate the subsequent one-vs-all PLS-based model building. We incorporate multi-modal information into the PLS model building stage to enhance cross-modal recognition. The performance of the proposed recognition algorithm is evaluated on three challenging datasets containing visible and thermal imagery acquired under different experimental scenarios: time-lapse, physical tasks, mental tasks, and subject-to-camera range. These scenarios represent difficult challenges relevant to real-world applications. We demonstrate that the proposed method performs robustly for the examined scenarios.

22 March 2019 in International Journal of Computer Vision
Verified
Journal Article
Closed access
PDF file
HZHe Zhang
He Zhang
BRBenjamin S. Riggan
Benjamin S. Riggan
SHShuowen (Sean) Hu
Shuowen (Sean) Hu
NSNathaniel J. Short
Nathaniel J. Short
VPVishal M. Patel
Vishal M. Patel

1 March 2016 in ISPRS Journal of Photogrammetry and Remote Sensing
Verified
Journal Article
Closed access
YSYun-Ting Su
Yun-Ting Su
JBJames Bethel
James Bethel
SHShuowen (Sean) Hu
Shuowen (Sean) Hu

1 October 2017
Verified
Conference Paper
Closed access
HZHe Zhang
He Zhang
VPVishal M. Patel
Vishal M. Patel
BRBenjamin S. Riggan
Benjamin S. Riggan
SHShuowen (Sean) Hu
Shuowen (Sean) Hu
The large domain discrepancy between faces captured in polarimetric (or conventional) thermal and visible domain makes cross-domain face recognition quite a challenging problem for both human-examiners and computer vision algorithms. Previous approaches utilize a two-step procedure (visible feature estimation and visible image reconstruction) to synthesize the visible image given the corresponding polarimetric thermal image. However, these are regarded as two disjoint steps and hence may hinder the performance of visible face reconstruction. We argue that joint optimization would be a better way to reconstruct more photo-realistic images for both computer vision algorithms and human-examiners to examine. To this end, this paper proposes a Generative Adversarial Network-based Visible Face Synthesis (GAN-VFS) method to synthesize more photo-realistic visible face images from their corresponding polarimetric images. To ensure that the encoded visible-features contain more semantically meaningful information in reconstructing the visible face image, a guidance sub-network is involved into the training procedure. To achieve photo realistic property while preserving discriminative characteristics for the reconstructed outputs, an identity loss combined with the perceptual loss are optimized in the framework. Multiple experiments evaluated on different experimental protocols demonstrate that the proposed method achieves state-of-the-art performance.

3 March 2015 in Optics Letters
Verified
Journal Article
Closed access
PDF file
NSNathaniel Short
Nathaniel Short
SHShuowen (Sean) Hu
Shuowen (Sean) Hu
PGPrudhvi Gurram
Prudhvi Gurram
KGKristan Gurton
Kristan Gurton
ACAlex Chan
Alex Chan
We investigate the performance of polarimetric imaging in the long-wave infrared (LWIR) spectrum for cross-modal face recognition. For this work, polarimetric imagery is generated as stacks of three components: the conventional thermal intensity image (referred to as S0), and the two Stokes images, S1 and S2, which contain combinations of different polarizations. The proposed face recognition algorithm extracts and combines local gradient magnitude and orientation information from S0, S1, and S2 to generate a robust feature set that is well-suited for cross-modal face recognition. Initial results show that polarimetric LWIR-to-visible face recognition achieves an 18% increase in Rank-1 identification rate compared to conventional LWIR-to-visible face recognition. We conclude that a substantial improvement in automatic face recognition performance can be achieved by exploiting the polarization-state of radiance, as compared to using conventional thermal imagery.

13 January 2017 in Applied Optics
Verified
Journal Article
Closed access
PDF file
SHShuowen (Sean) Hu
Shuowen (Sean) Hu
GGGeoffrey H. Goldman
Geoffrey H. Goldman
CBChristoph C. Borel-Donohue
Christoph C. Borel-Donohue
Unmanned aerial vehicles (UAVs) flown by adversaries are an emerging asymmetric threat to homeland security and the military. To help address this threat, we developed and tested a computationally efficient UAV detection algorithm consisting of horizon finding, motion feature extraction, blob analysis, and coherence analysis. We compare the performance of this algorithm against two variants, one using the difference image intensity as the motion features and another using higher-order moments. The proposed algorithm and its variants are tested using field test data of a group 3 UAV acquired with a panoramic video camera in the visible spectrum. The performance of the algorithms was evaluated using receiver operating characteristic curves. The results show that the proposed approach had the best performance compared to the two algorithmic variants.