I am an Applied Scientist in Amazon Rekognition working on face related projects. Prior of joining Amazon, I received the Ph.D. degreee from University of Houston advised by Prof. Ioannis A. Kakadiaris in May, 2019. My research is mainly focused on Computer Vision, Deep Learning, Machine Learning, Biometrics, and Face Recognition. I have worked on several projects covering a wide range of topics such as Face Detection, Face Alignment, 3D Face Reconstruction, and Face Recognition. You can find more projects and experiences in my resume.
Download my ResumeMy research is focused on computer vision, machine learning, and deep learning. As a Research Assistant at the University of Houston, my research is addressing on face recognition in the presence of viriances of pose, expressions, and occlusions.
A few well-developed face recognition pipelines have been reported in recent years. Most of the face-related work focuses on a specific module or demonstrates a research idea. The modern well-designed face recognition pipelines that can work in the real life is even less. In this paper, we present a pose-invariant 3D-aided 2D face recognition system (UR2D-E) that is robust to pose variations by leveraging deep learning technology. We describe the architecture and the interface of UR2D-E, and introduce each module in detail. Experiments are conducted on the UHDB31 and IJB-A, demonstrating that UR2D-E outperforms existing 2D face recognition systems such as VGG-Face, FaceNet, and a commercial off-the-shelf software (COTS) by at least 9% on UHDB31 and 3% on IJB-A dataset on average. It fills a gap by providing a 3D-aided 2D face recognition system that has compatible results with 2D face recognition systems using deep learning techniques.
The goal of this work is to seek principles of designing a deep neural network for face reconstruction from a single image. To make the evaluation simple, we generated a synthetic dataset with 10,000 identities and used a part of it for evaluation. We applied extensive experiments using an end-to-end face reconstruction algorithm using E2FAR and its variations, and analyzed the reason why it can be successfully applied for 3D face reconstruction. From the comparative studies, we conclude the feature aggregation from different layers is a key point to training better neural networks for 3D face reconstruction. Based on these observations, a face reconstruction feature aggregation network (FR-FAN) is proposed, which obtains significant improvements compared with all baselines on the synthetic validation set. We evaluate our model on existing popular indoor and in-the-wild 2D-3D datasets. Extensive experiments demonstrate that \name performs statistically significantly better than the existing state-of-the-art algorithms on several datasets. Finally, the sensitivity analysis is performed on controlled datasets demonstrates that our designed network is robust to large variations of pose, illumination, and expressions.
We explore global and local features obtained from Convolutional Neural Networks (CNN) for learning to estimate head pose and localize landmarks jointly. Because there is a high correlation between head pose and landmark locations, the head pose distributions from a reference database and learned local deep patch features are used to reduce the error in the head pose estimation and face alignment tasks. First, we train GNet on the detected face region to obtain a rough estimate of the pose and to localize the seven primary landmarks. The most similar shape is selected for initialization from a reference shape pool constructed from the training samples according to the estimated head pose. Starting from the initial pose and shape, LNet is used to learn local CNN features and predict the shape and pose residuals. We demonstrate that our algorithm, named JFA, improves both the head pose estimation and face alignment. To the best of our knowledge, this is the first system that explores the use of the global and local CNN features to solve head pose estimation and landmark detection tasks jointly.
This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. These works may not be reposted without the explicit permission of the copyright holder.
@inproceedings{Xu_2019_190309, address = {Long Beach, CA}, author = {Xu, Xiang and Zhou, Xiong and Venkatesan, Ragav and Swaminathan, Gurumurthy and Majumder, Orchid}, booktitle = {Proc. IEEE Conference on Computer Vision and Pattern Recognition}, month = {Jun. 16 -- 20}, title = {d-SNE: Domain Adaptation using Stochastic Neighborhood Embedding}, year = {2019} }
@inproceedings{Xu_2019_181105, address = {Waikoloa Village, HI}, author = {Xu, Xiang and Le, Ha A. and Kakadiaris, Ioannis A.}, booktitle = {Proc. IEEE Winter Conference on Applications of Computer Vision}, month = {Jan. 8 -- 10}, title = {On the Importance of Feature Aggregation for Face Reconstruction}, year = {2019} }
@inproceedings{Shi_2018_180923, address = {Los Angeles, CA, USA}, author = {Shi, Lei and Xu, Xiang and Kakadiaris, Ioannis A.}, booktitle = {Proc. IEEE International Conference on Biometrics: Theory Applications and Systems}, month = {Oct 22 -- 25}, title = {{SSFD: A Face Detector via a Single-Scale Feature Map}}, year = {2018} }
@inproceedings{Sarafianos_2018_180921, address = {Munich, Germany}, author = {Sarafianos, Nikolaos and Xu, Xiang and Kakadiaris, Ioannis A}, booktitle = {Proc. European Conference on Computer Vision}, month = {Sep. 8 --14}, title = {{Deep Imbalanced Attribute Classification using Visual Attention Aggregation}}, year = {2018} }
@inproceedings{Xu_2017_17643, address = {Denver, CO}, author = {Xu, X and Le, H and Dou, P and Wu, Y and Kakadiaris, I A}, booktitle = {Proc. International Joint Conference on Biometrics}, month = {Oct. 1--4}, pages = {446--455}, title = {{Evaluation of 3D-aided pose invariant 2D face recognition system}}, year = {2017} }
@inproceedings{Xu_2017_17394, address = {Washington, DC}, author = {Xu, X and Kakadiaris, I A}, booktitle = {Proc. $12^{th}$ IEEE Conference on Automatic Face {\&} Gesture Recognition}, month = {May}, pages = {642--649}, title = {{Joint head pose estimation and face alignment framework using global and local CNN features}}, year = {2017} }
@inproceedings{Xu_2016_16539, address = {Sendai, Japan}, author = {Xu, X and Shah, S and Kakadiaris, I A}, booktitle = {Proc. IEEE International Conference on Identity, Security and Behavior Analysis}, title = {{Face alignment via an ensemble of random ferns}}, year = {2016} }
@inproceedings{wu2015towards, title={Towards fitting a 3D dense facial model to a 2D image: A landmark-free approach}, author={Wu, Yuhang and Xu, Xiang and Shah, Shishir K and Kakadiaris, Ioannis A}, booktitle={Proc. IEEE International Conference on Biometrics Theory, Applications and Systems}, pages={1--8}, year={2015} }
Here, I list some codes or resources I contribute to the research communities.
I would be happy to talk to you if you need my assistance in your research or you have any suggestions or comments about my work.