Felix Juefei Xu     

{ { { Deep Learning } Machine Learning, Computer Vision } AI } Researcher

Ph.D. in Electrical and Computer Engineering, Carnegie Mellon University

M.S. in Machine Learning, Carnegie Mellon University

M.S. in Electrical and Computer Engineering, Carnegie Mellon University

B.S. in Electronic Engineering, Shanghai Jiao Tong University

Felix Juefei Xu is a Research Scientist with GenAI at Meta, based in New York City, where he works on robust perception and efficient learning problems in the domain of generative AI. He is also affiliated with New York University as an Adjunct Professor. Previously, he was a Research Scientist with Alibaba Group, based in Sunnyvale, CA. During his Ph.D. studies at CMU, he was working in a research group specializing in pattern recognition, machine learning, computer vision, and image processing, especially as applied to the field of biometrics, at Carnegie Mellon CyLab Biometrics Center under the supervision of Prof. Marios Savvides. CyLab is CMU's Security and Privacy Institute. Felix is published under Felix Juefei-Xu or F. Juefei-Xu.

Email  /  Bio  /  Google Scholar  /  Twitter  /  Linkedin  /  Photo

Research

Deep Learning, Computer Vision, Generative Models, Adversarial Robustness, AI Safety.

My research in general is focused on a fuller understanding of deep learning where I am actively exploring new methods in deep learning that are statistically efficient and adversarially robust. Through these efforts, I am investigating under what conditions deep learning starts to work and under what conditions deep learning starts to fail. I am now heavily involved in pushing the boundary of robust vision and safe AI by creating high-realism degradation-mimetic adversarial attacks such as adversarial motion/defocus blur, rain/haze, relighting/exposure/vignetting/shadow, etc., and then devising tools to defend against them through an attempt to situate both the natural corruption robustness and the adversarial robustness problems under the same dome; creating cunning DeepFakes that are detection-evasive to foster the development of next-generation robust detectors of real-world video/audio DeepFakes; seeking robust learners by creating novel neural architectures and computational modules that are statistically efficient and adversarially robust.

My Ph.D. thesis (here) is primarily focused on tackling the Pose, Expression, Resolution, Illumination, and Occlusion (Perio) challenges for unconstrained periocular face recognition using both shallow and deep discriminative and generative methods, especially under the dome of self-supervised predictive learning. I also have broader interests in pattern recognition, computer vision, machine learning, optimization, statistics, compressive sensing, and image processing.

I am the recipient of multiple best/distinguished paper awards, including the Best Poster Paper Award of the IEEE/IAPR International Joint Conference on Biometrics (IJCB) in 2011, the Best Paper Award of the IEEE Seventh International Conference on Biometrics: Theory, Applications and Systems (BTAS) in 2015, the Best Student Paper Award of the IEEE Eighth International Conference on Biometrics: Theory, Applications and Systems (BTAS) in 2016, the ACM SIGSOFT Distinguished Paper Award of the IEEE/ACM International Conference on Automated Software Engineering (ASE) in 2018, and the Best Student Paper Award of the 14th Asian Conference on Computer Vision (ACCV) in 2018.

CVPR'24 EDGE Workshop

Call for Papers

prl
Latest News
Research Highlights & Preprints
prl

Movie Gen: A Cast of Media Foundation Models
Movie Gen team at Meta
arXiv Preprint, 2024
arxiv / bibtex / project / teaser video


This paper introduces research for the most advanced media foundation AI models.

prl

Imagine yourself: Tuning-Free Personalized Image Generation
Zecheng He*, Bo Sun*, Felix Juefei-Xu*, Haoyu Ma, Ankit Ramchandani, Vincent Cheung, Siddharth Shah, Anmol Kalia, Harihar Subramanyam, Alireza Zareian, Li Chen, Ankit Jain, Ning Zhang, Peizhao Zhang, Roshan Sumbaly, Peter Vajda, Animesh Sinha*
arXiv Preprint, 2024
arxiv / bibtex


This paper introduces a state-of-the-art tuning-free personalizaed image generation method.

prl

Pixel-Space Post-Training of Latent Diffusion Models
Christina Zhang, Simran Motwani, Matthew Yu, Ji Hou, Felix Juefei-Xu, Sam Tsai, Peter Vajda, Zijian He, Jialiang Wang
arXiv Preprint, 2024
arxiv / bibtex


This paper shows that adding a pixel-space objective significantly improves both supervised quality fine-tuning and preference-based post-training by a large margin.

prl

LUNA: A Model-Based Universal Analysis Framework for Large Language Models
Da Song, Xuan Xie, Jiayang Song, Derui Zhu, Yuheng Huang, Felix Juefei-Xu, Lei Ma
arXiv Preprint, 2023
arxiv / bibtex


This paper delves into the challenges and concerns associated with LLMs and introduces a universal analysis framework to tackle these issues.

prl

Look Before You Leap: An Exploratory Study of Uncertainty Measurement for Large Language Models
Yuheng Huang, Jiayang Song, Zhijie Wang, Shengming Zhao, Huaming Chen, Felix Juefei-Xu, Lei Ma
arXiv Preprint, 2023
arxiv / bibtex


This paper investigates blackbox techniques to understand the confidence of inference by LLMs.

prl

Among Us: Adversarially Robust Collaborative Perception by Consensus
Yiming Li, Qi Fang, Jiamu Bai, Siheng Chen, Felix Juefei-Xu, Chen Feng
IEEE International Conference on Computer Vision (ICCV), 2023
arxiv / bibtex


We improve the adversarial robustness on collaborative perception via multi-robot consensus.

prl

ALA: Adversarial Lightness Attack via Naturalness-aware Regularizations
Yihao Huang, Liangru Sun, Qing Guo, Felix Juefei-Xu, Jiayi Zhu, Jincao Feng, Yang Liu, Geguang Pu
ACM International Conference on Multimedia (ACM MM), 2023
arxiv / bibtex


To generate unrestricted adversarial examples with high image quality and good transferability, we propose adversarial lightness attack (ALA), a white-box unrestricted adversarial attack that focuses on modifying the lightness of the images.

prl

Zero-Day Backdoor Attack against Text-to-Image Diffusion Models via Personalization
Yihao Huang, Qing Guo, Felix Juefei-Xu
arXiv Preprint, 2023
arxiv / bibtex


This work exposes the backdoor vulnerability in text-to-image diffusion models via personalizaation methods.

prl

Common Corruption Robustness of Point Cloud Detectors: Benchmark and Enhancement
Shuangzhi Li, Zhijie Wang, Felix Juefei-Xu, Qing Guo, Xingyu Li, Lei Ma
IEEE Transactions on Multimedia (TMM), 2023
arxiv / bibtex


We present a comprehensive benchmark for the common corruption robustness aspect of point cloud detectors, together with the enhancement therewithin.

prl

TFormer: A Transmission-Friendly ViT Model for IoT Devices
Zhichao Lu, Chuntao Ding, Felix Juefei-Xu, Vishnu Naresh Boddeti, Shangguang Wang, Yun Yang
IEEE Transactions on Parallel and Distributed Systems (TPDS), 2022
arxiv / bibtex / proceedings


We present a transmission-friendly ViT model (TFormer) for deployment on resource-constrained IoT devices by virtue of the proposed hybrid layer that consists of non-learnable modules and pointwise convolutions, which can obtain multitype and multiscale features with only a few parameters and FLOPs.

prl

DARTSRepair: Core-Failure-Set Guided DARTS for Network Robustness to Common Corruptions
Xuhong Ren, Jianlang Chen, Felix Juefei-Xu, Wanli Xue, Qing Guo, Lei Ma, Jianjun Zhao, Shengyong Chen
Pattern Recognition (PR), 2022
arxiv / bibtex / proceedings


We propose a novel core-failure-set guided DARTS (differentiable architecture search) that embeds a K-center-greedy algorithm for DARTS to select suitable corrupted failure examples to refine the model architecture.

prl

Towards Transmission-Friendly and Robust CNN Models over Cloud and Device
Chuntao Ding, Zhichao Lu, Felix Juefei-Xu, Vishnu Naresh Boddeti, Yidong Li, Jiannong Cao
IEEE Transactions on Mobile Computing (TMC), 2022
arxiv / bibtex / proceedings


We propose a cloud-assisted CNN training framework with low model parameter transmission and strong model robustness.

prl

Countering Malicious DeepFakes: Survey, Battleground, and Horizon
Felix Juefei-Xu, Run Wang, Yihao Huang, Qing Guo, Lei Ma, Yang Liu
International Journal of Computer Vision (IJCV), 2022
arxiv / bibtex / project / proceedings


We present a comprehensive survey covering the latest DeepFake generation and detection methods, as well as the battleground landscape.

prl

Can You Spot the Chameleon? Adversarially Camouflaging Images from Co-Salient Object Detection
Ruijun Gao*, Qing Guo*, Felix Juefei-Xu, Hongkai Yu, Huazhu Fu, Wei Feng, Yang Liu, Song Wang
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2022
arxiv / bibtex


We present the very first black-box joint adversarial exposure and noise attack (Jadena) on co-salient object detection. By camouflaging images from co-salient object detection, a new layer of privacy protection can be achieved.

prl

Uncertainty-Aware Cascaded Dilation Filtering for High-Efficiency Deraining
Qing Guo, Jingyang Sun, Felix Juefei-Xu, Lei Ma, Di Lin, Wei Feng, Song Wang
arXiv Preprint, 2022
arxiv / bibtex


The proposed EfDeRain+ method makes contributions to specifically address the residual rain traces, multi-scale, and diverse rain patterns, without harming the efficiency.

prl

Masked Faces with Faced Masks
Jiayi Zhu, Qing Guo, Felix Juefei-Xu, Yihao Huang, Yang Liu, Geguang Pu
arXiv Preprint, 2022
arxiv / bibtex


We devise realistic facial masks that exhibit partial face patterns (i.e., faced masks) and stealthily add adversarial textures that can not only lead to significant performance deterioration of the SOTA deep learning-based FRS, but also remain undetected by the SOTA facial mask detector, thus successfully fooling both systems at the same time.

prl

Adversarial Rain Attack and Defensive Deraining for DNN Perception
Liming Zhai, Felix Juefei-Xu, Qing Guo, Xiaofei Xie, Lei Ma, Wei Feng, Shengchao Qin, Yang Liu
arXiv Preprint, 2022
arxiv / bibtex


A factor-aware rain generation process is proposed to synthesize realistic adversarial rain streaks that can attack image classification and object detection. A defensive deraining stretegy is also proposed to complete the story of rain-robust perception studies.

prl

NPC: Neuron Path Coverage via Characterizing Decision Logic of Deep Neural Networks
Xiaofei Xie, Tianlin Li, Jian Wang, Lei Ma, Qing Guo, Felix Juefei-Xu, Yang Liu
ACM Transactions on Software Engineering and Methodology (TOSEM), 2022
arXiv Preprint, 2022
arxiv / bibtex / proceedings


We investigate the interpretable coverage criteria through constructing the decision structure of a DNN. Based on the control flow and data flow of the decision graph, we propose two variants of path coverage to measure the adequacy of the test cases in exercising the the decision logic.

prl

AdvBokeh: Learning to Adversarially Defocus Blur
Yihao Huang, Felix Juefei-Xu, Qing Guo, Weikai Miao, Yang Liu, Geguang Pu
arXiv Preprint, 2021
arxiv / bibtex


We present a novel adversarial bokeh attack via a depth-guided bokeh synthesis network (DebsNet) that is able to adversarially and flexibly synthesize, refocus, and adjust the level of bokeh of the image.

prl

Benchmarking Shadow Removal for Facial Landmark Detection and Beyond
Lan Fu, Qing Guo, Felix Juefei-Xu, Hongkai Yu, Wei Feng, Yang Liu, Song Wang
arXiv Preprint, 2021
arxiv / bibtex


We construct a novel benchmark to link two independent yet related tasks, shadow removal and facial landmark detection with extensive evaluations.

prl

ArchRepair: Block-Level Architecture-Oriented Repairing for Deep Neural Networks
Hua Qi*, Zhijie Wang*, Qing Guo, Jianlang Chen, Felix Juefei-Xu, Lei Ma, Jianjun Zhao
arXiv Preprint, 2021
arxiv / bibtex


We initiate a robust architecture-oriented search-based DNN repairing by jointly optimizing the architecture and weights at the block level.

prl

Dodging DeepFake Detection via Implicit Spatial-Domain Notch Filtering
Yihao Huang, Felix Juefei-Xu, Qing Guo, Lei Ma, Xiaofei Xie, Weikai Miao, Yang Liu, Geguang Pu
arXiv Preprint, 2021
arxiv / bibtex


We investigate how a learning-based approach can reproduce the notch filtering but solely in the spatial domain to reconstruct the noise-free fake images that can evade the SOTA DeepFake detectors.

prl

FakeLocator: Robust Localization of GAN-Based Face Manipulations
Yihao Huang, Felix Juefei-Xu, Qing Guo, Yang Liu, Geguang Pu
IEEE Transactions on Information Forensics and Security (TIFS), 2022
arxiv / bibtex / proceedings


We present robust simultaneous detection and localization of GAN-based DeepFake manipulations.

prl

Pasadena: Perceptually Aware and Stealthy Adversarial Denoise Attack
Yupeng Cheng*, Qing Guo*, Felix Juefei-Xu, Xiaofei Xie, Shang-Wei Lin, Weisi Lin, Wei Feng, Yang Liu
IEEE Transactions on Multimedia (TMM), 2021
arxiv / bibtex / proceedings


We investigate a new task with seemingly confliting objectives, the adversarial denoise attack, that stealthily embeds the attacks inside the image denoising operation.

prl

Fooling LiDAR Perception via Adversarial Trajectory Perturbation
Yiming Li, Congcong Wen, Felix Juefei-Xu, Chen Feng
IEEE International Conference on Computer Vision (ICCV), 2021
[Oral Presentation]
arxiv / bibtex / project


We present that adversarial spoofing of a self-driving car's trajectory with small perturbations is enough to jeopardize the LiDAR perception.

prl

Learning to Adversarially Blur Visual Object Tracking
Qing Guo, Ziyi Cheng, Felix Juefei-Xu, Lei Ma, Xiaofei Xie, Yang Liu, Jianjun Zhao
IEEE International Conference on Computer Vision (ICCV), 2021
arxiv / bibtex


We present adversarial blur attack on visual object trackers by online transferring input frames to their natural (and adversarially) motion-blurred counterparts while misleading the SOTA trackers during the tracking process.

prl

JPGNet: Joint Predictive Filtering and Generative Network for Image Inpainting
Xiaoguang Li*, Qing Guo*, Felix Juefei-Xu, Hongkai Yu, Yang Liu, Song Wang
ACM International Conference on Multimedia (ACM MM), 2021
arxiv / bibtex


We explore the interplay between predictive filtering and generative network for high-fidelity image inpainting.

prl

FakeTagger: Robust Safeguards against DeepFake Dissemination via Provenance Tracking
Run Wang, Felix Juefei-Xu, Meng Luo, Yang Liu, Lina Wang
ACM International Conference on Multimedia (ACM MM), 2021
arxiv / bibtex


We investigate the potentials of image tagging in serving the DeepFake provenance tracking by robustly embedding message to the facial image that is recoverable after drastic DeepFake transformation, with high confidence.

prl

AdvFilter: Predictive Perturbation-aware Filtering against Adversarial Attack via Multi-domain Learning
Yihao Huang, Qing Guo, Felix Juefei-Xu, Lei Ma, Weikai Miao, Yang Liu, Geguang Pu
ACM International Conference on Multimedia (ACM MM), 2021
arxiv / bibtex


Jointly considering predictive perturbation-aware filtering and adversarial training begets more robust CNNs against adversarial attacks.

prl

Adversarial Relighting against Face Recognition
Ruijun Gao*, Qing Guo*, Qian Zhang, Felix Juefei-Xu, Hongkai Yu, Wei Feng
arXiv Preprint, 2021
arxiv / bibtex


We present a novel physical model-based adversarial relighting attack to fool the face recognition systems.

prl

CarveNet: Carving Point-Block for Complex 3D Shape Completion
Qing Guo*, Zhijie Wang*, Felix Juefei-Xu, Di Lin, Lei Ma, Wei Feng, Yang Liu
arXiv Preprint, 2021
arxiv / bibtex


We propose the CarveNet, a novel 3D point cloud block carving method for complex 3D point cloud completion, by conducting exclusive convolution on each point of the block.

prl

DeepMix: Online Auto Data Augmentation for Robust Visual Object Tracking
Ziyi Cheng*, Xuhong Ren*, Felix Juefei-Xu, Wanli Xue, Qing Guo, Lei Ma, Jianjun Zhao
IEEE International Conference on Multimedia and Expo (ICME), 2021
[Oral Presentation]
arxiv / bibtex / VOT2021 challenge (ICCV'21)


We propose a one-step online data augmentation scheme for robust visual object tracking.

prl

Bias Field Poses a Threat to DNN-Based X-Ray Recognition
Binyu Tian, Qing Guo, Felix Juefei-Xu, Wen Le Chan, Yupeng Cheng, Xiaohong Li, Xiaofei Xie, Shengchao Qin
IEEE International Conference on Multimedia and Expo (ICME), 2021
[Oral Presentation]
arxiv / bibtex


We reveal that the bias field in X-ray imaging can be smoothly and adversarially attacked, posing a threat to the DNN-based X-ray automated diagnosis.

prl

Sparta: Spatially Attentive and Adversarially Robust Activation
Qing Guo, Felix Juefei-Xu, Changqing Zhou, Yang Liu, Song Wang
arXiv Preprint, 2021
arxiv / bibtex


We extend ReLU to a novel Sparta activation function (Spatially Attentive and Adversarially Robust Activation), which enables CNNs to achieve both higher robustness and higher accuracy than the existing SOTA activation functions.

prl

DeepRepair: Style-Guided Repairing for Deep Neural Networks in the Real-World Operational Environment
Bing Yu, Hua Qi, Qing Guo, Felix Juefei-Xu, Xiaofei Xie, Lei Ma, Jianjun Zhao
IEEE Transactions on Reliability (TR), 2021
arxiv / bibtex / proceedings


We present a novel and reliable style-guided repairing for DNNs deployed in the real-world operational environment.

prl

AVA: Adversarial Vignetting Attack against Visual Recognition
Binyu Tian, Felix Juefei-Xu, Qing Guo, Xiaofei Xie, Xiaohong Li, Yang Liu
International Joint Conference on Artificial Intelligence (IJCAI), 2021
arxiv / bibtex


We present a novel degradation-mimetic adversarial attack by capitalizing the image vignetting effect.

prl

Let There be Light: Improved Traffic Surveillance via Detail Preserving Night-to-Day Transfer
Lan Fu, Hongkai Yu, Felix Juefei-Xu, Jinlong Li, Qing Guo, Song Wang
IEEE Transactions on Circuits and Systems for Video Technology (TCSVT), 2021
arxiv / bibtex / proceedings


We present a robust nighttime traffic surveillance method by incorporating translation based StyleMix and object detail preservation via kernel prediction network.

prl

AdvHaze: Adversarial Haze Attack
Ruijun Gao, Qing Guo, Felix Juefei-Xu, Hongkai Yu, Wei Feng
arXiv Preprint, 2021
arxiv / bibtex


We adversarially synthesize haze into an image based on the atmospheric scattering model with high realisticity to fool the classifier.

prl

Adversarial Exposure Attack on Diabetic Retinopathy Imagery
Yupeng Cheng, Felix Juefei-Xu, Qing Guo, Huazhu Fu, Xiaofei Xie, Shang-Wei Lin, Weisi Lin, Yang Liu
arXiv Preprint, 2020
arxiv / bibtex


We have investigated how slight and adversarial exposure change on diabetic retinopathy imagery can severely affect its grading.

prl

Auto-Exposure Fusion for Single-Image Shadow Removal
Lan Fu*, Changqing Zhou*, Qing Guo, Felix Juefei-Xu, Hongkai Yu, Wei Feng, Yang Liu, Song Wang
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2021
arxiv / bibtex / code


We have proposed an auto-exposure fusion that achieves the SOTA single-image shadow removal.

prl

EfficientDeRain: Learning Pixel-wise Dilation Filtering for High-Efficiency Single-Image Deraining
Qing Guo*, Jingyang Sun*, Felix Juefei-Xu, Lei Ma, Xiaofei Xie, Wei Feng, Yang Liu
AAAI Conference on Artificial Intelligence (AAAI), 2021
arxiv / bibtex / code / poster


We have proposed the fastest single-image deraining method with the SOTA performance.

prl

Watch out! Motion is Blurring the Vision of Your Deep Neural Networks
Qing Guo, Felix Juefei-Xu, Xiaofei Xie, Lei Ma, Jian Wang, Bing Yu, Wei Feng, Yang Liu
Advances in Neural Information Processing Systems (NeurIPS), 2020
arxiv / bibtex / code / poster / proceedings


We newly identify an attacking method termed motion-based adversarial blur attack (ABBA) that can generate visually natural motion-blurred adversarial examples.

prl

Breaking Neural Reasoning Architectures with Metamorphic Relation-Based Adversarial Examples
Alvin Chan, Lei Ma, Felix Juefei-Xu, Yew Soon Ong, Xiaofei Xie, Minhui Xue, Yang Liu
IEEE Transactions on Neural Networks and Learning Systems (TNNLS), 2021
arxiv / bibtex / proceedings


First adversarial attack on differentiable neural computer.

prl

SPARK: Spatial-Aware Online Incremental Attack Against Visual Tracking
Qing Guo*, Xiaofei Xie*, Felix Juefei-Xu, Lei Ma, Zhongguo Li, Wanli Xue, Wei Feng, Yang Liu
European Conference on Computer Vision (ECCV), 2020
arxiv / bibtex / code / slides / proceedings


We propose a novel adversarial attack on visual tracking: online generating imperceptible perturbations that mislead trackers along an incorrect (untargeted attack) or specified trajectory (targeted attack).

prl

Amora: Black-box Adversarial Morphing Attack
Run Wang, Felix Juefei-Xu, Qing Guo, Yihao Huang, Xiaofei Xie, Lei Ma, Yang Liu
ACM International Conference on Multimedia (ACM MM), 2020
[Oral Presentation]
arxiv / bibtex / proceedings


In contrast to adversarial noise attack that perturbs pixel intensity values by adding human-imperceptible noise, our proposed adversarial morphing attack is a semantic attack that perturbs pixels spatially in a coherent manner.

prl

FakePolisher: Making DeepFakes More Detection-Evasive by Shallow Reconstruction
Yihao Huang, Felix Juefei-Xu, Run Wang, Qing Guo, Lei Ma, Xiaofei Xie, Jianwei Li, Weikai Miao, Yang Liu, Geguang Pu
ACM International Conference on Multimedia (ACM MM), 2020
[Oral Presentation]
arxiv / bibtex / slides / proceedings


We have successfully shown that simple shallow reconstruction can very effectively reduce the artifacts introduced during DeepFake image synthesis, making it more detection-evasive.

prl

DeepSonar: Towards Effective and Robust Detection of AI-Synthesized Fake Voices
Run Wang, Felix Juefei-Xu, Yihao Huang, Qing Guo, Xiaofei Xie, Lei Ma, Yang Liu
ACM International Conference on Multimedia (ACM MM), 2020
[Oral Presentation]
arxiv / bibtex / proceedings


This is the first attempt to robustly detect AI-synthesized fake voices based on monitoring neural behaviors.

prl

DeepRhythm: Exposing DeepFakes with Attentional Visual Heartbeat Rhythms
Hua Qi*, Qing Guo*, Felix Juefei-Xu, Xiaofei Xie, Lei Ma, Wei Feng, Yang Liu, Jianjun Zhao
ACM International Conference on Multimedia (ACM MM), 2020
arxiv / bibtex / slides / proceedings


This is the first attempt to robustly detect DeepFakes based on monitoring remote visual photoplethysmography (PPG) signals with the aid of dual-spatial-temporal attention mechanism.

prl

FakeSpotter: A Simple yet Robust Baseline for Spotting AI-Synthesized Fake Faces
Run Wang, Felix Juefei-Xu, Lei Ma, Xiaofei Xie, Yihao Huang, Jian Wang, Yang Liu
International Joint Conference on Artificial Intelligence (IJCAI), 2020
arxiv / bibtex / proceedings

Media Coverage: Synced Review


Monitoring neuron behavior can serve as an asset in detecting AI-synthesized fake faces since layer-by-layer neuron activation patterns may capture more subtle features that are important for the fake detector.

prl

DeepHunter: A Coverage-Guided Fuzz Testing Framework for Deep Neural Networks
Xiaofei Xie, Lei Ma, Felix Juefei-Xu, Minhui Xue, Hongxu Chen, Yang Liu, Jianjun Zhao, Bo Li, Jianxiong Yin, Simon See
ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA), 2019
arxiv / bibtex / proceedings


An automated fuzz testing framework for hunting potential defects of general-purpose DNNs. It performs metamorphic mutation to generate new semantically preserved tests, and leverages multiple plugable coverage criteria as feedback to guide the test generation.

prl

SSR2: Sparse Signal Recovery for Single-Image Super-Resolution on Faces with Extreme Low Resolutions
Ramzi Abiantun*, Felix Juefei-Xu*, Utsav Prabhu, Marios Savvides
Pattern Recognition, 2019
bibtex / proceedings


Extreme low-resolution face super-resolution based on traditional methods and its application to a real-world scenario.

prl

Secure Deep Learning Engineering: A Software Quality Assurance Perspective
Lei Ma, Felix Juefei-Xu, Minhui Xue, Qiang Hu, Sen Chen, Bo Li, Yang Liu, Jianjun Zhao, Jianxiong Yin, Simon See
arXiv Preprint, 2018
arxiv / bibtex


From a software quality assurance perspective, this work pinpoints challenges and future opportunities towards universal secure deep learning engineering.

prl

RankGAN: A Maximum Margin Ranking GAN for Generating Faces
Felix Juefei-Xu*, Rahul Dey*, Vishnu Naresh Boddeti, Marios Savvides
Asian Conference on Computer Vision (ACCV), 2018
[Oral Presentation][Best Student Paper Award]
arxiv / supplementary / bibtex


Progressive training of multiple GAN stages with maximum margin ranking loss for generating faces with improved fidelity.

prl

DeepCT: Tomographic Combinatorial Testing for Deep Learning Systems
Lei Ma, Felix Juefei-Xu, Minhui Xue, Bo Li, Li Li, Yang Liu, Jianjun Zhao
IEEE International Conference on Software Analysis, Evolution and Reengineering (SANER), 2019
bibtex


A set of combinatorial testing criteria specialized for deep learning systems, as well as a combinatorial testing coverage guided test generation technique.

DeepMutation: Mutation Testing of Deep Learning Systems
Lei Ma, Fuyuan Zhang, Jiyuan Sun, Minhui Xue, Bo Li, Felix Juefei-Xu, Chao Xie, Li Li, Yang Liu, Jianjun Zhao, Yadong Wang
IEEE International Symposium on Software Reliability Engineering (ISSRE), 2018
arxiv / bibtex


A mutation testing framework specialized for deep learning systems to measure the quality of test data.

DeepGauge: Multi-Granularity Testing Criteria for Deep Learning Systems
Lei Ma, Felix Juefei-Xu, Fuyuan Zhang, Jiyuan Sun, Minhui Xue, Bo Li, Chunyang Chen, Ting Su, Li Li, Yang Liu, Jianjun Zhao, Yadong Wang
ACM/IEEE International Conference on Automated Software Engineering (ASE), 2018
[Oral Presentation][ACM SIGSOFT Distinguished Paper Award]
arxiv / bibtex


A set of multi-granularity testing criteria for deep learning systems, which aims at rendering a multi-faceted portrayal of the testbed.

prl

Perturbative Neural Networks
Felix Juefei-Xu, Vishnu Naresh Boddeti, Marios Savvides
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018
arxiv / project / bibtex


The new perturbation layer does away with convolution in the traditional sense and instead computes its response as a weighted linear combination of non-linearly activated additive noise perturbed inputs.

prl

Local Binary Convolutional Neural Networks
Felix Juefei-Xu, Vishnu Naresh Boddeti, Marios Savvides
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017
[Spotlight Oral Presentation]
arxiv / project / bibtex


Convolutional neural network architecture with fixed randomized sparse binary convolutional kernels.

prl

Gang of GANs: Generative Adversarial Networks with Maximum Margin Ranking
Felix Juefei-Xu, Vishnu Naresh Boddeti, Marios Savvides
arXiv Preprint, 2017
arxiv / project / bibtex


A progressive training paradigm involving multiple GANs to contribute to the maximum margin ranking loss so that the GAN at later stages will improve upon early stages.

Dissertation

Unconstrained Periocular Face Recognition: From Reconstructive Dictionary Learning to Generative Deep Learning and Beyond
Felix Juefei-Xu
Advisor: Marios Savvides
Ph.D. Dissertation, Carnegie Mellon University, 2018
bibtex

Older Publications

DeepGender2: A Generative Approach Toward Occlusion and Low Resolution Robust Facial Gender Classification via Progressively Trained Attention Shift Convolutional Neural Networks (PTAS-CNN) and Deep Convolutional Generative Adversarial Networks (DCGAN)
Felix Juefei-Xu, Eshan Verma, Marios Savvides
Book Chapter, Deep Learning for Biometrics, Springer, 2017
bibtex

Learning to Invert Local Binary Patterns
Felix Juefei-Xu, Marios Savvides
British Machine Vision Conference (BMVC), 2016
bibtex

Fastfood Dictionary Learning for Periocular-Based Full Face Hallucination
Felix Juefei-Xu, Marios Savvides
IEEE 8th International Conference on Biometrics: Theory, Applications and Systems (BTAS), 2016
[Oral Presentation][Best Student Paper Award]
bibtex

Simultaneous Forgery Identification and Localization in Paintings Using Advanced Correlation Filters
Paul Buchana*, Irina Cazan*, Manuel Diaz-Granados*, Felix Juefei-Xu, Marios Savvides
IEEE International Conference on Image Processing (ICIP), 2016
[Oral Presentation]
bibtex

Discriminative Invariant Kernel Features: A Bells-and-Whistles-Free Approach to Unsupervised Face Recognition and Pose Estimation
Dipan K. Pal, Felix Juefei-Xu, Marios Savvides
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016
[Spotlight Oral Presentation]
slides / poster / bibtex

DeepGender: Occlusion and Low Resolution Robust Facial Gender Classification via Progressively Trained Convolutional Neural Networks with Attention
Felix Juefei-Xu*, Eshan Verma*, Parag Goel, Anisha Cherodian, Marios Savvides
Biometrics Workshop, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016
[Oral Presentation]
slides / poster / bibtex

Multi-class Fukunaga Koontz Discriminant Analysis for Enhanced Face Recognition
Felix Juefei-Xu, Marios Savvides
Pattern Recognition, 2016
bibtex / proceedings

Spartans: Single-sample Periocular-based Alignment-robust Recognition Technique Applied to Non-frontal Scenarios
Felix Juefei-Xu, Khoa Luu, Marios Savvides
IEEE Transactions on Image Processing (TIP), 2015
bibtex / proceedings

Pokerface: Partial Order Keeping and Energy Repressing Method for Extreme Face Illumination Normalization
Felix Juefei-Xu, Marios Savvides
IEEE 7th International Conference on Biometrics: Theory, Applications and Systems (BTAS), 2015
[Oral Presentation][Best Paper Award]
slides / bibtex

NIR-VIS Heterogeneous Face Recognition via Cross-Spectral Joint Dictionary Learning and Reconstruction
Felix Juefei-Xu, Dipan K. Pal, Marios Savvides
Perception Beyond the Visual Spectrum (PBVS) Workshop, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015
bibtex

Driver Cell Phone Usage Detection on Strategic Highway Research Program (SHRP2) Face View Videos
Keshav Seshadri, Felix Juefei-Xu, Dipan K. Pal, Marios Savvides
Computer Vision in Vehicle Technology (CVVT) Workshop, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015
bibtex

Electromyograph and Keystroke Dynamics for Spoof-Resistant Biometric Authentication
Shreyas Venugopalan*, Felix Juefei-Xu*, Benjamin Cowley, Marios Savvides
Biometrics Workshop, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015
bibtex

Hallucinating the Full Face from the Periocular Region via Dimensionally Weighted K-SVD
Felix Juefei-Xu, Dipan K. Pal, Marios Savvides
Biometrics Workshop, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2014
bibtex

Subspace Based Discrete Transform Encoded Local Binary Patterns Representations for Robust Periocular Matching on NIST's Face Recognition Grand Challenge
Felix Juefei-Xu, Marios Savvides
IEEE Transactions on Image Processing (TIP), 2014
bibtex / proceedings

Miscellaneous
 
 

I like this page :-)