Siwei Lyu

Siwei Lyu is affiliated with the University at Buffalo, State University of New York in the United States. Their research has primarily focused on areas within computer science, especially in computer vision and pattern recognition. Lyu's scholarly contributions encompass topics such as generative adversarial networks and image synthesis, digital media forensic detection, face recognition and analysis, adversarial robustness in machine learning, anomaly detection techniques and applications, advanced neural network applications, and advanced image processing techniques.

The scientist has authored numerous papers, with significant publications including:

• "UA-DETRAC: A new benchmark and protocol for multi-object detection and tracking" (2020) published in Computer Vision and Image Understanding
• "Anti-Forensics for Face Swapping Videos via Adversarial Training" (2021) published in IEEE Transactions on Multimedia
• "Eyes Tell All: Irregular Pupil Shapes Reveal GAN-Generated Faces" (2022) presented at ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
• "DE-GAN: Domain Embedded GAN for High Quality Face Image Inpainting" (2021) published in Pattern Recognition
• "Fusing Global and Local Features for Generalized AI-Synthesized Image Detection" (2022) presented at the 2022 IEEE International Conference on Image Processing (ICIP)

Lyu has collaborated frequently with several researchers, including Shu Hu, Yuezun Li, Ming-Ching Chang, Honggang Qi, and Xin Wang. These collaborations have resulted in multiple co-authored works.

Their publications are often found in leading venues such as arXiv (Cornell University), Computer Vision and Image Understanding, the IEEE International Conference on Image Processing (ICIP), the Proceedings of the AAAI Conference on Artificial Intelligence, and IEEE Transactions on Multimedia.

The academic fields covered by Lyu's work include:

• Computer Science

Within this broader field, the scientist focuses on subfields such as:

• Computer Vision and Pattern Recognition
• Artificial Intelligence
• Signal Processing
• Sociology and Political Science
• Control and Systems Engineering

Siwei Lyu's work addresses methods and challenges related to image and video analysis, forensic detection of synthetic media, and the application of adversarial training techniques in machine learning. The research efforts also include creating benchmarks and protocols for multi-object detection and tracking, as well as developing models for high-quality image inpainting.