Advanced Perception

Efficient Long-Form Video Understanding: Scaling Perception, Selecting Evidence, and Adapting Computation

Shuming Liu, Ph.D. Student, Computer Science

Nov 18, 19:00 - 21:00

Zoom Meeting 95938664864

scene understanding Advanced Perception reasoning Efficient AI visual computing multimodal models

This dissertation advances efficient long-form video understanding by developing methods that allocate limited computational resources across three stages: scaling video perception, selecting relevant visual evidence, and adapting reasoning effort to individual questions.

Advancing Monocular Depth Estimation: High-Resolution and Amodal Understanding

Zhenyu Li, Ph.D. Student, Computer Science

Nov 17, 15:00 - 17:00

B1 R2202

depth estimation geometric modeling computational imaging visual computing 3D reconstruction scene understanding Advanced Perception

This thesis advances monocular depth estimation to develop efficient methods for recovering fine geometric details, such as object boundaries and thin structures, from high-resolution images while preserving scene-wide consistency and metric scale.

Zhenyu Li

Ph.D. Student, Computer Science

3D reconstruction scene understanding depth estimation generative ai visual computing Advanced Perception

Zhenyu Li is a final-year Ph.D. candidate in Computer Science at King Abdullah University of Science and Technology (KAUST), where he is advised by Prof. Peter Wonka, specializing in computer vision, 3D vision, and generative AI, with particular emphasis on monocular depth estimation, high resolution 3D perception, and scene reconstruction.

Mostafa Kassem

Visiting Student, Cyber Security and Resilience Technology

Automous Vehicles Advanced Perception

Adonai Vera

Intern, Image and Video Understanding Lab

Computer Vision Advanced Perception multimodal AI deep learning

Machine Learning Engineer with 6+ years of experience specializing in computer vision, deep learning, and robotics. Led AI solutions for retail analytics, UAV systems, and edge devices. Author, educator, and TensorFlow Google Developer Expert.