Ismael Seidel
Universidade Federal de Santa Catarina (Brazil)
Coding the fourth dimension: Compressing light fields with JPEG Pleno
Language: Portuguese
Date: October 6th, 13h30
Abstract
Light fields are a plenoptic image modality that extend traditional 2D imaging by capturing both spatial and angular information, effectively introducing a fourth dimension to visual data. This additional dimensionality enables a range of advanced applications, including immersive media, post-capture refocusing, and enhanced low-light imaging. This talk introduces the fundamentals of light field imaging and presents the JPEG Pleno Light Field Coding solution, the first International Standard for light field compression. The discussion covers key aspects of the standardization process, including the design of the reference software, the conformance testing framework, and the current status within the JPEG committee. In addition to an overview of the standard, the talk highlights ongoing research directions and recent developments in compression techniques that contribute to the evolution of JPEG Pleno Light Field Coding standard.
Biography
Ismael Seidel received the B.S., M.S., and Ph.D. degrees in computer science from the Federal University of Santa Catarina (UFSC), Florianópolis, Brazil, in 2011, 2014, and 2019, respectively. From 2009 to 2019, he served as a Research Assistant at the Embedded Computing Laboratory (ECL) at UFSC. From 2019 to 2022, he served as a Research Scientist at the Samsung R&D Institute Brazil (SRBR) in Campinas, Brazil. Since 2022, he has been a professor in the Department of Informatics and Statistics at UFSC and has rejoined the ECL as a researcher. He is also a member of the Brazilian delegation in the ISO/IEC JTC1/SC29/WG1 (JPEG) Standardization Committee, where he actively participates in the JPEG Pleno light field coding activity, having served as coeditor of ISO/IEC 21794-3 (conformance testing) and ISO/IEC 21794-4 (reference software) standards. His research interests include visual signal processing and coding, as well as related low-power hardware architectures.