COMPROMISE project webinar A UNIVERSAL DATA COMPRESSION PARADIGM BASED ON PREDICTIONS AND DATA RESTORATION

The COMPROMISE project hosted a successful webinar on September 23th, titled “A Universal Data Compression Paradigm Based on Predictions and Data Restoration.” The event attracted over 60 participants from esteemed institutions across both national and European levels, highlighting the growing interest and relevance of data compression technologies. 

The webinar featured a series of expert-led presentations that explored novel methods and advances in data compression, spanning 1D, 2D, and 3D data. Topics included near-lossless compression techniques for EEG signals, image compression utilizing machine learning, and advanced methods for voxel and surface geometry compression. 

Notably, the event also attracted attention from industry representatives, with several engaging in discussions and raising questions that pointed to the practical implications of these new technologies. 

PROGRAM

10:00 Prof. dr. Borut Žalik: Overview of basic principles of data compression 

10:20 dr. David Podgorelec: Challenges, trends, and new paradigms in data compression 

10:40 assist. prof. dr. Josef Kohout: Near-lossless multi-channel EEG signal compression 

11:00 Luka Lukač: Machine learning for image compression 

11:20 assist prof. dr. Bogdan Lipuš: Preliminary research on image color compression 

11:40 assist prof. dr. Libor Váša: Efficient encoding of 3D triangle meshes 

12:00 Blaž Repnik: Voxelized surface representation with chain codes 

12:20 assoc. prof. dr. Damjan Strnad: Predictive coding and decoding of sparse voxel grids