About Abbas Cheddad
Abbas Cheddad is an Associate Professor of Computer Science at the University of Tartu, Estonia, and
Head of the Visual and Data Analytics Laboratory (VIDAL). His research focuses on computer vision,
artificial intelligence, machine learning, and pattern recognition, with particular emphasis on
developing computational methods for solving challenging real-world problems. These methods have
been successfully applied to medical imaging, industrial AI, digital heritage, scientific imaging, image
forensics, and intelligent data analytics.
He has over 18 years of international academic and industrial research experience, having held
academic appointments in Estonia, Sweden, and the United Kingdom. His research has been conducted
in close collaboration with major industrial partners, including Sony Mobile Communications,
Ericsson, GKN Aerospace, Axis Communications, ArkivDigital, and VITO, alongside an extensive
network of international academic collaborators.
Recent Comments
Given the promising results of the HCR framework for single drop regions, have you considered incorporating advanced sequence modeling techniques, such as Transformer architectures, to improve the reconstruction quality for multiple drop regions and longer gaps?
Thank you for your thoughtful response. Indeed, it could be possible for the SOTA Transformers architecture to capture better inferences from noisy and/or incomplete data. Thus, utilizing Transformer decoders for the reconstruction phase in conjunction with the introduced HCR (halftone-based compression and reconstruction) could prove beneficial to generate coherent outputs. However, for applications with limited computational resources, such as mobile apps, LSTM may still be competitive.
In our paper, we have set aside the exploration of these advanced techniques for future work.