6. – 10. December 2021 from 14:00 – 16:00 (CET) the Norwegian University of Science and Technology (NTNU) will host the CapsNetwork Training School 1 in Zoom. The programme will be updated.
|Monday 6th||14:00||Image Analysis in the Sphere of Capsule Endoscopy: Insights and Challenges||Dimitris Iakovidis |
Professor, deputy head of the department of Computer Science and Biomedical Informatics, director of the Biomedical Imaging Laboratory at the University of Thessaly.
|Monday 6th||15:00||Omnidirectional vision and applications to robotic navigation||Helder Araujo |
Professor at the Department of Electrical and Computer Engineering, University of Coimbra.
|Tuesday 7th||14:00 – 15:00||Evaluating the Possibility for 3D Reconstruction of GI-tract using Virtual Capsule Simulator (VR-CAPS)||Pål Anders Floor |
Researcher at the Department of Computer Science at the Norwegian University of Science and Technology.
|Wednesday 8th||14:00 – 16:00||What could go wrong? – Murphy’s law in the age of deep learning||David Völgyes |
Senior Scientific Software Developer at Science [&] Technology AS.
|Thursday 9th||14:00||Deep learning for color: fundamentals and applications||Simone Bianco |
Associate Professor at the Department of Informatics, Systems and Communication at the University of Milan-Bicocca.
|Thursday 9th||15:00||Deep learning for color: image-to-image color enhancement||Marco Buzzelli |
Postdoc at Department of Informatics, Systems and Communication at the University of Milan-Bicocca.
|Friday 10th||14:00 – 16:00||Computational Visual Perception for Image and Video Quality Enhancement and Assessment in the Era of Deep Learning||Azeddine Beghdadi |
Professor at the Laboratory of Information Processing and Transmission at Université Sorbonne Paris Nord.
Abstracts and Bios for the lectures
|Image Analysis in the Sphere of Capsule Endoscopy: Insights and Challenges |
An overview of the research performed in the field of image analysis, focusing on methods and techniques proposed in the context of capsule endoscopy. The significant role of color is highlighted. Insights and challenges are identified based on a 20-year research experience in the field, covering topics that include abnormality detection, localization, and measurement.
|Dr. Dimitris K. Iakovidis received his BSc in Physics in 1997, his MSc in Cybernetics in 2001, and his PhD in Informatics in 2004, all from the University of Athens in Greece. Today, he is a Professor of Signal Processing and Medical Decision Support Systems at the Dept. of Computer Science and Biomedical Informatics of the University of Thessaly in Greece. His research interests include signal and image processing, decision support systems, intelligent systems and applications. In this context he has co-authored over 170 papers in international journals, conferences, and books, many of which address endoscopic image processing and analysis.|
|Evaluating the Possibility for 3D Reconstruction of GI-tract using Virtual Capsule Simulator (VR-CAPS)|
Construction of 3D models based on capsule images can provide enhanced viewing for Gastroenterologists. As the quality of capsule images is very low the problem is challenging.
Recently, a virtual capsule endoscope environment has been developed that can be helpful when constructing post processing algorithms for capsules in general.
In this talk Pål Anders will present some of the challenges faced in 3D reconstruction for capsule endoscopy, then introduce an existing virtual capsule endoscope simulator, VR-CAPS, that can greatly help in solving this task.
|Pål Anders Floor is a researcher at the Department of Computer Science at the Norwegian University of Science and Technology.|
|What could go wrong? – Murphy’s law in the age of deep learning|
Deep Learning is the state of the art for object classification and semantic segmentation. There are stellar results for ImageNet,
medical challenges, satellite image analysis, etc. But Deep Learning is a black box. What could go wrong? What are the common mistakes and how can you avoid them? The lecture begins with a quick overview of
fundamentals, afterwards typical pitfalls will be analyzed through examples.
|David Völgyes obtained MSc degree as physicist at Eötvös Loránd University (Budapest, Hungary) in 2008, and he obtained PhD in Computer Science at NTNU in 2018.|
He was a postdoc at University of Oslo between 2018 and 2020 focusing on machine learning for medical image analysis. Currently he is a senior scientific software engineer and machine learning expert at Science & Technology AS, Oslo. He is also a guest researcher at the University of Oslo.
|Deep learning for color: fundamentals and applications|
This lecture will cover the fundamentals of deep learning, providing the terminology and concepts necessary to understand a wide variety of applications to the domain of color.
Example applications will include regression for image quality assessment, and multi-regression for color constancy.
|Simone Bianco obtained the PhD in Computer Science at DISCo (Dipartimento di Informatica, Sistemistica e Comunicazione) of the University of Milano-Bicocca, Italy, in 2010. He obtained the BSc and the MSc degree in Mathematics from the University of Milano-Bicocca, Italy, respectively in 2003 and 2006. He is currently Associate Professor and his research interests include computer vision, machine learning, optimization algorithms, and color imaging.|
|Deep learning for color: image-to-image color enhancement|
This lecture will cover the topic of image-to-image translation, focusing on color-oriented applications presented at different levels of detail.
The analyzed techniques will include general-purpose and domain-specific generative adversarial networks, and image enhancement with color curves.
|Marco Buzzelli obtained his Bachelor Degree and Master Degree in Computer Science respectively in 2012 and 2014, focusing on Image Processing and Computer Vision tasks. He got his PhD in Computer Science in 2019 at the University of Milano – Bicocca (Italy), where he is currently employed as a post-doctoral researcher. His main topics of research include characterization of digital imaging devices, and object recognition in complex scenes.|