12th Int'l Symposium on Image and Signal Processing and Analysis (ISPA 2021)

13-15 September 2021, Zagreb, Croatia

Special Sessions

On this page, you can see the list of ISPA 2021 special sessions and their scopes.

Interested authors are invited to submit papers for this special session through the ISPA online submission system. When submitting a paper indicate this special session in the online submission form.

Papers submitted to special sessions must adhere to the instructions for authors and will be subject to peer review. Unless otherwise stated, submissions should adhere to conference deadlines. Accepted special session papers will be published in the Symposium Proceedings in the same length and format as regular ISPA papers.

If you would like to organize a special session at ISPA 2021 please see the Call for Special Session Proposals.

Color vision and processing: where are we at?

Aim and scope: "In the last 30 years, image processing has seen many trends, roughly speaking: the PDE-based formalization of classical algorithms, variational methods, patch-based models and, nowadays, deep learning. Despite huge improvements in research fields such as, e.g., denoising, segmentation or inpainting, color image processing still seems to elude optimal treatment. Little or none improvement in the understanding of colorimetric attributes has been achieved, and many color algorithms simply perform separated computations in the RGB channels. In this section we aim at discussing both at highlighting problems and discussing solution proposals related to any aspect of color processing, e.g. color correction, enhancement and transfer, white balance, HDR imaging and tone mapping, to quote but a few."

Organiser: Edoardo Provenzi, professor at the University of Bordeaux, Institute of Mathematics, edoardo.provenzi@math.u-bordeaux.fr

Analyzing nondestructive evaluation data. Is it automated yet?

Aim and Scope


Nondestructive evaluation (NDE) of materials is a highly demanding task that accepts only the perfect results. While computer vision and signal processing methods have recently been thriving in numerous industries, NDE data analysis is still largely performed manually. However, a rather large group of scientists researches this topic on a quest to develop novel methods that could transform this area.


The goal of the special session is to discuss and share novel ideas regarding the automation of NDE data analysis. Proposals can be related to any nondestructive evaluation method of inspection. Topics of interest include, but are not limited to:

  • defect detection on images or signals

  • anomaly detection

  • synthetic data generation

  • data enhancement


Review, Proceedings and Submission

Submission deadline is June 13th 20th.

Organizers

  • Duje Medak, University of Zagreb, Faculty of Electrical engineering and computing, duje.medak@fer.hr

  • Luka Posilović, University of Zagreb, Faculty of Electrical engineering and computing, luka.posilovic@fer.hr

3D Surface Imaging and Structured Light Scanning

Aim and Scope

Modern optical 3D surface scanners including structured light scanners can do more than ever before with less size, weight, power and price, but often at the cost of increased system complexity. Most of the proposed scanning solutions are used exclusively in the lab and have yet to be tested in real world applications where their performance may be seriously affected by challenging scanning conditions. Therefore, there is a need to improve current 3D scanners and make them simultaneously simpler and more robust via incremental improvements. On the other hand machine learning, as one of the main artificial intelligence field driving forces, is attracting considerable interest and may be applied to 3D surface imaging and structured light scanning to achieve more robust performance outside of the lab.

The goal of this special session is to bring together researchers from the fields of computer vision, artificial intelligence and optics to facilitate diffusion of ideas and foster future research in this interdisciplinary area. Topics of interest include, but are not limited to:

  • robust signal processing methods for 3D surface imaging and structured light scanning particularly under extreme environmental conditions, for example, direct sunlight or underwater scanning

  • machine learning and artificial intelligence for 3D surface imaging and structured light scanning

  • optimal coding functions for structured light scanning

  • structured light scanning of moving objects

  • 3D imaging and structured light solutions on smartphones and tablets

Review, Proceedings and Submission

Submission deadline is June 20th.

Organizers

Advances in Machine Learning for Finance

Aim and Scope

With the development of financial technology and growth of available financial data, signal processing methods and machine learning tools have become increasingly important in many financial use cases. The need for robust and scalable methods in financial technology, risk management and portfolio optimization calls for interdisciplinary engineering approaches to the problems of modern finance.

The goal of this special session is to bring together researchers from the disciplines of signal processing, machine learning, mathematics, and finance to facilitate diffusion of ideas and foster future research. Topics of interest include, but are not limited to:

  • Signal processing methods for financial time series

  • Machine learning for financial risk modelling

  • Deep learning and reinforcement learning for financial data

  • Portfolio optimization algorithms and asset allocation

  • Fintech innovation and blockchain-based assets

Review, Proceedings and Submission

Submission deadline is June 20th.

Organizers

  • Zvonko Kostanjčar, University of Zagreb, zvonko.kostanjcar@fer.hr

  • Stjepan Begušić, University of Zagreb, stjepan.begusic@fer.hr

Combinatorial Image Processing and Digital Geometry

Aim and Scope

There are various ways to work with digital images, several of them use some combinatorial techniques. On the other hand, digital geometry is a theoretical foundation of digital image processing and computer graphics. Since computers use a discrete underlying grid for images (usually the square grid), some geometrical properties differ from the properties of Euclidean geometry.

This field includes theoretical aspects and applications of:

  • digital distances (e.g., distances based on neighbourhood sequences, chamfer distances),

  • discrete and binary tomography,

  • image generation (e.g., 2D grammars and languages, cell automata),

  • mathematical morphology, etc.

The aim of this special session is to gather researchers of these particular fields, exchange their ideas and develop further collaborations.

Review, Proceedings and Submission
Submission deadline is June 13th 28th.

Organizers

  • Péter Balázs, University of Szeged, pbalazs@inf.u-szeged.hu

  • Benedek Nagy, Eastern Mediterranean University, nbenedek.inf@gmail.com

Image and signal processing in education

Aim and Scope

The special session is targeted to the application of image and signal processing in learning, teaching and education, as well as teaching image and signal processing.

Topics of interest include (but are not limited to):

  • learning analytics,

  • 3D modeling in education,

  • animation and 3D systems,

  • VR / AR in teaching,

  • visual programming tools,

  • gamification,

  • mobile and tablet technologies in education,

  • computer vision in educational robotics,

  • computer vision in competitions,

  • assessment and evaluation of robotics – enhanced class activities,

  • evaluation criteria and tools for measuring the impact of robotics / eLearning on students’ learning,

  • methods for teaching image and signal processing.

Review, Proceedings and Submission

Submission deadline is May 25th.

Organizer

  • Ana Sović Kržić, University of Zagreb, ana.sovic.krzic@fer.hr