Keynote speakers

• Laura Leal-Taixé
• Professor, Technical University of Munich



• Prof. Laura Leal-Taixé is leading the Dynamic Vision and Learning group at the Technical University of Munich, Germany. She received her Bachelor and Master degrees in Telecommunications Engineering from the Technical University of Catalonia (UPC), Barcelona. She did her Master Thesis at Northeastern University, Boston, USA and received her PhD degree (Dr.-Ing.) from the Leibniz University Hannover, Germany.

  During her PhD she did a one-year visit at the Vision Lab at the University of Michigan, USA. She also spent two years as a postdoc at ETH Zurich, Switzerland and one year at the Technical University of Munich. In 2017, she won the Sofja Kovalevskaja Award of 1.65 million euros from the presitgious Humboldt Foundation for her project "socialMaps".



• Dynamic Scene Understanding - With our without time?

• If you ask any student nowadays what tools to use to solve an image recognition task, the most popular answer will be Deep Learning. But the real-world is not static but rather dynamic, and therefore better represented by videos than by still images. It is still an interesting open question how to deal with the temporal redundancy of video frames: shall a Neural Network exploit it or ignore it?

  In this talk, I will explore both strategies, i.e., to actually exploit the redundancy in the content of nearby frames, or to ignore it.

  In the first work on multiple object tracking, I show how to obtain temporally coherent results while not using temporal information during training. This also alleviates the need for full video annotations. From our recent work on video super resolution, I will present our new temporal discriminator which works in a Generative Adversarial Network training scheme in order to create temporally coherent image details.

• Lourdes Agapito
• Professor, University College London



• Lourdes Agapito holds the position of Professor of 3D Vision in the Department of Computer Science at University College London (UCL). Her research in Computer Vision has focused on the inference of 3D information from the video acquired from a single moving camera. While early research focused on static scenes, attention soon turned to the much more challenging problem of estimating the 3D shape of non-rigid objects (Non-Rigid Structure from Motion, NR-SFM) or complex dynamic scenes where an unknown number of objects might be moving, possibly deforming, independently. Prof. Agapito's research group investigates all theoretical and practical aspects of NRSFM: deformable tracking; dense optical flow estimation and non-rigid video registration; 3D reconstruction of deformable and articulated structure and dense 3D modelling of non-rigid dynamic scenes.

• Fred Hamprecht
• Professor, Heidelberg University
• Visiting Professor, Uppsala University



• Fred Hamprecht holds the Robert-Bosch endowed professorship for Image Analysis and Learning at Heidelberg University. His research interests lie in image processing and machine learning. His main focus is on the development of algorithms to solve interesting problems from the life sciences. Major applications include the tracing of all neurites in a brain, the tracking of all cells in a developing embryo, and quantitative analysis of high-throughput experiments. The group puts particular emphasis on the user-friendly training of such systems, and is actively developing open source libraries and programs such as ilastik. Prof. Hamprecht studied and earned his PhD at the Swiss Federal Institute of Technology (ETH), and became a Professor for Multidimensional Image Processing at Heidelberg University in 2001. He is a co-founder of the Heidelberg Collaboratory for Image Processing (HCI).



• Signed graph partitioning: an important computer vision primitive

• Perennial computer vision problems such as image partitioning, instance segmentation or tracking can be reduced to combinatorial graph partitioning problems.

  The majority of models developed in this context have relied on purely attractive interactions between graph nodes. To obtain more than a single cluster, it is then necessary to pre-specify a desired number of clusters, or set thresholds.

  A notable exception to the above is multicut partitioning / correlation clustering, which accommodates repulsive in addition to attractive interactions, and which automatically determines an optimal number of clusters. Unfortunately, the multicut problem is NP-hard.

  In this talk, I will characterize the combinatorial problem and discuss its representations in terms of node or edge labelings. I will discuss greedy algorithms that find approximate solutions, or even exact ones under certain conditions. One algorithm I will discuss is the "mutex watershed" which currently gives the best results on a connectomics challenge.

  Joint work with Steffen Wolf, Constantin Pape, Nasim Rahaman, Alberto Bailoni, Ullrich Koethe, Anna Kreshuk.

The main topics of SCIA 2019 include:

Call for papers

Call for papers can be downloaded here.

How to submit a paper to SCIA 2019

The papers can be submitted by following this link. Instructions for the submission will appear in the submission system.

Paper preparation

The maximum allowed length for a paper is 12 pages (single column), including references and appendices (if any). Paper submissions are anonymous, i.e. authors should not write their names and affiliations in the submitted paper.

Papers can be submitted in LaTeX or Word format:

• LateX2e: we recommend the use of LaTeX2e for the preparation of your camera-ready manuscript, together with the corresponding Springer class file. The ZIP archive with the LaTeX template and relative instructions can be downloaded here.
• Word: we do not encourage the use of Microsoft Word, particularly as the layout of the pages (the position of figures and paragraphs) can change between printouts. However, we do provide the relevant template. Please read the explanatory typing instructions "SPLNPROC Word 2007-2010 Technical Instructions.pdf" contained in the ZIP archive carefully. The template can be downloaded here.
• Word 2003: A document template has been prepared by Springer for use with Word 2003. Predefined style formats are available for all the types of content that are part of a computer science proceedings paper, and these formats can be easily accessed via special toolbars. The ZIP archive with the template can be downloaded here.

Camera ready submission

Papers accepted for presentation must be resubmitted by April 3 in the submission system. The resubmission must consists of 3 different files:

1. the source files (LaTeX or Word) of the revised version of the manuscript. If your manuscript is written in LaTeX, you must ensure to provide all the figures, bibliography and additional manuscript files necessary to compile the final PDF. All the files should be included into a single ZIP file;
2. a separate file containing the answers to the reviewer comments, explaining how you addressed their critics and suggestions in the revised version of the manuscript;
3. a signed version (in PDF) of Springer copyright, available for download at the link below.

Copyright submission

Authors whose paper has been accepted for presentation in SCIA 2019 must download and compile the Springer copyright form, necessary for the publication in the Springer proceedings. The copyright form must be compiled as follows:

• Title of the Book or Conference Name: SCIA 2019.
• Volume Editor(s): Jonas Unger, Michael Felsberg, Per-Erik Forssén, Ida-Maria Sintorn
• Title of the Contribution: title of the paper submitted to SCIA 2019.
• Author(s) Name(s): complete names of all the authors of the paper submitted to SCIA 2019.
• Corresponding Author's Name, Address, Affiliation and Email: information of the corresponding author of the paper submitted to SCIA 2019.

The copyright form must be signed by the corresponding author of the paper and the three boxes in the end must be checked, only if one of the categories apply. Once the form is filled, it must be scanned and converted ina PDF file, which has to be sent along with the files required for the camera ready version of the manuscipt (see above), before April 3.

The copyright form can be downloaded here.

Presentation

For the oral paper presentation, authors are asked to prepare a set of slides for a talk of 15-17 minutes.

Original research and double submissions

The research papers must neither been published nor submitted for publication elsewhere. Preprints (such as ArXiv) are accepted.

Proceedings

Papers accepted for the SCIA 2019 conference will be published in Lecture Notes in Computer Science Proceedings, Springer. All figures will be in black and white in the printed publication.

Registration to SCIA 2019 and SSDL is now open. Early bird registration is available until April 7.

Register to SCIA 2019, SSDL or SCIA 2019 + SSDL here.

Registration fees

SCIA 2019 will be held at Norrköping Visualization Center C at Linköping University, Campus Norrköping located in the old industrial landscape close to Norrköping city center:

Accomodation and travel

Norrköping offers a variety of hotels within walking distance from the conference venue. Book hotel at conference rate here.

Traveling to Norrköping from outside of Sweden is easy. The closest airport is in the nearby city of Linköping with flights from and to the Schiphol airport in Amsterdam. Linköping airport is accessed to and from Norrköping by taxi or by taking a shuttle or bus to Linköping train station. The local train is operated by Östgötatrafiken leaves from Linköping Resecentrum to Norrköping Resecentrum 3-4 times every hour. Another option is Skavsta airport which operates Ryanair flights from several destinations in Europe and is accessed by frequent airport busses to and from Norrköping. The Arlanda airport in Stockholm can be reached from most major airports and is accessed by several daily direct trains to and from Norrköping.

Social events

The SCIA 2019 conference reception on Tuesday June 11 will be held at the Visualization Center and include a mingle and visits to the public exhibitions and the dome theatre. The conference dinner in the evening on Wednesday June 12 will be held at in Östgöta kök's restaurant at Nya Torget close to Norrköping city center.

Directions to the conference dinner:

Norrköping

Norrköping is located approximately 1.5 hours south of Stockholm. From the 17th century to the middle of the 20th century, it was the place of a number of large industries located along the Motala river that floats through the city. The city was a large producer of textiles, and has often been nicknamed "Sweden's Manchester". Today, the former industries have been restorated and transformed into a unique environment that hosts a number of innovative companies, museums, conference center, the Visualization Center C, as well as the Norrköping Campus of Linköping University.

For more information about Norrköping, see for example the experience Norrköping website.

New for this year is that the sponsorship packages cover both SCIA and SSDL. Our sponsorship packages for SCIA and SSDL offer a unique opportunity to:

1. Network with the leading research groups in the Nordic countries
2. Gain visibility before, during, and after the conference
3. Meet and recruit undergraduate students, graduated or soon to graduate PhD students

For more information and booking, please contact Daniel Jönsson.

General Chair

Program Chairs

Web Chair

Sponsors Chair

International program committee