-
January 10, 2024
Paper/Abstract submission & Registration opens
-
January 30, 2024
Early-bird registration deadline
-
February 12, 2024
Paper submission deadline
-
March 5, 2024
Late registration deadline
-
March 11, 2024 (9-lunch)
PhD student day
-
March 11-13, 2024
Conference
-
about SSBA/SSDL 2024
We celebrate the 41st anniversary of the Swedish Symposium on Image Analysis (SSBA) this year. It is co-located with the 7th Swedish Symposium on Deep Learning (SSDL). The event takes place at Luleå University of Technology, Luleå, Sweden, which is home to the spectacular Northern Lights.
The SSBA symposium is the premier Swedish event where researchers, industrial professionals and students gather to learn about the recent developments in the areas of image processing, computer vision, pattern recognition and related fields. SSBA 2024 will feature keynote speakers, and oral presentations and posters of submitted papers.
SSDL is an important Swedish forum for leading research groups in industry and academia to meet and discuss the latest trends and developments in deep learning and related areas. SSDL 2024 will feature invited talks by leading researchers in deep learning as well as oral and poster presentations of submitted papers and abstracts.
Program
Keynote: Historical Document Analysis and the Humanities
Elisa Barney
Professor , Luleå University of Technology
Elisa Barney Smith is a professor at Luleå University of Technology (LTU). She received her Ph.D. in Electrical, Computer and Systems Engineering from Rensselaer Polytechnic Institute, Troy, NY, USA. Professor Barney’s main research interests are image processing and machine learning. She applies these primarily to document imaging. Here is her homepage
Keynote: XAI for representation learning
AbstractThe field of eXplainable AI (XAI) has received tremendous attention in recent years. Most work in the XAI literature focus on providing input attributions in some form. In computer vision, for instance, attribution maps are created by methods that aim to highlight the pixels that were relevant for a certain image classification by e.g. a neural network. However, basically nobody has tried to explain representations. Representation learning, for instance via self-supervised learning, is key to any machine learning system. This talk will outline a new proposed framework called RELAX to provide XAI for representation learning in computer vision.
Robert Jenssen
Director, Visual Intelligence
Robert Jenssen is a Professor at UiT, Machine Learning Group and the Director, Visual Intelligence. Visual Intelligence is a Centre for Research-based Innovation (SFI) funded by the Research Council of Norway and a consortium of private and public partners. We are at the international forefront in deep learning research for complex image analysis. Here is his homepage
Keynote: "beyond 100%" - Open research in Document Analysis
AbstractDocument image analysis has a long research history and thus is considered as a mature research field. In fact, recent OCR (optical character reader) systems can indeed recognize various document images, such as scene texts and decorated font images, with very high accuracy, because of deep learning technologies and large-scale datasets. It should be emphasized that this situation does not indicate the shrinkage of the research. Instead, it indicates that we can start new research by utilizing the highly-accurate OCR systems and deep learning technologies. In my talk, I will introduce several new research trials that use the textual information collected by the recent OCRs as well as trials that are first enabled by deep learning technologies.
Seiichi Uchida
Professor (Distinguished Professor of Kyushu University)
Seiichi Uchida received B.E. and M.E. and Dr. Eng. degrees from Kyushu University in 1990, 1992 and 1999, respectively. From 1992 to 1996, he joined SECOM Co., Ltd., Japan. Currently, he is a distinguished professor at Kyushu University, Japan. His research interests include pattern recognition and image processing. He received 2007 IAPR/ICDAR Best Paper Award and many international and domestic awards. Here is his homepage
Keynote: Diagnosis and prognostication in medical imaging using AI
AbstractThe field of medical image analysis is making great strides in the era of deep learning (DL), with a wide range of problems being addressed using such techniques. Two considerable limitations to the use of DL in medical imaging is the lack of annotated data needed for supervised learning, and the oftentimes low level of explainability and missing uncertainty estimation of DL predictions. In my presentation, I will talk about the use of weakly labeled data for cancer diagnosis and prognostication, and ways to improve prognostication with new survival modelling approaches. Furthermore, I will talk about the combination of DL and rule-based approaches for explainability and uncertainty estimation in the detection of bone structures in medical images.
Ida Häggström
Associate Professor, Signal Processing and Biomedical Engineering, Electrical Engineering
Ida's research is focused on medical image analysis using machine learning techniques. She collaborates closely with medical doctors on projects to diagnose, predict and prognosticate different diseases, mainly cancer. She works mainly with images from positron emission tomography (PET), but also computed tomography (CT) and magnetic resonance imaging (MR). Here is her homepage
| Schedule | ||||
|---|---|---|---|---|
| Practical Info: Please use a dark/black background for your presentation slides, as study has shown this to be better for the eyes. | ||||
| Call for contributions by mail: (1) Short tutorial submissions - suggest a topic you would like to give a short (20mins) tutorial on for the PhD-student day.
(2) Bring your own problem – pitch short abstracts for 3 minutes about a problem or method your would like to share with the SSBA/SSDL audience. Deadline: 28 February. Please, use the same template as for abstract submissions below (or Word) and mail it to the above address or ssdl2024@ssba.org.se. |
||||
| Monday, March 11 | Location | |||
| 7.30 | Registration | LKAB (A117) | ||
| 8.00 - 10.00 | PhD student day Chair: Marcus Liwicki |
LKAB (A117) | ||
| 10.00 - 10.30 | Fika break (only for PhD students) | |||
| 10.30 - 11.00 | PhD student day Chair: Marcus Liwicki |
Short tutorials | LKAB (A117) | |
| 11.00 - 12.00 | PhD student day | Visit to the robotics lab and mingle | ||
| 12.00 - 13.00 | Lunch (and SSBA board meeting - A110) | Stuk (C hus) | ||
| 13.00 - 13.20 | Opening: Marcus Liwicki | LKAB (A117) | ||
| 13.20 - 14.20 | Keynote: Seiichi Uchida Chair: Marcus Liwicki |
"beyond 100%" - Open research in Document Analysis | LKAB (A117) | |
| 14.20 - 14.40 | Bring-Your-Own-Problem Pitches Chair: Marcus Liwicki |
Beyond traditional >> Spherical neurons >> Automatic production quality >> Reconstructing Hadron |
LKAB (A117) | |
| 14.40 - 15.00 | Fika break | |||
| 15.00 - 15.40 | SSDL Pitches: Cut-and-Paste >> Generative Models >> Instruction >> EmergAI >> Deep Ontology >> Generative AI >> Cloth-Splatting >> Exploring Temporal >> Intermodal >> GMSF >> Enhancing >> In Defense >> Enabling Diffusion Chair: Tosin Adewumi |
|||
| 15.40 - 16.40 | SSDL Poster session | LKAB corridor | ||
| 17.30 | Light dinner starts | LKAB corridor | ||
| 18.00 - 19.00 | SSBA Annual General Meeting | Start of SSBA | LKAB (A117) | |
| Tuesday, March 12 | ||||
| 8.00 - 9.00 | Keynote: Ida Häggström Chair: Tosin Adewumi |
Diagnosis and prognostication in medical imaging using AI | LKAB (A117) | |
| 9.00 - 10.15 |
SSBA oral session 1 Chair: György Kovács |
(1) 9.00 - 9.25: Computing the Lipschitz constant needed for fast scene recovery from CASSI measurements (2) 9.25 - 9.50: Learned Trajectory Embedding for Subspace Clustering (3) 9.50 - 10.15: MOoSE: Multi-Orientation Sharing Experts for Open-set Scene Text Recognition |
LKAB (A117) | |
| 10.15 - 10.30 | Fika break | |||
| 10.30 - 12.00 | SSBA oral session 2 Chair: Simon Corbillé |
(4) 10.30 - 11.00: Navigating Seasonal Changes: Neural Radiance Fields for Satellite Image-Based 3D Reconstruction (5) 11.00 - 11.30: Adaptive Sampling for BRDF Acquisition (6) 11.30 - 12.00: Steerers |
LKAB (A117) | |
| 12.00 - 13.00 | Lunch | Stuk (C hus) | ||
| 13.00 - 14.00 | Keynote: Elisa Barney Chair: Homam Mokayed |
Historical Document Analysis and the Humanities | LKAB (A117) | |
| 14.00 - 14.20 | Short Talk: George Nikolakopoulos Chair: Homam Mokayed | Robotics and AI group in Luleå | LKAB (A117) | |
| 14.20 - 14.40 | Bring-Your-Own-Problem Pitches Chair: Homam Mokayed |
Interoperable Visual Localization >> Denoising atomically | LKAB (A117) | |
| 14.40 - 15.00 | Fika break | |||
| 15.00 - 18.00 | Social event | Underground mining lab in Mjölkluddsberget & Teknikenshus |
||
| 18.00 - 21.00 | Dinner & Award (Most Industry-Relevant Paper) |
Teknikenshus | ||
| Wednesday, March 13 | ||||
| 8.00 - 9.00 | Keynote: Robert Jenssen Chair: Tommaso Dorigo |
XAI for representation learning | LKAB (A117) | |
| 9.00 - 10.15 | SSBA oral session 3 Chair: Chang Liu |
(7) 9.00 - 9.25: Robust AI based prostate cancer grading exceeding human performance (8) 9.25 - 9.50: Comparing Diffusion Models and GANs in Synthesizing Brain MRI and Chest X-Ray Images: A Focus on Memorization Risks (9) 9.50 - 10.15: Recognising Swedish Handwritten Stenography |
LKAB (A117) | |
| 10.15 - 10.30 | Fika break | |||
| 10.30 - 12.00 | SSBA oral session 4 Chair: Yushan Zhang |
(10) 10.30 - 11.00: Geometry from a Single Motion Blurred Image (11) 11.00 - 11.30: Classifying femur fractures using federated learning (12) 11.30 - 12.00: Federated training of segmentation models for radiotherapy treatment planning |
LKAB (A117) | |
| 12.00 - 13.00 | Lunch | Stuk (C hus) | ||
| 13.00 - 14.40 | Industry session | Panel discussion Chair: Marcus Liwicki | The progress of AI adoption in Sweden, Challenges and Next teps | LKAB (A117) | |
| The PhD student day includes a guided tour to the Robotics lab and is the perfect oppurtunity for PhD students to network and get to know peers within the community. | ||||
Proceedings
You can download the proceedings (based on authors' permission) here .
Important Dates
Submissions
We invite researchers from academy and industry to contribute to our joint symposium. The symposium is meant to be a snapshot of interesting activity in Sweden, and thus we welcome previously published material, and material that you plan to publish elsewhere. Contributions to SSDL and SSBA will not be made openly available online.
SSBA submissions
SSBA is open for (the usual) short papers (up to 6 pages, including references) on any topic in image analysis, computer vision or pattern recognition. Contributions will primarily be presented orally.A template with paper formatting is available here.
SSBA short papers are submitted by email, to: ssba2024@ssba.org.se
SSDL submissions
SSDL is open for extended abstracts (2-3 pages, including references) on any topic of interest within AI, machine learning or deep learning. The contributions will primarily be presented as posters (portrait format with a maximum width of 80 cm). A number of the submitted contributions will be chosen for oral presentation alongside the keynote talks at the symposium.Please use the same template as for the SSBA short papers.
SSDL extended abstracts are submitted by email, to: ssdl2024@ssba.org.se
Registration
Please, register in advance. There will be no registration on site.
Registration fees (excluding VAT, includes lunch, fika, social event & dinner on Tuesday)
| Type | Early bird (deadline January 30) | Regular registration (deadline March 5) |
|---|---|---|
| Members | 2000 SEK | 2400 SEK |
| Students | 1500 SEK | 1900 SEK |
Prices are excluding VAT.
General registration
Venue
SSBA/SSDL 2024 will be held at Luleå University of Technology (LTU), Luleå.
Click on the map to search for a specific room. House A is the location.
Practical Information and Public Transportation
It is very easy to plan your journey and transportation in Luleå. LLT has an app called journey planner and there you can plan and book your bus tickets. Tickets must be purchased before boarding the bus. You can travel with a bus pass or a mobile ticket. Types of tickets: 72-hour ticket is the perfect choice for when you're just visiting Luleå for a few days. This pass will allow you to take unlimited trips for 72 hours.
Bus numbers 4, 5, 6 and 7 get to LTU from the city center. Bus 4 gets to LTU from the airport. These buses stop at "Universitetsentren" (the closest stop), "Porsön" and "Kårhusvägen", which are all within 2 minutes walk to the location in the University. You may download the LLT app to buy the ticket of your choice.
Attractions in Luleå
Norra hamn, ice breaker, Gammaelstad Church Town. For further information, you may refer to this link
Sponsoring
A. Exhibitor, SEK 20,000
- Demonstration and exhibition space at your disposal.
- Free registration for 1 company representative (incl. accommodation).
B. Silver sponsorship, SEK 15,000
- Your logo on the symposium website.
- Public acknowledgement of your support during the conference.
- Your company is mentioned with a logo in the conference program sheet.
C. Gold sponsorship, SEK 30,000
- Your logo on the symposium website.
- Public acknowledgement of your support during the conference.
- Your company is mentioned with a logo in the conference program sheet.
- Demonstration and exhibition space at your disposal.
- Free registration for 1 company representative (incl. accommodation).
- You will have the opportunity to contribute material to our goodie bags.
- Guaranteed time for presentation during the industry session.