From Generative Models to Game Engines
Brief Description:
After the successful version 1.0 of the Synthetic Data Generation Session at the D3A conference in February 2024, we would like to continue by introducing a version 2.0. With this version we would like to focus on synthetic data generation from both game engines, but also in the use of generative models both for creating synthetic images, but also for generating tabular data, annotations, 3D models, point clouds, etc.
We saw a lot of interest in synthetic data generation for medical use cases, for surface inspection, for damage detection, for surveillance, among others. We also saw a lot of interest in creating synthetic data for very specific use cases, like vehicle and ship tracking, as well as the use of Nerfs and Gaussian Splatting for generation of novel synthetic views. This demonstrated that researchers and companies in Denmark and the Nordics are interested in the limitless possibilities coming from generating their own custom datasets and cutting the time from ideation to deployment.
In version 2 of the session, we will aim to expand the demonstrated use cases and bring even more researchers together with interested company representatives. We would also like to give Ph.D. students, postdocs and researchers working with synthetic data generation and those needing customized solutions a venue to discuss and build connections.
We invite researchers interested, ESPECIALLY ones that will also present their work in the poster session, to give a short talk around their work, interesting insights and workflows. Please send their application to Ivan Nikolov (iani@create.aau.dk) containing:
- Name and affiliation
- Area of research and work
- A short description of the topic of the proposed presentation
- Deadline is 01.10.2024.
Session Format:
- Session starts with a 30-minute keynote talk by an expert in synthetic data creation and integration for deep learning from Milestone System
- The second part of the parallel session will be separated into 10–15-minute presentations from participants in the D3A conference that are interested in sharing their research in the field.
- Depending on the number of participants, each person will be given 10-15 minutes to present their work plus 5 minutes for questions.
- The session will end will free time for discussion between guests and participants and the possibility to form new networking opportunities.
Target Audience and Size:
The target audience are researchers and PhDs both from the industry and university that are using or generating synthetic data, as well as people interested in utilizing synthetic data to solve problems that they are facing like lack of specialized datasets, not enough data or data variations.
We envision that the session will be of interest to both deep learning and computer vision specialists, as well as researchers working with computer graphics, interactive applications and even game development. We will have to post an invitation to researchers in Denmark working with generating and using synthetic data to present their work.
Speakers:
- Peter Damm, Director of Applied Intelligence, Milestone – “Trustworthy AI with GraphRAG and synthetic embedding generation”
- Chosen researchers that will present their research topic in 10 to 15 minutes as part of the second part of the session.
Organizers:
Ivan Nikolov – (Main organizer) – Assistant Professor of Computer Graphics and Computer Vision, Aalborg University – iani@create.aau.dk
Kamal Nasrollahi – Director of Research, Milestone Systems/ Professor of Computer Vision and Machine Learning, Aalborg University – kn@create.aau.dk
