As deep learning is becoming a more and more integral part of many companies’ workflows and people’s everyday life, the need for data increases exponentially. Moreover, as more specialized tasks are automated through AI, the required data is becoming more specific and general open-source datasets do not contain the needed information in large enough quantities. Deep learning tasks such as object detection, segmentation, depth estimation, anomaly detection, among others greatly benefit from transfer learning and additional tuning on specialized data.
The problem is that a lot of the time gathering specialized data can prove costly and time-consuming, as creating such datasets can take many months and even years and requires the purchase and installation of costly hardware. In other cases, capturing footage can be deemed outright dangerous or impossible without the help of specialists or in very small intervals of time.
This leaves a large gap in how such data can be captured, with current research focusing on modeling 3D digital twins and using them to capture data, augmenting synthetic elements in real datasets or using generative networks to synthesize the required information. These research fields perfectly encapsulate the modern need for combining knowledge from deep learning, computer vision, computer graphics and interactive systems, to create functional and useful systems for synthesizing and augmenting such data.
This parallel session will aim to bring together researchers and PhD students working in the intersection between computer vision, deep learning and computer graphics. By showcasing their current work through a series of short talks we aim to garner a better understanding of the work done in this field in Denmark and establish the possibility for collaborations and knowledge sharing. We hope to be able to establish a network of researchers from universities and companies interested in or working in the fields of synthetic data generation and augmentation in Denmark.
Researchers that would like to showcase their work should send their application to Ivan Nikolov (email@example.com) containing:
- Name and affiliation
- Area of research and work
- Topic of the proposed presentation
- Up to 1 page abstract describing the topic of the presentation.
- Deadline is 14.01.2024.
- The parallel session will include a 30-minute keynote speech by Dr. Barry Norton, VP of Research at Milestone Systems, explaining the company’s experience with synthetic data from a responsible tech perspective. Dennis Schou Jørgensen, Principle Software Engineer at Milestone Systems, will give a demo of these works.
- 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. A call for participation will be sent out through relevant channels. Depending on the number of participants, each person will be given 10 minutes to present their work plus 5 minutes for questions.
- The session will end with 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. The time constraints dictate that there will be maybe time for 3-4 talks of 10-15 minutes, plus a possible 30-minute invited talk. Other than that, no maximum caps of participant listeners are needed, as people can come to listen in on the talks.
- Dr. Barry Norton, VP of Research at Milestone Systems
- Dennis Schou Jørgensen, Principle Software Engineer at Milestone Systems
- Chosen researchers who will present their research topic in 10 to 15 minutes as part of the second part of the session.
Kamal Nasrollahi – Director of Research, Milestone Systems/ Professor of Computer Vision and Machine Learning, Aalborg University – firstname.lastname@example.org
Ivan Nikolov – (Main organizer) – Assistant Professor of Computer Graphics and Computer Vision, Aalborg University – email@example.com