Program
| Sunday, 03 July 2022 | Monday, 04 July 2022 | Tuesday, 05 July 2022 |
| 08:00 Registration and Preparation | 08:30 Registration and Preparation | |
| 09:00 Welcome and Keynote | 09:00 Session 6: Image Coding | |
| 10:00 Session 1: Video Compression | 10:20 Coffee Break and Demos | |
| 10:40 Coffee Break and Demos | 11:00 Session 7: Film Look | |
| 11:20 Session 2: SQoE | 11:40 Session 8: 360° Video and Point Clouds | |
| 12:00 Session 3: VVC Codec | 12:20 Summary and Closure | |
| 13:00 Lunch Break at Mensa | 13:00 Networking Lunch at ML2 | |
| 14:00 Session 4: Image and Networks | ||
| 15:20 Poster Session + Coffee break | Departure | |
| Arrival | 16:20 Session 5: Video | |
| 17:45 Networking Hike to Gabelbach | ||
| 19:00 Pre-Event at BC club | 19:45 Social Event at Gabelbach |
Pre-School Event
The pre-school event will be held on Sunday, 03 July 2022 at BC-club (student club of Ilmenau).
Location. Starting time: 19:00
You will receive a Welcome Snack from the barbecue and will get free entry to the club’s Sunday event. This event is for PhD students and others to meet and greet and exchange before the actual summer school starts.
Keynote/Invited Lecture
| Speaker: Christian Feldmann |  |
| Title: Video Coding in the Cloud – Challenges and Solutions | |
| Abstract: After the rise of video streaming over the internet (OTT) more and more workflows for video transcoding and distribution are moving to the cloud. While this combination can be a match made in heaven, it is full of challenges like a multitude of different video and codecs, streaming container formats, subtitle formats, etc. The talk will give an overview of the transcoding landscape and the biggest challenges that we and our customers are facing. We will talk about codecs and device compatibility, cloud infrastructure, encoders and how some of the biggest names in streaming are trying to solve these problems. | |
| About Christian Feldmann: He is a video coding engineer at Bitmovin working on next generation video coding technologies in the most recent video coding standards, such as HEVC, VVC and AV1. His main focus points at Bitmovin are practical application of video and transport standards in the scalable cloud transcoding solution. After his study of computer engineering at RWTH University Aachen, he completed his doctoral degree (PhD) at the Institut für Nachrichtentechnik (Institute for Communication Technologies) in Aachen. With his detailed experience in video coding, he is developing video coding technologies for the future of video coding. Christian participates in the standardization activities of the Moving Picture Experts Group (MPEG). |
Social Event
It is planned to have a 2 hour networking hike towards Hotel Gabelbach right after Session 5 on Monday. There, an included joint dinner will take place.
Session 1: Video Compression Performance
- Christian Herglotz (FAU Erlangen-Nürnberg): Beyond Bjøntegaard: Limits of Video Compression Performance Comparisons
- Sophie Pientka (Fraunhofer HHI): Rate-Distortion-Optimization for Deep Image Compression
Session 2: Sustainable QoE
- Matthias Kränzler (FAU Erlangen-Nürnberg): Energy Efficient Video Decoding for VVC Using a Greedy Strategy Based Design Space Exploration
- Geetha Ramasubbu (FAU Erlangen-Nürnberg): Investigation of the Impacts of Video and Encoding Parameters on the Encoding Energy of the HEVC Software Encoding Process
Session 3: Usage of the VVC Codec
- Nicolas Horst (RWTH Aachen): New Reference Subpicture Padding Techniques for Use with Independently Decodable Subpictures
- Martin Benjak (Leibniz Universität Hannover): Neural Network-based Error Concealment for B-Frames in VVC
- Adam Wieckowski (Fraunhofer HHI): Encoding complexity analysis through the measurement of the search space
Session 4: Image and Networks
- Simon Grosche (FAU Erlangen-Nürnberg): Applications of Non-Regular Image Sampling using Locally Fully Connected Reconstruction Network
- Frank Sippel (FAU Erlangen-Nürnberg): Camera Array for Multispectral Imaging
- Konstantin Schmidt (Fraunhofer IIS): Comparing Generative Adversarial Networks and Normalizing Flow for Super-Resolution
- Ulrike Pestel-Schiller (Leibniz Universität Hannover): Determination of Relevant Hyperspectral Bands using a Spectrally Constrained CNN
Session 5: Video
- Goluck Konuko (CentraleSupélec Paris): Ultra-low Bitrate Video Compression using Deep Image Animation
- Marta Milovanovic (Telecom Paris): Hybrid Method for Patch-depth Transmission in MPEG Immersive Video
- Smitha Lingadahalli Ravi (INSA Rennes): DIBR for Immersive Videos: Traditional Depth Estimators versus Learning-Based Depth Estimators
Session 6: Image Coding
- Maximiliane Gruber (FAU Erlangen-Nürnberg): Domain Adaptation for Unknown Image Distortions
- Shabhrish Reddy Uddehal (Leipzig University of Telecommunications): Image Segmentation for Improved Lossless Screen Content Compression
- Marwa Tarchouli (INSA Rennes): Patch-based Image Coding with End-to-End Learned Codec using Overlapping
- Daniel Gritzner (Leibniz Universität Hannover): Using Binary Space Partitioning for the Semantic Segmentation of Aerial Images
Session 7: Film Look
- Dominik Keller (Technische Universität Ilmenau): Application of sensory evaluation techniques in film - How do expert and non-expert evaluations differ?
- Zoubida Ameur (INSA Rennes): Deep-learning based film grain generation and removal
Session 8: 360° Video and Point Clouds
- Viktoria Heimann (FAU Erlangen-Nürnberg): Point Cloud Upsampling Using Frequency-Based Models
- Andy Regensky (FAU Erlangen-Nürnberg): Motion-Plane-Adaptive Inter Prediction in 360-Degree Video Coding
Poster Session
- Fabian Brand (FAU Erlangen-Nürnberg): Rate-Distortion Optimized Deep Image Compression
- Kristian Fischer (FAU Erlangen-Nürnberg): Boosting Neural Image Compression for Machines Using Latent Space Masking
- Thibaut Meyer (RWTH Aachen): Point Cloud Attribute Coding using Graph Representations
- Melan Vijayaratnam (Telecom Paris): Towards zero latency video transmission through frame extrapolation
- Dominik Mehlem (RWTH Aachen): Entropy Coding of Neural Network Latent Space Coefficients for Point Clouds
- Tomasz Konaszynski (AGH Krakow): Analysis of context factors affecting human perceptions of 360º video quality
- Ashutosh Singla (Technische Universität Ilmenau): Holistic Assessment of 360° Video QoE
Demos
Hearing in Mixed Reality
A current trend in research and development is the extension of reality by adding artificial, virtual objects to realize mixed reality and augmented reality. The lab tour deals with methods and possibilities of measurement and auralization for the creation of virtual audio objects in the real listening environment. An audio system is presented in a practical way, in which all essential components for audio augmented reality are included, from recording to signal analysis and signal processing to playback. In addition to the headphone-based dynamic binaural audio reproduction, an (ambisonic-based) 22.2 loudspeaker reproduction will be presented, which gives an insight into the current loudspeaker-based sound reproduction of e.g. cinemas.
IVE for educational research
The Immersive Virtual Environments (IVE) presented in this demo are developed at TU Ilmenau to investigate the suitability of audio-visual IVEs for a “real-world” assessment of the influence of the visuospatial and acoustic environment on cognitive performance of adults and children in classroom-type environments. There are two different visual representations: 360° video and a CGI model, while we will compare both visualizations with regard to the influence on quality, listening effort in noise and role of spatial audio for different paradigms. The demo is part of a DFG-funded project within the SPP 2236 AUDICTIVE.
Co-Presence for Seniors
An overview of the goals of the CO-HUMANICS project funded by the Carl-Zeiss-Stiftung (www.co-humanics.de) and its current research stage will be given, focusing on the factors impacting QoE in AR/MR and robot-based telepresence.
Immersive Audiovisual Communication - Meeting in VR
IVEs promise to be an ideal medium for performing and studying group communication in a variety of situations. In an ongoing DFG project (as part of SPP 2236 AUDICTIVE), we aim to identify and gain a robust understanding of factors influencing plausibility and experience in multi-party audiovisual communication and interactive shared virtual environments. In this context, we assess plausibility, presence, quality as well as Quality of Experience (QoE) and develop methods for evaluation of such immersive systems. This demo shows a VR communication system that represents the basis for user studies conducted in our audiovisual test labs.
Mobile Groovability: How well does your phone groove?
In an industry-funded project from 2019 - 2021, we investigated the impact of smartphone acoustic reproduction on the perceived groove of music. Groove in this context is the human experience that motivates dance movements such as rhythmic nodding or foot tapping. In the present test, 8 different smartphones can be directly compared with each other based on various pieces of music.
HD studio direction in action
We are happy to show our fully functional HD studio suitable for offline and live production.
Student film productions
We encourage students to produce short films, movies, music videos, and audio productions in our labs. With regard to audio, voice-overs can be recorded and media productions can be edited. Some students projects are shown.
4K HDR Reference Screen
A professional reference monitor offers 4K HDR image quality, accuracy and consistency, making it perfect for demanding on-set, studio and post-production applications, including 4K digital film production. The capabilities are demonstrated with a few test sequences.