The 26th International Symposium on Multimedia (ISM) was held from December 11-13, 2024 at the Hitachi Central Research Laboratory in Tokyo, Japan. This event brought together researchers from around the world to discuss advances in multimedia informatics, covering areas such as image processing, video quality assessment and artificial intelligence. Steve and Dominik reported on their latest findings in the field of image attractiveness and high frame rate video quality.

Steve's research on predicting the attractiveness of upscaled images using AI

Steve presented a new dataset of 1,496 upscaled images, each rated for visual attractiveness. The study compared five modern upscaling methods - Real-ESRGAN, BSRGAN, waifu2x, KXNet and Lanczos - and evaluated their performance. In addition, Steve presented predictive models that can identify which method was used for upscaling and estimate the attractiveness of the resulting images. This research offers valuable insights for the development of AI-driven image enhancement tools.

More details can be found in the publication: Appeal Prediction for AI Upscaled Images.

Dominik's study on the quality of high frame rate videos

Dominik focused on the impact of frame rate on perceived video quality. Using the SAMVIQ method, his research showed that higher frame rates significantly improve the video quality perceived by viewers. Dominik also presented open source software to facilitate high frame rate (HFR) research so that the community can conduct similar experiments and extend the results of the study.

For more information, please read the publication: Investigating the Impact of High Frame Rate on Video Quality: A SAMVIQ Approach.

Access the software here: AVRateNG on GitHub.