Talks:
- On SMPTE-2094-50: A New Dynamic Metadata Standard for HDR, Chris Cameron, Google
Integrating high dynamic range (HDR) content into composited systems and dynamic viewing conditions has proven difficult because initial standards were focused on viewing a single content source in a theater-like environment. This talk discusses how SMPTE ST 2094-50 addresses these challenges by introducing metadata for an HDR reference white anchor and by characterizing HDR content and displays by their HDR headroom.
- Benchmarking UGC Video Quality Assessment Beyond Mean Opinion Scores, Yilin Wang, Google
The rapid expansion of User-Generated Content (UGC) underscores the need for robust Video Quality Assessment (VQA) models with practical applicability. Existing No-Reference (NR) VQA metrics excel at evaluating absolute video quality across varied content (inter-content accuracy). However, they often struggle to distinguish fine-grained quality variations—such as those caused by compression or enhancement—among variants originating from the same source (intra-content accuracy). To overcome these limitations and accelerate research in fine-grained UGC quality evaluation, this talk presents the UGC Fine-grained Quality (UGC-FQ) Dataset. This large-scale benchmark contains 700 distinct source videos with 36 compression and enhancement variations per source, accompanied by comprehensive subjective quality annotations for both inter- and intra-content differences. Additionally, we introduce an enhanced UGC Video Quality (UVQ) training framework engineered to enable NR-VQA models to reliably measure both overall perceptual quality and subtle quality nuances.
- AI Extension Layer on Standard Codecs, In-Suk Chong, Jianle Chen, Debargha Mukherjee, Google
This talk will present the challenges in AI based video compression, and propose a middle-ground scalable coding approach where the base layer is conventionally coded while the enhancement layer is AI based. This model can be deployed today with some benefits while still leaving ample room to grow into a full AI codec in the future as the technology becomes more mature.
