In the past few years, Gaussian Splatting has become the most promising volumetric video representation. For real-world scenarios, a great deal of research effort has been directed toward better spatiotemporal multimodal scene reconstruction and more effective deployment. In this talk, we will first introduce the fundamental representation of Gaussian Splatting, including its attributes, construction, rendering, and optimization.
We will then overview the recent development of Gaussian Splatting across the end-to-end ecosystem—from content capture, creation, and delivery to content consumption. More specifically, at the content capture stage, we will address issues and solutions for multiview camera setups in both the spatial and temporal domains to facilitate high-quality model building. At the content creation stage, to enrich the multimodal experience, learned audio and semantic attributes from foundation models (such as CLIP, DINO, etc.) are embedded into the Gaussian Splatting primitives to enable a joint audio-visual-semantic representation. Volumetric video editing to enhance the perceived experience is also an important tool along this pipeline.
At the content delivery stage, instead of explicitly coding dozens of attributes per Gaussian, implicit methods—such as using coarse geometry representations, 2D plane projection, and MLPs to leverage conventional 2D video codecs—will be discussed to achieve photorealistic quality at streamable bitrates. At the content consumption stage, language- and semantics-guided methods are presented to enable interactive 3D scene navigation and efficient, physics-aware multimodal rendering.
Finally, we will present the latest international standardization efforts and highlight future research trends. In summary, the goal of this talk is to enable ICASSP 2026 attendees to understand the latest Gaussian Splatting technologies in both theory and practice. A secondary goal is to foster further discussion on volumetric video research at ICASSP 2026, with the hope of motivating attendees to identify potential research topics and inspiring innovative technical papers for ICASSP 2027.
