Sep 2026· Italian National Conference on Sensors· 0 citations· 21 references
TL;DR
This paper proposes SEELE, a Lyapunov-guided online control algorithm that represents accumulated budget pressure with two virtual queues and converts the long-term constrained problem into lightweight per-slot decisions, and provides steady-state QoE statistically comparable to a pretrained PPO policy while significantly reducing synchronization violations and improving composition stability.
Abstract
Immersive media services increasingly rely on edge-cloud rendering to deliver interactive visual content under dynamic network, computing, and mobility conditions. Rendering an entire scene as a single service couples interaction-sensitive foreground content with context-oriented background content, making it difficult to jointly control latency, quality, synchronization, and migration overhead. This paper studies sense-driven edge-cloud foreground–background split rendering for immersive media services. We formulate an online decision problem in which foreground and background rendering layers can be independently controlled under long-term system and migration cost budgets. The formulation turns structural scene separation into a coupled layer-state control problem by preserving asymmetric QoE roles and a common composition requirement. We propose SEELE, a Lyapunov-guided online control algorithm that represents accumulated budget pressure with two virtual queues and converts the long-term constrained problem into lightweight per-slot decisions. The resulting per-slot rule balances immediate QoE loss against queue-weighted system and migration costs. Under sustained resource and network stress, SEELE provides steady-state QoE statistically comparable to a pretrained PPO policy while significantly reducing synchronization violations and improving composition stability. It also improves steady-state QoE and system debt over deterministic and QoE-prioritized baselines. A prototype implementation and controlled characterization further validate split-stream deployment, runtime observability, practical control hooks, and the latency–capacity tradeoff of layered rendering.
3D Gaussian Splatting (3DGS) enables high-quality and efficient novel-view rendering, making it a promising representation for immersive scene streaming. However, supporting interactive 3DGS viewing on lightweight clients over wireless edge networks remains challenging, as view-dependent rendering workload, video encod...
Di-Hao Hu, Jinkai Zheng, T. Luan et al.· 2026 IEEE/CIC International...· 0 citations
Mobile augmented reality (AR) supports rich interactive applications by enabling high-fidelity 3D content to be displayed in real time on mobile devices. However, under timevarying wireless conditions and fluctuating device-side rendering contention, reconstructing such high-fidelity 3D content in a timely manner remai...
Shi-Dong Wang, Bang-Cai Wei, Chun-Sheng Wang et al.· 2026 IEEE/CIC International...· 0 citations
Manifold4D is proposed, which injects the render directly into the initial noise of flow matching, so that generation no longer departs from standard Gaussian noise but from a geometry-bearing noise distribution that is term the point cloud rendered manifold, leaving the source video as the only visual condition.
Yongbo Mao, Zijia Dai, Zhi-Shu Liu et al.· 0 citations
This work proposes Streaming4D, a tightly coupled synchronous pipeline that integrates block-wise autoregressive video generation with incremental 3D reconstruction, and instantiates it using a Self-Forcing-style autoregressive generator and an incremental reconstruction backend.
Xiao-Yan Liu, Jia-Xin Liu, Kang Li et al.· 0 citations
To address the issues of spatial positioning drift, unstable semantic overlay, and insufficient real-time rendering efficiency in complex dynamic scenes of immersive Japanese translation screens in augmented reality environments, which limit the development potential of related applications, this paper proposes a multi...
Li Wang· International Conference Com...· 0 citations
A unified closed-loop cross-layer framework is formulates that coordinates prediction, tiling, bitrate allocation, request timing, transport configuration, buffering, and edge processing under bandwidth, latency, and resource constraints.
Muhammad Farooq, G. Manfredi, L. De Cicco et al.· Network· 1 citation
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduNov 17, 2026
By studying the computational foundations of language learning, Associate Professor Jacob Andreas builds intelligent systems that can communicate effectively with humans.
Microsoft Research Blog· microsoft.comOct 29, 2026
Host Sowbhagya Suresh shares some of the latest news and cutting-edge research from across Microsoft—from understanding the challenges facing coding agents to uncovering how people use AI at work and improving extended-range weather forecasting. The post Microsoft Research Focus – Edition #97 appeared first on Microsoft Research.
MIT News · Artificial Intelligence· news.mit.eduOct 28, 2026
Professor Paris Smaragdis develops computers with superhuman hearing for applications like speech recognition, music understanding, and environmental monitoring.