Thomas Schierl

Soonbin Lee, Fangwen Shu, Yago Sanchez et al.

Explicit feature-grid based NeRF models have shown promising results in terms of rendering quality and significant speed-up in training. However, these methods often require a significant amount of data to represent a single scene or object. In this work, we present a compression model that aims to minimize the entropy in the frequency domain in order to effectively reduce the data size. First, we propose using the discrete cosine transform (DCT) on the tensorial radiance fields to compress the feature-grid. This feature-grid is transformed into coefficients, which are then quantized and entropy encoded, following a similar approach to the traditional video coding pipeline. Furthermore, to achieve a higher level of sparsity, we propose using an entropy parameterization technique for the frequency domain, specifically for DCT coefficients of the feature-grid. Since the transformed coefficients are optimized during the training phase, the proposed model does not require any fine-tuning or additional information. Our model only requires a lightweight compression pipeline for encoding and decoding, making it easier to apply volumetric radiance field methods for real-world applications. Experimental results demonstrate that our proposed frequency domain entropy model can achieve superior compression performance across various datasets. The source code will be made publicly available.

Soonbin Lee, Yeong-Gyu Kim, Simon Sasse et al.

Existing 3D Gaussian Splatting simplification methods commonly use importance scores, such as blending weights or sensitivity, to identify redundant Gaussians. However, these scores are not driven by visual error metrics, often leading to suboptimal trade-offs between compactness and rendering fidelity. We present GaussianPOP, a principled simplification framework based on analytical Gaussian error quantification. Our key contribution is a novel error criterion, derived directly from the 3DGS rendering equation, that precisely measures each Gaussian's contribution to the rendered image. By introducing a highly efficient algorithm, our framework enables practical error calculation in a single forward pass. The framework is both accurate and flexible, supporting on-training pruning as well as post-training simplification via iterative error re-quantification for improved stability. Experimental results show that our method consistently outperforms existing state-of-the-art pruning methods across both application scenarios, achieving a superior trade-off between model compactness and high rendering quality.

Robert Skupin, Christian Bartnik, Adam Wieckowski et al.

The user experience in adaptive HTTP streaming relies on offering bitrate ladders with suitable operation points for all users and typically involves multiple resolutions. While open GOP coding structures are generally known to provide substantial coding efficiency benefit, their use in HTTP streaming has been precluded through lacking support of reference picture resampling (RPR) in AVC and HEVC. The newly emerging Versatile Video Coding (VVC) standard supports RPR, but only conversational scenarios were primarily investigated during the design of VVC. This paper aims at enabling usage of RPR in HTTP streaming scenarios through analysing the drift potential of VVC coding tools and presenting a constrained encoding method that avoids severe drift artefacts in resolution switching with open GOP coding in VVC. In typical live streaming configurations, the presented method achieves -8.57% BD-rate reduction compared to closed GOP coding while in a typical Video on Demand configuration, -1.89% BD-rate reduction is reported. The constraints penalty compared to regular open GOP coding is 0.65% BD-rate in the worst case. The presented method was integrated into the publicly available open source VVC encoder VVenC v0.3.

Soonbin Lee, Fangwen Shu, Yago Sanchez et al.

3D Gaussian Splatting is a recognized method for 3D scene representation, known for its high rendering quality and speed. However, its substantial data requirements present challenges for practical applications. In this paper, we introduce an efficient compression technique that significantly reduces storage overhead by using compact representation. We propose a unified architecture that combines point cloud data and feature planes through a progressive tri-plane structure. Our method utilizes 2D feature planes, enabling continuous spatial representation. To further optimize these representations, we incorporate entropy modeling in the frequency domain, specifically designed for standard video codecs. We also propose channel-wise bit allocation to achieve a better trade-off between bitrate consumption and feature plane representation. Consequently, our model effectively leverages spatial correlations within the feature planes to enhance rate-distortion performance using standard, non-differentiable video codecs. Experimental results demonstrate that our method outperforms existing methods in data compactness while maintaining high rendering quality. Our project page is available at https://fraunhoferhhi.github.io/CodecGS

Serhan Gül, Sebastian Bosse, Dimitri Podborski et al.

Volumetric video allows viewers to experience highly-realistic 3D content with six degrees of freedom in mixed reality (MR) environments. Rendering complex volumetric videos can require a prohibitively high amount of computational power for mobile devices. A promising technique to reduce the computational burden on mobile devices is to perform the rendering at a cloud server. However, cloud-based rendering systems suffer from an increased interaction (motion-to-photon) latency that may cause registration errors in MR environments. One way of reducing the effective latency is to predict the viewer's head pose and render the corresponding view from the volumetric video in advance. In this paper, we design a Kalman filter for head motion prediction in our cloud-based volumetric video streaming system. We analyze the performance of our approach using recorded head motion traces and compare its performance to an autoregression model for different prediction intervals (look-ahead times). Our results show that the Kalman filter can predict head orientations 0.5 degrees more accurately than the autoregression model for a look-ahead time of 60 ms.

Serhan Gül, Dimitri Podborski, Jangwoo Son et al.

Volumetric video is an emerging technology for immersive representation of 3D spaces that captures objects from all directions using multiple cameras and creates a dynamic 3D model of the scene. However, processing volumetric content requires high amounts of processing power and is still a very demanding task for today's mobile devices. To mitigate this, we propose a volumetric video streaming system that offloads the rendering to a powerful cloud/edge server and only sends the rendered 2D view to the client instead of the full volumetric content. We use 6DoF head movement prediction techniques, WebRTC protocol and hardware video encoding to ensure low-latency in different parts of the processing chain. We demonstrate our system using both a browser-based client and a Microsoft HoloLens client. Our application contains generic interfaces that allow for easy deployment of various augmented/mixed reality clients using the same server implementation.

Serhan Gül, Dimitri Podborski, Thomas Buchholz et al.

Volumetric video is an emerging key technology for immersive representation of 3D spaces and objects. Rendering volumetric video requires lots of computational power which is challenging especially for mobile devices. To mitigate this, we developed a streaming system that renders a 2D view from the volumetric video at a cloud server and streams a 2D video stream to the client. However, such network-based processing increases the motion-to-photon (M2P) latency due to the additional network and processing delays. In order to compensate the added latency, prediction of the future user pose is necessary. We developed a head motion prediction model and investigated its potential to reduce the M2P latency for different look-ahead times. Our results show that the presented model reduces the rendering errors caused by the M2P latency compared to a baseline system in which no prediction is performed.

Dimitri Podborski, Jangwoo Son, Gurdeep Singh Bhullar et al.

Virtual Reality (VR) and 360-degree video streaming have gained significant attention in recent years. First standards have been published in order to avoid market fragmentation. For instance, 3GPP released its first VR specification to enable 360-degree video streaming over 5G networks which relies on several technologies specified in ISO/IEC 23090-2, also known as MPEG-OMAF. While some implementations of OMAF-compatible players have already been demonstrated at several trade shows, so far, no web browser-based implementations have been presented. In this demo paper we describe a browser-based JavaScript player implementation of the most advanced media profile of OMAF: HEVC-based viewport-dependent OMAF video profile, also known as tile-based streaming, with multi-resolution HEVC tiles. We also describe the applied workarounds for the implementation challenges we encountered with state-of-the-art HTML5 browsers. The presented implementation was tested in the Safari browser with support of HEVC video through the HTML5 Media Source Extensions API. In addition, the WebGL API was used for rendering, using region-wise packing metadata as defined in OMAF.

Nils Strodthoff, Barış Göktepe, Thomas Schierl et al.

We investigate Early Hybrid Automatic Repeat reQuest (E-HARQ) feedback schemes enhanced by machine learning techniques as a path towards ultra-reliable and low-latency communication (URLLC). To this end, we propose machine learning methods to predict the outcome of the decoding process ahead of the end of the transmission. We discuss different input features and classification algorithms ranging from traditional methods to newly developed supervised autoencoders. These methods are evaluated based on their prospects of complying with the URLLC requirements of effective block error rates below $10^{-5}$ at small latency overheads. We provide realistic performance estimates in a system model incorporating scheduling effects to demonstrate the feasibility of E-HARQ across different signal-to-noise ratios, subcode lengths, channel conditions and system loads, and show the benefit over regular HARQ and existing E-HARQ schemes without machine learning.