Dawid Mieloch

Dawid Mieloch, Stuart Perry

Immersive video delivery is bottlenecked by pixel-rate constraints, making the transmission of high-resolution depth maps or explicit 3D volumetric data expensive. Decoder-Side Depth Estimation (DSDE) shifts depth computation to the client, but struggles with complex geometries, inter-view flickering, and non-Lambertian reflections. Conversely, 3D Gaussian Splatting (3DGS) offers state-of-the-art view synthesis, but transmitting splats (or their projected 2D maps) incurs prohibitive bandwidth costs and is poorly aligned with standard video codecs. We propose Decoder-Side Gaussian Splatting (DSGS), a framework that natively replaces the depth-estimation stage of DSDE with feed-forward 3DGS inference, optimizing volumetric scenes entirely on the decoder side from compressed textures and metadata. A central, counterintuitive finding is that lossy compression acts as an implicit low-pass filter stabilizing feed-forward splat prediction: compressed bitstreams exceed lossless quality while shrinking tenfold. Under extreme view sparsity (one 2D atlas comprising 4 input views), DSGS achieves a +5.79 dB BD-PSNR and +0.054 BD-SSIM gain over the DSDE anchor while reducing maximum inter-view Delta IV-PSNR from 17.2 dB to 6.4 dB, minimizing the domain shift between transmitted and virtual viewports.

Dawid Mieloch, Olgierd Stankiewicz, Marek Domański

The paper presents a new method of depth estimation dedicated for free-viewpoint television (FTV). The estimation is performed for segments and thus their size can be used to control a trade-off between the quality of depth maps and the processing time of their estimation. The proposed algorithm can take as its input multiple arbitrarily positioned views which are simultaneously used to produce multiple inter view consistent output depth maps. The presented depth estimation method uses novel parallelization and temporal consistency enhancement methods that significantly reduce the processing time of depth estimation. An experimental assessment of the proposals has been performed, based on the analysis of virtual view quality in FTV. The results show that the proposed method provides an improvement of the depth map quality over the state of-the-art method, simultaneously reducing the complexity of depth estimation. The consistency of depth maps, which is crucial for the quality of the synthesized video and thus the quality of experience of navigating through a 3D scene, is also vastly improved.