# From Explicit References to Scene Manifolds: Distributional Fidelity and Realism for Radiance Field Quality Assessment

arXiv: https://arxiv.org/abs/2609.07346 · HTML: https://mlsanity.com/paper/2609.07346/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 24 of its 44 references that are in the map.
- First posted on arXiv 7 Sep 2026.

## Builds on (most cited first, 24 of 24)

- [The Unreasonable Effectiveness of Deep Features as a Perceptual Metric](https://mlsanity.com/paper/1801.03924/) (2018)
- [3D Gaussian Splatting for Real-Time Radiance Field Rendering](https://mlsanity.com/paper/2308.04079/) (2023)
- [Image-to-Image Translation with Conditional Adversarial Networks](https://mlsanity.com/paper/1611.07004/) (2016)
- [MUSIQ: Multi-scale Image Quality Transformer](https://mlsanity.com/paper/2108.05997/) (2021)
- [Image Quality Assessment: Unifying Structure and Texture Similarity](https://mlsanity.com/paper/2004.07728/) (2020)
- [PaDiM: a Patch Distribution Modeling Framework for Anomaly Detection and Localization](https://mlsanity.com/paper/2011.08785/) (2020)
- [Blind Image Quality Assessment via Vision-Language Correspondence: A Multitask Learning Perspective](https://mlsanity.com/paper/2303.14968/) (2023)
- [TOPIQ: A Top-down Approach from Semantics to Distortions for Image Quality Assessment](https://mlsanity.com/paper/2308.03060/) (2023)
- [Compact 3D Scene Representation via Self-Organizing Gaussian Grids](https://mlsanity.com/paper/2312.13299/) (2023)
- [ARNIQA: Learning Distortion Manifold for Image Quality Assessment](https://mlsanity.com/paper/2310.14918/) (2023)
- [3DGS.zip: A survey on 3D Gaussian Splatting Compression Methods](https://mlsanity.com/paper/2407.09510/) (2024)
- [Toward Generalized Image Quality Assessment: Relaxing the Perfect Reference Quality Assumption](https://mlsanity.com/paper/2503.11221/) (2025)
- [Good, Cheap, and Fast: Overfitted Image Compression with Wasserstein Distortion](https://mlsanity.com/paper/2412.00505/) (2024)
- [Opinion-Unaware Blind Image Quality Assessment using Multi-Scale Deep Feature Statistics](https://mlsanity.com/paper/2405.18790/) (2024)
- [NeRF-NQA: No-Reference Quality Assessment for Scenes Generated by NeRF and Neural View Synthesis Methods](https://mlsanity.com/paper/2412.08029/) (2024)
- [GS-QA: Comprehensive Quality Assessment Benchmark for Gaussian Splatting View Synthesis](https://mlsanity.com/paper/2502.13196/) (2025)
- [Wasserstein Distortion: Unifying Fidelity and Realism](https://mlsanity.com/paper/2310.03629/) (2023)
- [Explicit-NeRF-QA: A Quality Assessment Database for Explicit NeRF Model Compression](https://mlsanity.com/paper/2407.08165/) (2024)
- [CrossScore: Towards Multi-View Image Evaluation and Scoring](https://mlsanity.com/paper/2404.14409/) (2024)
- [3DGS-IEval-15K: A Large-scale Image Quality Evaluation Database for 3D Gaussian-Splatting](https://mlsanity.com/paper/2506.14642/) (2025)
- [NVS-SQA: Exploring Self-Supervised Quality Representation Learning for Neurally Synthesized Scenes without References](https://mlsanity.com/paper/2501.06488/) (2025)
- [3DGS-VBench: A Comprehensive Video Quality Evaluation Benchmark for 3DGS Compression](https://mlsanity.com/paper/2508.07038/) (2025)
- [Puzzle Similarity: A Perceptually-guided Cross-Reference Metric for Artifact Detection in 3D Scene Reconstructions](https://mlsanity.com/paper/2411.17489/) (2024)
- [Non-Aligned Reference Image Quality Assessment for Novel View Synthesis](https://mlsanity.com/paper/2511.08155/) (2025)

## Abstract (arXiv)

Radiance field representations such as 3D Gaussian Splatting (3DGS) enable high-quality novel view synthesis but can introduce complex, view-dependent artifacts from reconstruction, rendering, and compression. Reliable perceptual quality assessment (QA) is thus essential for evaluating rendered views and guiding the design of perceptually faithful scene representations. Existing full-reference QA metrics require an aligned reference image, while recent cross-reference metrics relax this requirement by comparing a test view with non-aligned references. However, under wide-baseline radiance field settings, selecting a reliable nearby reference can be difficult, particularly when evaluating views along arbitrary trajectories and poses. We propose SCODA, a lightweight scene-conditioned objective QA method that shifts QA from explicit image-to-image comparison to scene-manifold modeling. High-quality observations of each scene are represented as a multivariate Gaussian distribution in deep feature space, producing a semantic fidelity score that measures deviation from the scene distribution. A weakly-supervised distortion-aware patch discriminator provides a complementary realism signal, and both cues are combined through an unsupervised bounded fusion strategy. Experiments on multiple benchmarks show strong agreement with human judgments and robust generalization across GS- and NeRF-generated views and trajectories. Code is publicly available at https://gitlab.com/saeedmp/scoda.