Gaussflaw dataset and model improve 3d gaussian splatting diagnosis
GaussFlaw: Fine-Grained Perceptual Diagnosis for 3D Gaussian Splatting
Computer Vision and Pattern Recognition
Summary
3D Gaussian Splatting is a way to create realistic 3D images quickly, but it can suffer from small, hard-to-find image flaws. The authors made a big dataset called GaussFlaw-80K that shows exactly where and what kind of problems happen in these images. They also built GeoCritic, a model that learns to spot and name these flaws by looking at both colors and depth information. This helps improve 3D image quality by fixing specific problems more precisely, not just judging the overall picture.
What this means in practice
- •For 3d graphics engineers: Automatically identify and classify localized image artifacts in 3D Gaussian splatting renders to guide targeted improvements in rendering quality.
- •For augmented reality developers: Use fine-grained artifact detection to enhance real-time rendering quality in AR applications based on 3D Gaussian splatting.
Authors
Shaocheng Shen, Zihan Zheng, Kejin Li, Lu Liu, Xiaoyun Zhang, Qiang Hu, Wenjun Zhang
Abstract
3D Gaussian Splatting (3DGS) enables photorealistic novel- view synthesis with real-time rendering, yet fine-grained di- agnosis of its rendering artifacts remains challenging. These artifacts are typically spatially localized and exhibit diverse degradation patterns, making them difficult to character- ize using global quality metrics and image-level assessment benchmarks, which provide no explicit indication of where they occur or what types of artifacts they represent. To ad- dress this limitation, we introduce GaussFlaw-80K, the first large-scale dataset for fine-grained perceptual diagnosis in 3DGS. Constructed through a model-assisted human-in-the- loop pipeline, GaussFlaw-80K contains 80,049 artifact re- gions across 14,289 rendered images, each annotated with a pixel-level mask and one of eight flaw labels spanning ge- ometric, blending, and photometric degradations. These an- notations turn 3DGS artifact diagnosis into an explicit and measurable task. Building on this dataset, we propose Ge- oCritic, a geometry-guided multi-task model that integrates RGB appearance with depth-derived cues to jointly localize artifacts and classify their flaw types. Beyond evaluation, we further demonstrate the downstream benefits of fine-grained diagnosis through three complementary targeted operations: Gaussian densification, attribute correction, and outlier prun- ing. Experiments show that GeoCritic achieves an AUC-J of 0.9507 and an Acc@1 of 80.08%, while diagnosis-guided re- finement improves all reported global and crop-level metrics across diverse datasets and 3DGS backbones.