# Semantic Ray: Learning a Generalizable Semantic Field with Cross-Reprojection Attention

arXiv: https://arxiv.org/abs/2303.13014 · HTML: https://mlsanity.com/paper/2303.13014/

## Key facts

- Cited by 29 papers in the map (data through 14 Oct 2026).
- Builds on 54 of its 68 references that are in the map.
- First posted on arXiv 23 Mar 2023.

## Builds on (most cited first, 30 of 54)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://mlsanity.com/paper/1912.01703/) (2019)
- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://mlsanity.com/paper/1505.04597/) (2015)
- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [The Unreasonable Effectiveness of Deep Features as a Perceptual Metric](https://mlsanity.com/paper/1801.03924/) (2018)
- [PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation](https://mlsanity.com/paper/1612.00593/) (2016)
- [Fully Convolutional Networks for Semantic Segmentation](https://mlsanity.com/paper/1411.4038/) (2014)
- [PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space](https://mlsanity.com/paper/1706.02413/) (2017)
- [DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs](https://mlsanity.com/paper/1606.00915/) (2016)
- [Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation](https://mlsanity.com/paper/1802.02611/) (2018)
- [ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenes](https://mlsanity.com/paper/1702.04405/) (2017)
- [DeepSDF: Learning Continuous Signed Distance Functions for Shape Representation](https://mlsanity.com/paper/1901.05103/) (2019)
- [Rethinking Atrous Convolution for Semantic Image Segmentation](https://mlsanity.com/paper/1706.05587/) (2017)
- [Occupancy Networks: Learning 3D Reconstruction in Function Space](https://mlsanity.com/paper/1812.03828/) (2018)
- [Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fields](https://mlsanity.com/paper/2103.13415/) (2021)
- [pixelNeRF: Neural Radiance Fields from One or Few Images](https://mlsanity.com/paper/2012.02190/) (2020)
- [SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation](https://mlsanity.com/paper/1511.00561/) (2015)
- [Direct Voxel Grid Optimization: Super-fast Convergence for Radiance Fields Reconstruction](https://mlsanity.com/paper/2111.11215/) (2021)
- [Neural Sparse Voxel Fields](https://mlsanity.com/paper/2007.11571/) (2020)
- [PlenOctrees for Real-time Rendering of Neural Radiance Fields](https://mlsanity.com/paper/2103.14024/) (2021)
- [The Replica Dataset: A Digital Replica of Indoor Spaces](https://mlsanity.com/paper/1906.05797/) (2019)
- [Scene Representation Networks: Continuous 3D-Structure-Aware Neural Scene Representations](https://mlsanity.com/paper/1906.01618/) (2019)
- [NeRF++: Analyzing and Improving Neural Radiance Fields](https://mlsanity.com/paper/2010.07492/) (2020)
- [GIRAFFE: Representing Scenes as Compositional Generative Neural Feature Fields](https://mlsanity.com/paper/2011.12100/) (2020)
- [MVSNeRF: Fast Generalizable Radiance Field Reconstruction from Multi-View Stereo](https://mlsanity.com/paper/2103.15595/) (2021)
- [Depth-supervised NeRF: Fewer Views and Faster Training for Free](https://mlsanity.com/paper/2107.02791/) (2021)
- [CCNet: Criss-Cross Attention for Semantic Segmentation](https://mlsanity.com/paper/1811.11721/) (2018)
- [IBRNet: Learning Multi-View Image-Based Rendering](https://mlsanity.com/paper/2102.13090/) (2021)
- [GRAF: Generative Radiance Fields for 3D-Aware Image Synthesis](https://mlsanity.com/paper/2007.02442/) (2020)
- [KiloNeRF: Speeding up Neural Radiance Fields with Thousands of Tiny MLPs](https://mlsanity.com/paper/2103.13744/) (2021)

## Built on by (most cited first, 29 of 29)

- [Gaussian Graph Network: Learning Efficient and Generalizable Gaussian Representations from Multi-view Images](https://mlsanity.com/paper/2503.16338/) (2025)
- [GP-NeRF: Generalized Perception NeRF for Context-Aware 3D Scene Understanding](https://mlsanity.com/paper/2311.11863/) (2023)
- [GSNeRF: Generalizable Semantic Neural Radiance Fields with Enhanced 3D Scene Understanding](https://mlsanity.com/paper/2403.03608/) (2024)
- [Lift3D: Zero-Shot Lifting of Any 2D Vision Model to 3D](https://mlsanity.com/paper/2403.18922/) (2024)
- [FreeSplat++: Generalizable 3D Gaussian Splatting for Efficient Indoor Scene Reconstruction](https://mlsanity.com/paper/2503.22986/) (2025)
- [UniForward: Unified 3D Scene and Semantic Field Reconstruction via Feed-Forward Gaussian Splatting from Only Sparse-View Images](https://mlsanity.com/paper/2506.09378/) (2025)
- [GNeSF: Generalizable Neural Semantic Fields](https://mlsanity.com/paper/2310.15712/) (2023)
- [Semantically-aware Neural Radiance Fields for Visual Scene Understanding: A Comprehensive Review](https://mlsanity.com/paper/2402.11141/) (2024)
- [GOV-NeSF: Generalizable Open-Vocabulary Neural Semantic Fields](https://mlsanity.com/paper/2404.00931/) (2024)
- [Semantic Flow: Learning Semantic Field of Dynamic Scenes from Monocular Videos](https://mlsanity.com/paper/2404.05163/) (2024)
- [LangScene-X: Reconstruct Generalizable 3D Language-Embedded Scenes with TriMap Video Diffusion](https://mlsanity.com/paper/2507.02813/) (2025)
- [RT-GS2: Real-Time Generalizable Semantic Segmentation for 3D Gaussian Representations of Radiance Fields](https://mlsanity.com/paper/2405.18033/) (2024)
- [SparseLGS: Sparse View Language Embedded Gaussian Splatting](https://mlsanity.com/paper/2412.02245/) (2024)
- [H3R: Hybrid Multi-view Correspondence for Generalizable 3D Reconstruction](https://mlsanity.com/paper/2508.03118/) (2025)
- [TK-Planes: Tiered K-Planes with High Dimensional Feature Vectors for Dynamic UAV-based Scenes](https://mlsanity.com/paper/2405.02762/) (2024)
- [MV2Cyl: Reconstructing 3D Extrusion Cylinders from Multi-View Images](https://mlsanity.com/paper/2406.10853/) (2024)
- [InfoNorm: Mutual Information Shaping of Normals for Sparse-View Reconstruction](https://mlsanity.com/paper/2407.12661/) (2024)
- [SemGS: Feed-Forward Semantic 3D Gaussian Splatting from Sparse Views for Generalizable Scene Understanding](https://mlsanity.com/paper/2603.02548/) (2026)
- [LangFlash: Feed-forward 3D Language Gaussian Splatting from Sparse Unposed Images](https://mlsanity.com/paper/2605.23287/) (2026)
- [Semantic-Human: Neural Rendering of Humans from Monocular Video with Human Parsing](https://mlsanity.com/paper/2308.09894/) (2023)
- [uSF: Learning Neural Semantic Field with Uncertainty](https://mlsanity.com/paper/2312.08012/) (2023)
- [OPONeRF: One-Point-One NeRF for Robust Neural Rendering](https://mlsanity.com/paper/2409.20043/) (2024)
- [TeX-NeRF: Neural Radiance Fields from Pseudo-TeX Vision](https://mlsanity.com/paper/2410.04873/) (2024)
- [M3DMap: Object-aware Multimodal 3D Mapping for Dynamic Environments](https://mlsanity.com/paper/2508.17044/) (2025)
- [ESCAPE: Episodic Spatial Memory and Adaptive Execution Policy for Long-Horizon Mobile Manipulation](https://mlsanity.com/paper/2604.13633/) (2026)
- [Omni-Recon: Harnessing Image-based Rendering for General-Purpose Neural Radiance Fields](https://mlsanity.com/paper/2403.11131/) (2024)
- [Neural Radiance and Gaze Fields for Visual Attention Modeling in 3D Environments](https://mlsanity.com/paper/2503.07828/) (2025)
- [Gaussian Swaying: Surface-Based Framework for Aerodynamic Simulation with 3D Gaussians](https://mlsanity.com/paper/2512.01306/) (2025)
- [First Shape, Then Meaning: Efficient Geometry and Semantics Learning for Indoor Reconstruction](https://mlsanity.com/paper/2605.03463/) (2026)

## Abstract (arXiv)

In this paper, we aim to learn a semantic radiance field from multiple scenes that is accurate, efficient and generalizable. While most existing NeRFs target at the tasks of neural scene rendering, image synthesis and multi-view reconstruction, there are a few attempts such as Semantic-NeRF that explore to learn high-level semantic understanding with the NeRF structure. However, Semantic-NeRF simultaneously learns color and semantic label from a single ray with multiple heads, where the single ray fails to provide rich semantic information. As a result, Semantic NeRF relies on positional encoding and needs to train one specific model for each scene. To address this, we propose Semantic Ray (S-Ray) to fully exploit semantic information along the ray direction from its multi-view reprojections. As directly performing dense attention over multi-view reprojected rays would suffer from heavy computational cost, we design a Cross-Reprojection Attention module with consecutive intra-view radial and cross-view sparse attentions, which decomposes contextual information along reprojected rays and cross multiple views and then collects dense connections by stacking the modules. Experiments show that our S-Ray is able to learn from multiple scenes, and it presents strong generalization ability to adapt to unseen scenes.