Research map / 3D vision / 3D generation
AdaSFormer: Adaptive Serialized Transformers for Monocular Semantic Scene Completion from Indoor Environments
Xuzhi Wang, Xinran Wu, Song Wang, Lingdong Kong, Ziping Zhao
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add Adaptive Serialized Attention (ablation) | conv-only baseline · NYUv2 test | SSC mIoU (%): 28.68 -> 30.98 | Improved |
| add Center-Relative Positional Encoding (ablation) | baseline + ASA · NYUv2 test | SSC mIoU (%): 30.98 -> 31.66 | Improved |
| add Conv-Modulated Layer Normalization block (ablation) | baseline + ASA + CRPE · NYUv2 test | SSC mIoU (%): 31.66 -> 32.50 | Improved |
| learn shifts with Gumbel-Softmax instead of direct discrete parameters (ablation) | ASA · NYUv2 test | SSC mIoU (%): 30.61 -> 31.96 | Improved |
| add temperature annealing to Gumbel-Softmax shift learning (ablation) | ASA · NYUv2 test | SSC mIoU (%): 31.96 -> 32.50 | Improved |
| use projected voxel mean as scene center instead of physical center (ablation) | CRPE · NYUv2 test | SSC mIoU (%): 32.11 -> 32.50 | Improved |
| add relative yaw and pitch positional encoding (ablation) | CRPE · NYUv2 test | SSC mIoU (%): 32.23 -> 32.50 | Improved |
| add ASA component (latency cost) (ablation) | AdaSFormer · Latency | latency (ms): 156.0 -> 178.9 | Hurt |
| use AdaSFormer instead of ISO | ISO · Latency | latency (ms): 535.2 -> 178.9 | Improved |
| use AdaSFormer instead of ISO on Occ-ScanNet | ISO · Occ-ScanNet val | mIoU (%): 28.71 -> 45.33 | Improved |
| use AdaSFormer instead of MonoMRN on NYUv2 scene completion | MonoMRN · NYUv2 test | SC IoU (%): 53.16 -> 51.33 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 4 on most shared results.
- SurroundOCC — SurroundOcc: Multi-camera 3D Occupancy Prediction for Autonomous Driving (ICCV 2023) · better on 2 of 2 shared results
- GaussianFormer — GaussianFormer: Scene as Gaussians for Vision-Based 3D Semantic Occupancy Prediction (ECCV 2024) · better on 1 of 1 shared result
- NDC-Scene — NDC-Scene: Boost Monocular 3D Semantic Scene Completion in Normalized Device Coordinates Space (ICCV 2023) · better on 1 of 1 shared result
- MonoScene — MonoScene: Monocular 3D Semantic Scene Completion (CVPR 2022) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- H2GFormer: Horizontal-to-Global Voxel Transformer for 3D Semantic Scene Completion (AAAI 2024 · similarity 0.88)
- Context and Geometry Aware Voxel Transformer for Semantic Scene Completion (NeurIPS 2024 · similarity 0.87)
- IPFormer: Visual 3D Panoptic Scene Completion with Context-Adaptive Instance Proposals (NeurIPS 2025 · similarity 0.87)
- InstructScene: Instruction-Driven 3D Indoor Scene Synthesis with Semantic Graph Prior (ICLR 2024 · similarity 0.86)
- MiKASA: Multi-Key-Anchor & Scene-Aware Transformer for 3D Visual Grounding (CVPR 2024 · similarity 0.86)
- Geometric Exploitation for Indoor Panoramic Semantic Segmentation (NeurIPS 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
