Research map / Probabilistic and scientific modeling / Molecules and proteins
Physics Inspired Optimization on Semantic Transfer Features: An Alternative Method for Room Layout Estimation
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 semantic transfer to VGG16 edge labelling (ablation) | VGG16 edge labelling · LSUN validation | ODS: 0.243 -> 0.271 | Helped |
| use ResNet-101 instead of VGG16 with semantic transfer (ablation) | semantic transfer edge labelling · LSUN validation | ODS: 0.271 -> 0.314 | Helped |
| use semantic transfer instead of standard SUNRGBD pre-training and last-layer reinitialization (ablation) | VGG16 with all parameters fine-tuned · LSUN validation | ODS: 0.268 -> 0.271 | No clear effect |
| add semantic transfer rather than standard pre-training when freezing parameters before layer 5b (ablation) | VGG16 edge labelling · LSUN validation | ODS: 0.202 -> 0.233 | Helped |
| apply PIO refinement to the best generated proposal instead of only ranking proposals (ablation) | LSUN room-layout proposal pipeline · LSUN validation | e_pixel (%): 11.28 -> 6.31 | Helped |
| remove dependence on edge-cue-generated proposals and optimize directly with PIO (ablation) | LSUN room-layout inference · LSUN validation | e_pixel (%): 6.31 -> 5.48 | Helped |
| use top 10 proposals rather than the single best proposal for PIO (ablation) | PIO proposal refinement · LSUN validation | e_pixel (%): 6.31 -> 5.75 | Helped |
| replace proposal ranking with proposal-free PIO (ablation) | LSUN room-layout inference · LSUN validation | e_pixel (%): 11.28 -> 5.48 | Helped |
| use PIO instead of NOB for inference | room-layout optimization · LSUN validation | Average running time per frame (artpf, s): 35.41 -> 1.79 | Helped |
| increase the gradient window size from 1 pixel to 10 pixels (ablation) | PIO · LSUN validation | e_pixel and e_corner | Hurt |
Closest papers by meaning
- OptiScene: LLM-driven Indoor Scene Layout Generation via Scaled Human-aligned Data Synthesis and Multi-Stage Preference Optimization (NeurIPS 2025 · similarity 0.86)
- Hierarchically-Structured Open-Vocabulary Indoor Scene Synthesis with Pre-trained Large Language Model (AAAI 2025 · similarity 0.85)
- S-INF: Towards Realistic Indoor Scene Synthesis via Scene Implicit Neural Field (AAAI 2025 · similarity 0.85)
- No More Ambiguity in 360deg Room Layout via Bi-Layout Estimation (CVPR 2024 · similarity 0.84)
- Direct Numerical Layout Generation for 3D Indoor Scene Synthesis via Spatial Reasoning (NeurIPS 2025 · similarity 0.84)
- DeBaRA: Denoising-Based 3D Room Arrangement Generation (NeurIPS 2024 · similarity 0.84)
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.