Research map / 3D vision / Depth and stereo
Iris: Integrating Language into Diffusion-based Monocular Depth Estimation
Ziyao Zeng, Jingcheng Ni, Daniel Wang, Patrick Rim, Younjoon Chung, Fengyu Yang, Byung-Woo Hong, Alex Wong
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 text conditioning in training and inference | Marigold · NYUv2 | AbsRel (%): 6.1 -> 5.9 | Improved |
| add text conditioning in training and inference | Lotus-G · NYUv2 | AbsRel (%): 6.7 -> 5.9 | Improved |
| add text conditioning in training and inference | E2E-FT · KITTI | AbsRel (%): 10.5 -> 9.7 | Improved |
| add text conditioning in training and inference | Lotus-D · KITTI | AbsRel (%): 8.7 -> 8.4 | Improved |
| add text only during training (ablation) | Marigold · NYUv2 | AbsRel (%): 6.1 -> 6.2 | Hurt |
| add text only during training (ablation) | Lotus-G · KITTI | AbsRel (%): 8.9 -> 9.2 | Hurt |
| add text only during inference (ablation) | Marigold · NYUv2 | AbsRel (%): 6.1 -> 6.1 | No clear effect |
| add text only during inference (ablation) | E2E-FT · NYUv2 | AbsRel (%): 6.9 -> 6.4 | Improved |
| add text conditioning | E2E-FT · DIODE | AbsRel (%): 31.0 -> 31.4 | Hurt |
| add text conditioning | Marigold · NYUv2 small areas (<5% of image) | AbsRel (%): 9.0 -> 8.3 | Improved |
| add text conditioning | Marigold · NYUv2 standard | AbsRel (%): 6.1 -> 5.9 | Improved |
| add text conditioning (fewer denoising steps) | Marigold · Inference denoising steps | convergence steps: 25 steps -> 10 steps | Improved |
Closest papers by meaning
- Repurposing Diffusion-Based Image Generators for Monocular Depth Estimation (CVPR 2024 · similarity 0.90)
- WorDepth: Variational Language Prior for Monocular Depth Estimation (CVPR 2024 · similarity 0.90)
- More Than Generation: Unifying Generation and Depth Estimation via Text-to-Image Diffusion Models (NeurIPS 2025 · similarity 0.89)
- DepthFM: Fast Generative Monocular Depth Estimation with Flow Matching (AAAI 2025 · similarity 0.89)
- Pixel-Perfect Depth with Semantics-Prompted Diffusion Transformers (NeurIPS 2025 · similarity 0.88)
- On the Robustness of Language Guidance for Low-Level Vision Tasks: Findings from Depth Estimation (CVPR 2024 · similarity 0.88)
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.
