mlsanityData to 14 Oct 2026

Pixel-Perfect Depth with Semantics-Prompted Diffusion Transformers

arXiv 2510.07316 ↗First posted 8 Oct 2025cs.CVcited by 51

On the map: Diffusion models → 3D, point cloud and scene generation

Abstract (arXiv)

This paper presents Pixel-Perfect Depth, a monocular depth estimation model based on pixel-space diffusion generation that produces high-quality, flying-pixel-free point clouds from estimated depth maps. Current generative depth estimation models fine-tune Stable Diffusion and achieve impressive performance. However, they require a VAE to compress depth maps into latent space, which inevitably introduces textitflying pixels at edges and details. Our model addresses this challenge by directly performing diffusion generation in the pixel space, avoiding VAE-induced artifacts. To overcome the high complexity associated with pixel-space generation, we introduce two novel designs: 1) Semantics-Prompted Diffusion Transformers (SP-DiT), which incorporate semantic representations from vision foundation models into DiT to prompt the diffusion process, thereby preserving global semantic consistency while enhancing fine-grained visual details; and 2) Cascade DiT Design that progressively increases the number of tokens to further enhance efficiency and accuracy. Our model achieves the best performance among all published generative models across five benchmarks, and significantly outperforms all other models in edge-aware point cloud evaluation.

Key facts. Cited by 51 papers in the map (data through 14 Oct 2026). Placed in 3D, point cloud and scene generation, one of 14 directions in Diffusion models. Builds on 84 of its 98 references that are in the map. First posted on arXiv 8 Oct 2025.

Builds on

84 of 98 references.

Closest ancestors · 12

Same direction, newest first

  1. 2025From Editor to Dense Geometry Estimator4 Sep
  2. 2025GeometryCrafter: Consistent Geometry Estimation for Open-world Videos with Diffusion Priors1 Apr
  3. 2025Jasmine: Harnessing Diffusion Prior for Self-supervised Depth Estimation20 Mar
  4. 2025DepthMaster: Taming Diffusion Models for Monocular Depth Estimation5 Jan
  5. 2024FiffDepth: Feed-forward Transformation of Diffusion-Based Generators for Detailed Depth Estimation1 Dec
  6. 2024Depth Any Video with Scalable Synthetic Data14 Oct
  7. 2024Learning Temporally Consistent Video Depth from Video Diffusion Priors3 Jun
  8. 2024GeoWizard: Unleashing the Diffusion Priors for 3D Geometry Estimation from a Single Image18 Mar
  9. 2023Zero-Shot Metric Depth with a Field-of-View Conditioned Diffusion Model20 Dec
  10. 2023Repurposing Diffusion-Based Image Generators for Monocular Depth Estimation4 Dec
  11. 2023The Surprising Effectiveness of Diffusion Models for Optical Flow and Monocular Depth Estimation2 Jun
  12. 2023Monocular Depth Estimation using Diffusion Models28 Feb

Other references · 72

  1. 2020Denoising Diffusion Probabilistic Modelscited by 21,076
  2. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,863
  3. 2020Denoising Diffusion Implicit Modelscited by 9,781
  4. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,974
  5. 2021Masked Autoencoders Are Scalable Vision Learnerscited by 7,403
  6. 2022Scalable Diffusion Models with Transformerscited by 6,054
  7. 2022Classifier-Free Diffusion Guidancecited by 5,847
  8. 2022Flow Matching for Generative Modelingcited by 5,702
  9. 2023SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesiscited by 4,356
  10. 2024Scaling Rectified Flow Transformers for High-Resolution Image Synthesiscited by 4,181
  11. 2022Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flowcited by 3,700
  12. 2017ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenescited by 3,117
38 more
  1. 2024CogVideoX: Text-to-Video Diffusion Models with An Expert Transformercited by 2,261
  2. 2021Vision Transformers for Dense Predictioncited by 1,803
  3. 2025VGGT: Visual Geometry Grounded Transformercited by 1,539
  4. 2019Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-shot Cross-dataset Transfercited by 1,495
  5. 2024Depth Anything V2cited by 1,422
  6. 2024Depth Anything: Unleashing the Power of Large-Scale Unlabeled Datacited by 1,360
  7. 2014Depth Map Prediction from a Single Image using a Multi-Scale Deep Networkcited by 1,274
  8. 2021Cascaded Diffusion Models for High Fidelity Image Generationcited by 1,163
  9. 2022Building Normalizing Flows with Stochastic Interpolantscited by 995
  10. 2018MegaDepth: Learning Single-View Depth Prediction from Internet Photoscited by 737
  11. 2024Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Thinkcited by 703
  12. 2023ZoeDepth: Zero-shot Transfer by Combining Relative and Metric Depthcited by 657
  13. 2020Hypersim: A Photorealistic Synthetic Dataset for Holistic Indoor Scene Understandingcited by 604
  14. 2020TartanAir: A Dataset to Push the Limits of Visual SLAMcited by 506
  15. 2019BlendedMVS: A Large-scale Dataset for Generalized Multi-view Stereo Networkscited by 490
  16. 2014Predicting Depth, Surface Normals and Semantic Labels with a Common Multi-Scale Convolutional Architecturecited by 483
  17. 2023Simple diffusion: End-to-end diffusion for high resolution imagescited by 422
  18. 2024Depth Pro: Sharp Monocular Metric Depth in Less Than a Secondcited by 404
  19. 2024Metric3Dv2: A Versatile Monocular Geometric Foundation Model for Zero-shot Metric Depth and Surface Normal Estimationcited by 395
  20. 2025Reconstruction vs. Generation: Taming Optimization Dilemma in Latent Diffusion Modelscited by 393
  21. 2024UniDepth: Universal Monocular Metric Depth Estimationcited by 382
  22. 2023Metric3D: Towards Zero-shot Metric 3D Prediction from A Single Imagecited by 365
  23. 2020Virtual KITTI 2cited by 353
  24. 2021Omnidata: A Scalable Pipeline for Making Multi-Task Mid-Level Vision Datasets from 3D Scanscited by 336
  25. 2024MoGe: Unlocking Accurate Monocular Geometry Estimation for Open-Domain Images with Optimal Training Supervisioncited by 299
  26. 2023Unleashing Text-to-Image Diffusion Models for Visual Perceptioncited by 269
  27. 2025MoGe-2: Accurate Monocular Geometry with Metric Scale and Sharp Detailscited by 261
  28. 2019DIODE: A Dense Indoor and Outdoor DEpth Datasetcited by 240
  29. 2020Learning to Recover 3D Scene Shape from a Single Imagecited by 216
  30. 2025Video Depth Anything: Consistent Depth Estimation for Super-Long Videoscited by 207
  31. 2024DepthCrafter: Generating Consistent Long Depth Sequences for Open-world Videoscited by 204
  32. 2023Iterative Geometry Encoding Volume for Stereo Matchingcited by 190
  33. 2024Lotus: Diffusion-based Visual Foundation Model for High-quality Dense Predictioncited by 166
  34. 2023DDP: Diffusion Model for Dense Visual Predictioncited by 148
  35. 2021SMD-Nets: Stereo Mixture Density Networkscited by 120
  36. 2023Towards Zero-Shot Scale-Aware Monocular Depth Estimationcited by 101
  37. 2024Selective-Stereo: Adaptive Frequency Information Selection for Stereo Matchingcited by 99
  38. 2023DiffusionDepth: Diffusion Denoising Approach for Monocular Depth Estimationcited by 82

50 most cited of 72 All 72 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 3; 2026: 48.

51 papers.

Sort:

By area:

  1. 2026NeoVerse: Enhancing 4D World Model with in-the-wild Monocular Videoscited by 46
  2. 2026InfiniDepth: Arbitrary-Resolution and Fine-Grained Depth Estimation with Neural Implicit Fieldscited by 23
  3. 20264RC: 4D Reconstruction via Conditional Querying Anytime and Anywherecited by 22
  4. 2026DriveDreamer-Policy: A Geometry-Grounded World-Action Model for Unified Generation and Planningcited by 17
  5. 2026MoGe-3: Fine-Detail Monocular Geometry Estimation with Self-Guided Sparse Volumetric Refinementcited by 8
  6. 2026PointDiT: Pixel-Space Diffusion for Monocular Geometry Estimationcited by 7
  7. 2026PromptStereo: Zero-Shot Stereo Matching via Structure and Motion Promptscited by 6
  8. 2026Driving with DINO: Vision Foundation Features as a Unified Bridge for Sim-to-Real Generation in Autonomous Drivingcited by 5
  9. 2026MoE3D: A Mixture-of-Experts Module for 3D Reconstructioncited by 4
  10. 2026Modality Forcing for Scalable Spatial Generationcited by 4
  11. 2026UniX: Unifying Autoregression and Diffusion for Chest X-Ray Understanding and Generationcited by 3
  12. 2026DAGE: Dual-Stream Architecture for Efficient and Fine-Grained Geometry Estimationcited by 3
38 more
  1. 2026FoundationGeo: Learning Spatial Pixel-Wise Fields for Monocular Metric Geometrycited by 3
  2. 2026Guardians of the Hair: Rescuing Soft Boundaries in Depth, Stereo, and Novel Viewscited by 2
  3. 2026SurGe: Improved Surface Geometry in Point Mapscited by 2
  4. 2026Modeling Depth Ambiguity: A Mixture-Density Representation for Flying-Point-Free Depth Estimationcited by 2
  5. 2025How Robot Dogs See the Unseeable: Improving Visual Interpretability via Peering for Exploratory Robotscited by 1
  6. 2025Boosting Monocular Metric Depth Estimation via Bokeh Renderingcited by 1
  7. 2026Iris: Bringing Real-World Priors into Diffusion Model for Monocular Depth Estimationcited by 1
  8. 2026WildPose: A Unified Framework for Robust Pose Estimation in the Wildcited by 1
  9. 2026Unlocking Diffusion Hierarchies: Adaptive Timestep Selection for Zero-Shot Segmentationcited by 1
  10. 2026SurroundNEXO: Ego-Centric Metric Bridging for Spatially Consistent Geometry in Autonomous Drivingcited by 1
  11. 2026Pixel-Space Diffusion Transformerscited by 1
  12. 2026PXDepth: Pixel-Space Modeling for Structure Preserving Monocular Depth Estimationcited by 1
  13. 2026OptiGeo: Efficient Monocular Geometry for Embodied Perception in Optically Challenging Scenescited by 1
  14. 2026Efficient and High-Quality Depth Estimation via Pixel-Space Diffusion with Linear Attentioncited by 1
  15. 2026Marigold V2: Revisiting Diffusion Transformers for Monocular Depth Estimationcited by 1
  16. 2025Generalized Geometry Encoding Volume for Real-time Stereo Matchingcited by 0
  17. 2026GVGS: Gaussian Visibility-Aware Multi-View Geometry for Accurate Surface Reconstructioncited by 0
  18. 2026Need for Speed: Zero-Shot Depth Completion with Single-Step Diffusioncited by 0
  19. 2026Divide and Conquer: Decoupled Representation Alignment for Multimodal World Modelscited by 0
  20. 2026GemDepth: Geometry-Embedded Features for 3D-Consistent Video Depthcited by 0
  21. 2026VolFill: Single-View Amodal 3D Scene Reconstruction with Volumetric Flow Matchingcited by 0
  22. 2026αDepth: Learning Single-Pass Soft Boundary Decomposition for Stereo Conversioncited by 0
  23. 2026Vision-Reasoning-Guided Occlusion Removal from Light Fieldscited by 0
  24. 2026DermDepth: Toward Monocular Metric Scale 3D Reconstruction Models for Dermatologycited by 0
  25. 2026STEREOFLOW: Progressive Stereo Matching with StereoDiT and Transition Flow Matchingcited by 0
  26. 2026GraspHOI: Full-Body 3D Human-Object Reconstruction with Finger-Level Grasps from a Single In-the-Wild Imagecited by 0
  27. 2026RealOOB: A Definition-Consistent Real-World Oriented Occlusion Boundary Benchmarkcited by 0
  28. 2026Monocular Depth Estimation from a Single Image: Progress and Opportunitiescited by 0
  29. 2026Zero-Shot Novel Depth Synthesis Using 3D Foundation Models Scene Representationscited by 0
  30. 2026TransNormal-2: Geometry-Grounded Rectified Flow with Edge-Aware Decoding for Precise Normal Estimationcited by 0
  31. 2026Mira-Scene: Pixel-Aligned Layouts for Generative 3D Scene Reconstructioncited by 0
  32. 2026HelloWorld: Towards Practical Applications of Generative Driving World Modelscited by 0
  33. 2026SkillWeaver: Agentic Exploration over Neural Interaction Skills for Scalable Robot Data Generationcited by 0
  34. 2026DMA2: Pixel-space Distribution Matching with Adversarial and Anchor Lossescited by 0
  35. 2026EagleDepth: Efficient Fine-Grained Depth Estimation via Pixel Diffusion Decodercited by 0
  36. 2026DistScene: Object-to-Scene Distillation for 3D Scene Generationcited by 0
  37. 2026CRISP: Fixing Flying Pixels in Latent LiDAR Generation via Diffusion Decodingcited by 0
  38. 2026ScenePlacer: Reformulating Scene-Level Object Pose Estimation as Visibility Field Learningcited by 0

50 most cited of 51 All 51 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2026Pixel-Perfect Visual Geometry Estimation
  2. 2024Diffusion Models for Monocular Depth Estimation: Overcoming Challenging Conditions
  3. 2025Leveraging Stable Diffusion for Monocular Depth Estimation via Image Semantic Encoding
  4. 2025Guided and Unguided Conditional Diffusion Mechanisms for Structured and Semantically-Aware 3D Point Cloud Generation
  5. 2024Iris: Integrating Language into Diffusion-based Monocular Depth Estimation
  6. 2026ARDepth: Auto-regressive Monocular Depth Estimation with Progressive Visual Conditioning
  7. 2024Novel View Synthesis with Pixel-Space Diffusion Models
  8. 2023MVDD: Multi-View Depth Diffusion Models