Deva Ramanan
Professor
Studies computer vision, including 3D perception for autonomous vehicles, recognition and dynamic scene reconstruction.
Deva Ramanan's lab at CMU works on 3D vision & neural rendering, embodied & active vision and robot perception & SLAM. 77 papers since 2023, 63 at top venues, mostly CVPR and ECCV. 7 PhD students. 50 alumni. Applications close .
At a glance
- papers, 2023–2026
- 20/ year
- at T1 venues
- 82%
- PhD students
- 7
- MS students
- 0
- alumni listed
- 50
Papers by yearfrom the lab homepage, 2023–2026
- 16 papers2026 · T1 14 · collecting
- 21 papers2025 · T1 17
- 18 papers2024 · T1 15
- 22 papers2023 · T1 17
T1 venuesOtherVenues: CVPR, ECCV, NeurIPS, 3DV, ICCV
Research focuspaper titles that use each topic's words
- 3D vision & neural rendering19 papers
- Embodied & active vision11 papers
- dynamic5 papers
- Robot perception & SLAM2 papers
- scene reconstruction1 paper
Paper list
From the lab homepage.
202616 papers
- CVPR (T1)Any4D: Unified Feed-Forward Metric 4D Reconstruction (opens in a new tab)
- CVPR (T1)Building a Precise Video Language with Human-AI Oversight (opens in a new tab)
- ICLR (T1)CRISP Contact-guided Real2Sim from Monocular Video with Planar Scene Primitives (opens in a new tab)
- ECCV (T1)DetPO: In-Context Learning with Multi-Modal LLMs for Few-Shot Object Detection (opens in a new tab)
- ECCV (T1)FixAnything: 3D-Consistent Rendering Refinement via Video Generative Priors (opens in a new tab)
- ECCV (T1)FrameCrafter: Novel View Synthesis as Video Completion (opens in a new tab)
- 3DVMapAnything: Universal Feed-Forward Metric 3D Reconstruction (opens in a new tab)
- ECCV (T1)Mechanistic Finetuning of Vision-Language-Action Models via Few-Shot Demonstrations (opens in a new tab)
- CVPR (T1)PAI-Bench: A Comprehensive Benchmark For Physical AI (opens in a new tab)
- NeurIPS (T1)Point4D: Long-range 4D Motion Reconstruction (opens in a new tab)
- NeurIPS (T1)PointZero: 3D Point Track Completion for Learning Transferable 3D Dynamics (opens in a new tab)
- 3DVRadarSim: Simulating Single-Chip Radar via Multimodal Neural Fields
202521 papers
- CVPR (T1)AerialMegaDepth: Learning Aerial-Ground Reconstruction and View Synthesis (opens in a new tab)
- ICRA (T1)BETTY Dataset: A Multi-modal Dataset for Full-Stack Autonomy (opens in a new tab)
- NeurIPS (T1)CameraBench: Towards Understanding Camera Motions in Any Video (opens in a new tab)
- CVPR (T1)DiffusionSfM: Predicting Structure and Motion via Ray Origin and Endpoint Diffusion (opens in a new tab)
- 3DVDressRecon: Freeform 4D Human Reconstruction from Monocular Video (opens in a new tab)
- 3DVDynOMo: Online Point Tracking by Dynamic Online Monocular Gaussian Reconstruction" (opens in a new tab)
- ICCV (T1)Enhancing Few-Shot Vision-Language Classification with Large Multimodal Model Features (opens in a new tab)
- ICCV. [Best Paper Award]Generating Physically Stable and Buildable Brick Structures from Text (opens in a new tab)
- ICCV (T1)MonoFusion: Sparse-View 4D Reconstruction via Monocular Fusion (opens in a new tab)
- ICLR (T1)Neural Eulerian Scene Flow Feilds (opens in a new tab)
- ICCV (T1)ONLY: One-Layer Intervention Sufficiently Mitigates Hallucinations in Large Vision-Language Models (opens in a new tab)
- ICRA (T1)Planning with Adaptive World Models for Autonomous Driving (opens in a new tab)
202418 papers
- ECCV (T1)Better Call SAL: Towards Learning to Segment Anything in Lidar (opens in a new tab)
- ICLR (T1)Cameras as Rays: Sparse-view Pose Estimation via Ray Diffusion (opens in a new tab)
- 3DVDynamic 3D Gaussians: Tracking by Persistent Dynamic View Synthesis (opens in a new tab)
- ECCV (T1)Evaluating Text-to-Visual Generation with Image-to-Text Generation (opens in a new tab)
- ECCV (T1)FlashTex: Fast Relightable Mesh Texturing with LightControlNet (opens in a new tab)
- CVPR (T1)HybridNeRF: Efficient Neural Rendering via Adaptive Volumetric Surfaces (opens in a new tab)
- ECCV (T1)I Can't Believe It's Not Scene Flow! (opens in a new tab)
- CVPR (T1)Language Models as Black-Box Optimizers for Vision-Language Models (opens in a new tab)
- IJCV (T1)Lidar Panoptic Segmentation in an Open World (opens in a new tab)
- NeurIPS (T1)NaturalBench: Evaluating Vision-Language Models on Natural Adversarial Samples (opens in a new tab)
- WACV. [Best Paper Finalist]Re-Evaluating LiDAR Scene Flow (opens in a new tab)
- 3DVRelPose++: Recovering 6D Poses from Sparse-view Observations (opens in a new tab)
202322 papers
- ICCV WorkshopAn Empirical Analysis of Range for 3D Object Detection (opens in a new tab)
- WACVBURST: A Benchmark for Unifying Object Recognition, Segmentation and Tracking in Video (opens in a new tab)
- CVPR (T1)Distilling Neural Fields for Real-Time Articulated Reconstruction from Video (opens in a new tab)
- WACVFar3Det: Towards Far-Field 3D Detection (opens in a new tab)
- ICCV (T1)Joint Metrics Matter: A Better Standard for Trajectory Forecasting (opens in a new tab)
- 3D-Aware Conditional Image Synthesis, CVPRK. Deng, G. Yang, D. Ramanan, J. Zhu
- ICML (T1)Learning Lightweight Object Detectors via Multi-Teacher Progressive Distillation (opens in a new tab)
- CVPR (T1)Learning to Zoom and Unzoom (opens in a new tab)
- IROS (T1)Lidar Panoptic Segmentation and Tracking without Bells and Whistles (opens in a new tab)
- MobiCom (T1)Low-Bandwidth Self-Improving Transmission of Rare Training Data (opens in a new tab)
- CVPR (T1)Multimodality Helps Unimodality: Cross-Modal Few-Shot Learning with Multimodal Models (opens in a new tab)
- CVPR (T1)Pix2map: Cross-modal Retrieval for Inferring Street Maps From Images (opens in a new tab)
Where alumni went50 listed
Apply to CMUFall 2027
CMU SCS Computer Science PhD
- Round
- Single round
- Opens
- GRE
- See program page
- English
- TOEFL / IELTS / Duolingo (F-1/J-1 non-native speakers)
- Fee
- USD 100 (USD 80 by early deadline)
Confirmed on the program page.
Similar labsnearest by research summary
- USUSC
Yue Wang
- 3D vision & neural rendering
- CHETH Zurich
Marc Pollefeys
- 3D vision & neural rendering
- GBImperial
Rolandos Potamias
- Embodied & active vision
- USGeorgia Tech
James Hays
- Recognition, detection & segmentation
- CNZhejiang
Qi Ye
- Embodied & active vision
- CNZhejiang
Guofeng Zhang
- 3D vision & neural rendering
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
CMUcloses