datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
eccv2026-cad-challenge-data
ECCV 2026 CAD Challenge Data
This challenge is part of the workshop The Path to Manufacturing: Evolving
3D Generation to Intelligent Computer-Aided Design.
Workshop homepage: https://3dgen-cad-workshop.github.io/
Challenge submission Space: https://huggingface.co/spaces/jingwei-xu-00/eccv2026-cad-challenge
Dataset rendering and preparation code (only .step files are required): https://github.com/DavidXu-JJ/eccv2026-cad-challenge-data-render
This repository contains the public… See the full description on the dataset page: https://huggingface.co/datasets/jingwei-xu-00/eccv2026-cad-challenge-data.ECCV_Event_Video_Depth_Estimation
Event-Guided Video Depth Estimation Workshop Dataset
This dataset is a mirrored and aligned workshop-ready version of the DVD event-guided video depth estimation data.
It packages each scene into a canonical folder tree that aligns:
low-light RGB frames
per-frame event slices
a scene-level lowlight_event.npz
the matched depth ground truth copied from inference_results/*/normal/depth.npz
The dataset is designed for direct upload to Hugging Face as a dataset repository.
The official… See the full description on the dataset page: https://huggingface.co/datasets/Ethanliang99/ECCV_Event_Video_Depth_Estimation.AerialMetric-ECCV2026
AerialMetric-ECCV2026
This repository is the main dataset hub for the paper AerialMetric: Benchmarking and Adapting UAV Monocular Metric Depth Estimation in the Real World.
It includes more than 65k aerial image-depth pairs, and all depth maps are metric.
Project homepage: https://kuieless.github.io/AerialMetric-ECCV2026-page/
Repository: https://github.com/kuieless/AerialMetric-ECCV2026
It contains the training data, test data, and the moge2-aerial weights used by the… See the full description on the dataset page: https://huggingface.co/datasets/Kuiee/AerialMetric-ECCV2026.papercli-papers-eccveccv2026-cad-challenge-data
ECCV 2026 CAD Challenge Data
This challenge is part of the workshop The Path to Manufacturing: Evolving
3D Generation to Intelligent Computer-Aided Design.
Workshop homepage: https://3dgen-cad-workshop.github.io/
Challenge submission Space: https://huggingface.co/spaces/jingwei-xu-00/eccv2026-cad-challenge
This repository contains the public data package for the challenge. The
evaluation Space accepts STEP predictions for the private evaluation split and
updates the leaderboard… See the full description on the dataset page: https://huggingface.co/datasets/Qiao123rvvr/eccv2026-cad-challenge-data.SYNTOM
SYNTOM: Synthetic Tomato Greenhouse Segmentation
77,217 photorealistic renders of greenhouse tomato plants (68,328 train / 8,889 val,
1920x1080) with pixel perfect ground truth for two tasks:
Semantic segmentation: 4 organ classes plus background, single channel PNG masks
Instance segmentation: whole plant instances in COCO format
Released with Text-conditioned Segmentation for Tomato Phenotyping via Procedural
Synthetic Data, where it is used to fine-tune SAM 3
for… See the full description on the dataset page: https://huggingface.co/datasets/ECCV26-Tomato-Phenotyping/SYNTOM.EgoDyn-Bench-ECCV2026
EgoDyn-Bench
A physics-grounded VQA benchmark for evaluating Vision-Language Models on trajectory-based dynamics reasoning in autonomous driving.
Project page | Paper | GitHub
This repository contains the data artifacts for the benchmark. The evaluation harness, baselines, and reference implementations live in the companion GitHub repository.
Note on licensing. The nuScenes-derived portion of this dataset is released under CC BY-NC-SA 4.0 to comply with nuScenes' upstream… See the full description on the dataset page: https://huggingface.co/datasets/TUM-AVS/EgoDyn-Bench-ECCV2026.safesae-vla-eccv2026
SafeSAE-VLA: Rollouts and SAE Checkpoint (ECCV 2026)
Artifacts for the paper "SafeSAE-VLA: Interpreting OpenVLA Progress Dynamics with
Sparse Feature Analysis" (ECCV 2026), by Socrates Osorio and Joy Zheyun Yang.
Code: https://github.com/socratesosorio/safesae-vla-eccv2026
This repository provides the OpenVLA rollouts and the trained sparse autoencoder (SAE)
used in the paper's progress analysis, so the reported results can be reproduced.
Contents
rollouts/… See the full description on the dataset page: https://huggingface.co/datasets/socratesosorio/safesae-vla-eccv2026.eccv-testECCV2024-papersMedVidU_ECCV2026_TrainVal
ECCV 2026 Workshop on Medical Video Understanding (MedVidU @ ECCV 2026) — Train / Val Split
This is the public train / val split for the MedVidU Challenge at the ECCV 2026 Workshop on Medical Video Understanding. This split is derived from the benchmark introduced in MedGRPO: Multi-Task Reinforcement Learning for Heterogeneous Medical Video Understanding (CVPR 2026).
Participants are free to use this split for any combination of training and local validation.
Final challenge… See the full description on the dataset page: https://huggingface.co/datasets/UII-AI/MedVidU_ECCV2026_TrainVal.ECCV2026compositionality_eccv_captioneccv_doll_moveECCV26-ARAECCV_OCR
ECCV_OCR
OCR Data split by Venue.
eccv_rebuttaleccv2024metasurface-real-eccv2026
Metasurface Depth: real training and evaluation scenes
Real encoded image pairs and the corresponding original depth-label arrays used
by the selected mixed-training checkpoints for
Physically Grounded Monocular Depth via Nanophotonic Wavefront Encoding.
Contents
Split
Scenes
Input PNGs
Depth NPYs
Stored resolution (height × width)
train
5
10
5
1200 × 1600
test
42
84
42
1190 × 1596
The archive contains 141 original data files plus portable CSV… See the full description on the dataset page: https://huggingface.co/datasets/Bingxuan111/metasurface-real-eccv2026.ECCV2024vid_motion_mag_eccv18ECCV26_PhysAI_Challenge_NVS_Syn4D_subset
ECCV'26 PhysAI Challenge — Novel-View Synthesis (Syn4D subset)
A small, self-contained novel-view-synthesis (NVS) evaluation package for the
ECCV'26 PhysAI workshop challenge (video-diffusion NVS under camera control). It ships
the held-out Syn4D inputs (source rgb + depth + camera + segmentation mask) and the
matching RecamMaster raw-resolution predictions, so you can reproduce the reference
metrics end-to-end with no dataset access and no inference.
👉 The… See the full description on the dataset page: https://huggingface.co/datasets/yslan/ECCV26_PhysAI_Challenge_NVS_Syn4D_subset.AerialMetric-ECCV2026
AerialMetric-ECCV2026
This repository is the main dataset hub for the paper AerialMetric: Benchmarking and Adapting UAV Monocular Metric Depth Estimation in the Real World.
It includes more than 65k aerial image-depth pairs, and all depth maps are metric.
Project homepage: https://kuieless.github.io/AerialMetric-ECCV2026-page/
Repository: https://github.com/kuieless/AerialMetric-ECCV2026
It contains the training data, test data, and the moge2-aerial weights used by the… See the full description on the dataset page: https://huggingface.co/datasets/Addy0811/AerialMetric-ECCV2026.LAM_Cases_For_ECCVECCV2020ECCV2018ECCV2022scout-eccv-pab-annotations
PAB hard-negative-pair annotations
Supporting data for SCOUT (Sim-to-Real Text-Based Person Retrieval by Embedding-Space Prediction
over Frozen Video Features), an ECCV 2026 workshop paper on AI City Challenge Track 4 (Text-Based
Person Re-Identification, Sim2Real). Code: https://github.com/abtraore/SCOUT-ECCV
What this is
The AI City Challenge Track 4 release of PAB (Pedestrian Anomaly Behavior) strips several fields
from the dataset authors' original CMP… See the full description on the dataset page: https://huggingface.co/datasets/Abdrah/scout-eccv-pab-annotations.attribute_project_eccvECCV_2024_OCR
ECCV_2024_OCR
OCR Data.