datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
MIT-Indoor-ScenesIndoorSceneRecognition
Dataset Card for IndoorSceneRecognition
The database contains 67 Indoor categories, and a total of 15620 images. The number of images varies across categories, but there are at least 100 images per category. All images are in jpg format.
This is a FiftyOne dataset with 15620 samples.
Installation
If you haven't already, install FiftyOne:
pip install -U fiftyone
Usage
import fiftyone as fo
import fiftyone.utils.huggingface as fouh
# Load the dataset… See the full description on the dataset page: https://huggingface.co/datasets/Voxel51/IndoorSceneRecognition.indoor-safety-hazard-detection-and-work-zone-monitoring
Indoor Safety Hazard Detection & Work-Zone Monitoring
Generated by datapack-import.ts
This dataset mirrors public data-pack render outputs from Physicl.
Each row represents one render view. The image column contains a stable URL to the primary render image uploaded under /data; image_path stores the relative repository path and data_commit_sha pins the Hugging Face dataset commit used by those URLs. Files are uploaded as downloaded unless optional PNG recompression is enabled by… See the full description on the dataset page: https://huggingface.co/datasets/physicl/indoor-safety-hazard-detection-and-work-zone-monitoring.infinigen2-flying-indoors
Infinigen2 Flying Indoors
Click here to see a 150 scene video preview on YouTube
This release provides 16000 stereo videos. Each is 24 frames in duration and has ground truth annotations for several computer vision tasks.
Each group of 4 stereo videos provides random synchronized views of the same procedural dynamic 3D scene.
This datarelease is split into three parts, based on release date:
Part A - released 2026-08-19 - 2000 stereo videos, using a… See the full description on the dataset page: https://huggingface.co/datasets/infinigen/infinigen2-flying-indoors.IndoorCAD
IndoorCAD Dataset
This repo contains our IndoorCAD Dataset.There are 4400+ Furniture models in the Furniture_Data part of the dataset,each has CAD model in STEP form,B-rep information,mesh model and 32-angle multi-view picture,text discription generated by qwenVLM and alignment label.There are 4500+ scene model in the Scene_Data part of the dataset,wich include CAD model in STEP form and pictures of different perspective.
And we provide a small sample of the dataset in… See the full description on the dataset page: https://huggingface.co/datasets/anon-neurips-2026/IndoorCAD.indoor-anomaly-detection-path-obstruction-monitoring
Indoor Anomaly Detection & Path Obstruction Monitoring
Generated by datapack-import.ts
This dataset mirrors public data-pack render outputs from Physicl.
Each row represents one render view. The image column contains a stable URL to the primary render image uploaded under /data; image_path stores the relative repository path and data_commit_sha pins the Hugging Face dataset commit used by those URLs. Files are uploaded as downloaded unless optional PNG recompression is enabled… See the full description on the dataset page: https://huggingface.co/datasets/physicl/indoor-anomaly-detection-path-obstruction-monitoring.3d-front-indoor-renders
Indoor Scene Renders from 3D-FRONT / 3D-FUTURE
20,240 photo-realistic indoor scene renders with instance segmentation, 6DoF
object poses, camera intrinsics and depth — rendered from the 3D-FRONT scene
layouts and 3D-FUTURE furniture models.
This is not a copy of the original 3D-FUTURE render set. It is a separate
render set built from the same assets. See Differences from the original below.
Why this exists
The 3D-FUTURE technical report describes 20,240 rendered… See the full description on the dataset page: https://huggingface.co/datasets/Spatial1ntelligence/3d-front-indoor-renders.spaq-cityscape-indoor-scene
SPAQ Cityscape + Indoor Scene Export
Images are copied without modification from SPAQ.
Selection rule: Cityscape > 0 OR Indoor scene > 0 from Scene category labels.xlsx.
Corrupt or missing images are skipped and recorded in failed_files.jsonl.
ITLP-Campus-Indoor🏢 ITLP Campus Indoor is a multimodal dataset focused on indoor Place Recognition across five floors of a university building. It features synchronized RGB images from front and back cameras, LiDAR point clouds, manually annotated scene text, and strategically placed ArUco markers for accurate localization. Semantic segmentation masks were automatically generated using the OneFormer model. Captured during night and twilight conditions, the dataset reflects real-world challenges in indoor… See the full description on the dataset page: https://huggingface.co/datasets/OPR-Project/ITLP-Campus-Indoor.UAV-IndoorCL
UAV-IndoorCL
This repository contains the dataset presented in the paper Learning on the Fly: Replay-Based Continual Object Perception for Indoor Drones.
Project Page | GitHub
Dataset Summary
UAV-IndoorCL is an indoor video dataset consisting of 14,400 frames capturing inter-drone and ground vehicle footage. It was specifically designed to support and benchmark Class-Incremental Learning (CIL) research for resource-constrained aerial platforms. The frames were… See the full description on the dataset page: https://huggingface.co/datasets/kaochuang/UAV-IndoorCL.train_dataADS-B Training Dataset (dump1090 Output)
Overview
This dataset contains ADS-B data captured using the dump1090 program.
The data is organized into multiple directories, each representing a separate data collection session or environment.
The dataset is designed for use in training and evaluating models related to aircraft signal analysis, localization,
or reception quality prediction.
Structure
The dataset directory has the following structure:
train_data/
│
├── ali/
│ ├── home/
│… See the full description on the dataset page: https://huggingface.co/datasets/IndoorOutdoor/train_data.indoor_mmwave
Indoor FireRescue Rada (IFR) dataset
The Indoor FireRescue Rada (IFR) dataset is a novel firerescue dataset containing 27,000 frames of synchronized and calibrated
16-layer LiDAR-, RGB camera-, and 4D radar-data (raw ADC data and point clouds) acquired in multiple buildings in UBC.
It consists of 3D bounding box annotations for building layout objects such as doors, waste containers, chairs, desks, fire hydrants.
Example scenario from the Indoor FireRescue Rada (IFR)… See the full description on the dataset page: https://huggingface.co/datasets/yysd123/indoor_mmwave.indoor-scene-classification
Dataset Labels
['meeting_room', 'cloister', 'stairscase', 'restaurant', 'hairsalon', 'children_room', 'dining_room', 'lobby', 'museum', 'laundromat', 'computerroom', 'grocerystore', 'hospitalroom', 'buffet', 'office', 'warehouse', 'garage', 'bookstore', 'florist', 'locker_room', 'inside_bus', 'subway', 'fastfood_restaurant', 'auditorium', 'studiomusic', 'airport_inside', 'pantry', 'restaurant_kitchen', 'casino', 'movietheater', 'kitchen', 'waitingroom', 'artstudio', 'toystore'… See the full description on the dataset page: https://huggingface.co/datasets/keremberke/indoor-scene-classification.indoor-scene-state-environmental-context-understanding-next-pack-0c894b1f-effc97c5
Indoor Robot Navigation and Toy Grasping
Training dataset for a mobile robot in home environments (kids rooms and playrooms). Renders show furniture to navigate between and toys to grasp, with metric depth and world-space normals for contact geometry, plus per-frame annotations, at 1024x1024 with the environments' authored lighting.
This dataset mirrors public data-pack render outputs from Physicl.
Each row represents one render view. The image column contains a stable URL to… See the full description on the dataset page: https://huggingface.co/datasets/physicl-community/indoor-scene-state-environmental-context-understanding-next-pack-0c894b1f-effc97c5.bb_indoor_pullThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "so100_follower",
"total_episodes": 59,
"total_frames": 10450,
"total_tasks": 1,
"total_videos": 236,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:59"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/andlyu/bb_indoor_pull.IndoorCrowd
Dataset Card for IndoorCrowd
Dataset Summary
IndoorCrowd is a multi-scene dataset designed for indoor human detection, instance segmentation, and multi-object tracking. It captures diverse challenges such as viewpoint variation, partial occlusion, and varying crowd density across four distinct campus locations (ACS-EC, ACS-EG, IE-Central, R-Central). Faces are explicitly blurred to preserve privacy, making it suitable for safe research into intelligent crowd… See the full description on the dataset page: https://huggingface.co/datasets/sebnae/IndoorCrowd.IndoorOutdoorNet-20K
IndoorOutdoorNet-20K
IndoorOutdoorNet-20K is a labeled image dataset designed for the task of image classification, particularly focused on distinguishing between indoor and outdoor scenes. The dataset is publicly available on Hugging Face Datasets and is useful for scene understanding, transfer learning, and model benchmarking.
Dataset Summary
Task: Image Classification
Modalities: Image
Labels: Indoor, Outdoor (2 classes)
Total Images: 19,998
Split: Train (100%)… See the full description on the dataset page: https://huggingface.co/datasets/prithivMLmods/IndoorOutdoorNet-20K.secret_leaderboard_resultsreside-indoorIndoorVG-coco-format
IndoorVG — Indoor Visual Genome (COCO format)
IndoorVG is a curated split of
Visual Genome
targeting real-world indoor scenarios (kitchens, offices, living rooms, …).
It was proposed in
Neau et al. (2024)
and reformatted here in standard COCO-JSON format.
It was produced as part of the
SGG-Benchmark framework and used to train
the models described in the REACT paper
(Neau et al., BMVC 2025).
The 84 object classes and 37 predicate classes were manually selected and
semi-automatically… See the full description on the dataset page: https://huggingface.co/datasets/maelic/IndoorVG-coco-format.pintest_indoor_2025table_indoor-object3D-SynthPlace_indoor_scenes_dataset
3D-SynthPlace indoor scene dataset in OptiScene (NeurIPS2025)
This is the 3D-SynthPlace dataset in the paper OptiScene: LLM-driven Indoor Scene Layout Generation via Scaled Human-aligned Data Synthesis and Multi-Stage Preference Optimization (NeurIPS2025).
3D-SynthPlace dataset JSON File Format Specification
The basic format of the scene description is JSON. This format is used to describe room floor and interior object layouts. You can refer prompts_all_scenes.json to… See the full description on the dataset page: https://huggingface.co/datasets/B3rrYang/3D-SynthPlace_indoor_scenes_dataset.Indoor_Real_1024mit-indoor-outdoorWIFI_RSSI_Indoor_Positioning_Dataset
WIFI RSSI Indoor Positioning Dataset
A reliable and comprehensive public WiFi fingerprinting database for researchers to implement and compare the indoor localization’s methods.The database contains RSSI information from 6 APs conducted in different days with the support of autonomous robot.
We use an autonomous robot to collect the WiFi fingerprint data. Our 3-wheel robot has multiple sensors including wheel odometer, an inertial measurement unit (IMU), a LIDAR, sonar sensors and a… See the full description on the dataset page: https://huggingface.co/datasets/Brosnan/WIFI_RSSI_Indoor_Positioning_Dataset.umi-okra-indoor-20260804
UMI Okra Harvesting — Indoor Session (2026-08-04)
Unitree Dex1-1 グリッパに改造した UMI (Universal Manipulation Interface)
で、屋内に並べた観葉植物を対象にオクラ収穫動作を撮影したデモンストレーションデータ。
English summary: Human demonstrations of okra-harvesting motions recorded indoors with a
hand-held UMI gripper (modified with a Unitree Dex1-1) and a GoPro. Contains a
diffusion-policy-ready replay buffer (14 episodes / 3,500 frames) plus the raw ORB-SLAM3
trajectories and calibration files. Read the Limitations section… See the full description on the dataset page: https://huggingface.co/datasets/Kota0612/umi-okra-indoor-20260804.bb_indoor_augpull_train_v2indoor-laptop-detection-next-pack-e6bd6a07-356e7b31
Indoor Plant Detection for YOLO
Synthetic training dataset of 5 renders (640x640) staged in home and office interiors — libraries and residential lounges — each framing an indoor plant. Includes RGB, albedo and metric depth passes with per-frame bounding-box annotations, for training a YOLO object-detection model to detect indoor plants.
This dataset mirrors public data-pack render outputs from Physicl.
Each row represents one render view. The image column contains a stable URL… See the full description on the dataset page: https://huggingface.co/datasets/physicl-community/indoor-laptop-detection-next-pack-e6bd6a07-356e7b31.Stairs_Image_Dataset_Parts_of_House_Indoor
Stairs Image Dataset — Parts of House Indoor (Sample)
⚠️ This is a free sample subset (100 images) for evaluation purposes only.The full dataset (3,000+ HD images) is available for commercial licensing.Contact: sales@datacluster.ai · datacluster.ai
Dataset Summary
This dataset is an extremely challenging collection of original stair images, crowdsourced from over 500 urban and rural areas. Every image is manually reviewed and verified by computer vision… See the full description on the dataset page: https://huggingface.co/datasets/Dataclusterlabspvtltd/Stairs_Image_Dataset_Parts_of_House_Indoor.