We support the open AI research community by providing free monthly sample packs. All public datasets are hosted permanently and censorship-resistant on the decentralized Autonomi Network.
Dataset Focus: FPV Mailbox Delivery & Hand-Paper Interaction (Sofia, Bulgaria)
Contents: High-FPS raw video footage capturing multi-sheet A6/A5/A4 flyer insertions, varied hand approaches, and outdoor entrance backgrounds.
Autonomi Address: 2e486330189f82bf6c0e0bb4d7b678b7f6881a120b3c8fc20de7f4a01e5e7a5c
Chunks: 277 (276 + 1 data map)
Size: 1.06 GB
Cost: 25.09 ANT (gas: 0.000032 ETH)
Time: 1492.6s
The Autonomi Network allows direct peer-to-peer data retrieval without subscriptions, central servers, or account creation.
Install the Autonomi Client
Download the latest ant executable for your operating system (Windows, macOS, or Linux) from the official Autonomi portal: https://autonomi.com/
For linux: curl -sSL https://raw.githubusercontent.com/WithAutonomi/ant-client/main/install.sh | bash
Open Terminal / Command Prompt
Navigate to your local storage folder:
Bash
cd /path/to/your/datasets
Execute the Download Command
Run the following command using the dataset's XOR address:
ant file download 2e486330189f82bf6c0e0bb4d7b678b7f6881a120b3c8fc20de7f4a01e5e7a5c -o drujba_1.zip
Access Your Files
Once the download completes, your folder will contain raw .avi video files.
Engineered specifically for Computer Vision (CV), Autonomous Robotics, Optical Character Recognition (OCR), and Human Action Recognition models.
Feature
Specification Details
Location
Sofia, Bulgaria (EU Residential & Urban Outdoor/Indoor Entrances)
Perspective
First-Person / Ego-centric (FPV Bodycam)
Primary Action
Fine-grained hand insertion of paper materials into outdoor metallic mailboxes
Media Formats
Multi-sheet glossy paper leaflets, promotional brochures, and flyers
Paper Sizes
European Standard Sizes: A6, A5, and A4 (Single to 2–3 sheets per drop)
Computer Vision Challenges
High motion blur, variable daylight, hand occlusions, paper friction/bending, Cyrillic text OCR
Email: dimitar@autonomidata.eu