Midv-180 -

: Thousands of individual frames extracted from the 180 videos, each with precise coordinate mapping.

: Teaching AI to distinguish between the "document" and the "table." midv-180

MIDV-180 is a cornerstone for training several types of machine learning models: : Thousands of individual frames extracted from the

The is a specialized collection of video data designed for the development and benchmarking of document analysis and recognition systems. Created by researchers at the Smart Engines company and the Institute for Information Transmission Problems (RAS), this dataset addresses a critical need in the field of mobile document scanning: the ability to recognize identity documents in diverse, real-world conditions. What is MIDV-180? What is MIDV-180

Researchers can often access this dataset through GitHub repositories maintained by Smart Engines (who also host the MIDV-500 and MIDV-2020 datasets) or through scientific data hosting platforms like or ResearchGate , where the accompanying papers describe the methodology in detail. Applications in AI and Computer Vision

The code refers to a specific entry in the Japanese Adult Video (JAV) industry. It is associated with the production company MOODYZ .

By providing a standardized "testing ground," MIDV-180 allows developers to compare their software's performance against others, driving innovation in the and digital identity verification industries.