MTD-360 is a meticulously curated dataset designed to evaluate and benchmark table detection models across diverse languages and document types. It comprises 360 images spanning 14 languages. The dataset encompasses a wide range of document formats, such as textbooks, dictionaries, bank reports, bills, and multi-column layouts, capturing varied table structures commonly found in real-world scenarios. Serving as both a test set and a benchmark.
© 2026 - Copyright AIKosh. All rights reserved.