metadata

language:
  - th
  - en
tags:
  - OCR
  - dataset
  - evaluation
  - multilingual
  - handwritten
license: cc-by-sa-4.0

ThaiTROCR Evaluation Dataset

Dataset Description

The ThaiTROCR Evaluation Dataset is designed for evaluating Optical Character Recognition (OCR) models across various domains. It includes images and textual data derived from various open-source websites. This dataset aims to provide a comprehensive evaluation resource for researchers and developers working on OCR systems, particularly in Thai language processing.

Data Fields

Each sample in the dataset contains the following fields:

image: Path to the image file.
text: Ground truth text extracted from the image.
category: The domain/category of the image (e.g., "handwritten", "document", "scene_text").

Usage

To load the dataset, you can use the following code:

from datasets import load_dataset

dataset = load_dataset("suchut/thaitrocr-eval-dataset-beta")