|
Download README.md from t-Shr/Detect-AD: direct link, hf CLI and curl.
- Browser
- Download file 1.6 kB
-
https://huggingface.co/datasets/t-Shr/Detect-AD/resolve/main/README.md
- Command line
-
hf download hf://datasets/t-Shr/Detect-AD/README.md
-
curl -L -o README.md https://huggingface.co/datasets/t-Shr/Detect-AD/resolve/main/README.md
1.6 kB
metadata
license: other
task_categories:
- object-detection
- image-classification
language:
- en
- hi
tags:
- newspaper
- advertisement-detection
- yolo
- object-detection
- indian-newspapers
- computer-vision
- OCR
- document-layout
size_categories:
- 1K<n<10K
Detech-AD: Indian Newspaper Advertisement Detection Dataset
Overview
Detech-AD is a dataset created for research and development of an automated advertisement detection pipeline for Indian e-newspapers.
The dataset contains newspaper PDFs, rendered newspaper pages, and YOLO-format annotations for advertisement detection.
The primary objective is to detect advertisements appearing in Indian newspapers and identify their locations on newspaper pages.
The dataset is intended to support computer vision, document analysis, advertisement detection, and OCR-based newspaper analysis.
Newspapers
The dataset currently contains data from the following Indian newspapers:
- The Times of India (TOI)
- The Hindu
- Dainik Jagran
Additional newspapers may be added in future versions.
Dataset Structure
The repository is organized as follows:
data/
│
├── raw/
│ ├── toi/
│ ├── the_hindu/
│ └── dainik_jagran/
│
├── pages/
│ ├── toi/
│ ├── the_hindu/
│ └── dainik_jagran/
│
└── yolo_dataset/
├── images/
│ ├── train/
│ ├── val/
│ └── test/
│
└── labels/
├── train/
├── val/
└── test/