StegPDF-21: A Feature-Engineered Dataset for PDF Steganography Detection

StegPDF-21 is a structured dataset designed for research in PDF steganography detection and document-level steganalysis. The dataset contains feature-based representations of clean and steganographic PDF documents for machine learning applications. Clean samples are labeled as 0, while steganographic samples containing intentionally embedded hidden data are labeled as 1.

To improve dataset traceability and reproducibility, each document is associated with a unique file_id, enabling consistent identification and mapping of samples throughout preprocessing, feature extraction, and evaluation stages.