Yolo Dataset

Published: 12 October 2026| Version 1 | DOI: 10.17632/d3t8wccnkx.1
Contributor:

Description

This dataset provides the experimental evaluation of a multi-layer digital certification security framework integrating YOLO26-EasyOCR, 6D Hyper-chaotic DNA encryption, SPHINCS+ digital signatures, and Quantum-Resistant Ledger (QRL) anchoring. It consists of 3,000 evaluated digital certificate samples (1,500 authentic and 1,500 adversarially attacked) across multiple formats and resolutions. The dataset captures key evaluation metrics including AI vision extraction accuracy (CER), computational latency, cryptographic randomness (Entropy, NPCR, UACI), and steganographic visual fidelity (PSNR, SSIM, MSE).

Files

Steps to reproduce

To reproduce the experimental findings presented in the study, follow these steps: 1. Data Loading: Import both Original Testing Result.csv and Attack Testing Result.csv using data analysis tools like Python (pandas) or R. 2. AI Vision & Scalability Analysis: Calculate the mean of CER Score for extraction accuracy. To observe the O(M x N) linear time complexity, group Exec Time Sec and Peak RAM MB by Resolution. 3. Cryptographic Randomness Evaluation: Extract the descriptive statistics (mean) of Entropy, NPCR, UACI, and Chi Square from the authentic dataset to validate resistance against statistical and differential attacks. 4. Adversarial Robustness Assessment: Compare the visual quality metrics (PSNR, SSIM, MSE) and Verification Result (Authentic vs. Counterfeit) between the original and attacked datasets to evaluate the SPHINCS+ zero-tolerance tamper detection.

Institutions

Categories

Image Database, Meta Dataset

Licence