This repository contains the output of the AuSR algorithms with USC-SIPI, Kodak-PCD0992, and UCID-1338 datasets.
The datasets can be accessed at ColorImageDatasets
The USC-SIPI dataset contains 8 color images with sizes of 512x512 pixels.
The Kodak-PCD0992 dataset contains 24 color images with sizes of 512x768 pixels.
The UCID-1338 dataset contains 1.338 color images with sizes of 512x384 pixels.
The results are then combined into two CSV files: watermarked-v3.csv and recovered-v3.csv
The statistical analysis consists of two processes:
- Parametric analysis (parametric or non-parametric).
- Statistical analysis of the PSNR and SSIM grouped by the AuSR algorithms.
The statistical analysis of the watermarked images can be accessed at watermarked-v3.ipynb
The statistical analysis of the recovered images can be accessed at recovered-v3.ipynb
Aminuddin, Afrig and Ernawan, Ferda "AuSR1: Authentication and self-recovery using a new image inpainting technique with LSB shifting in fragile image watermarking." Journal of King Saud University-Computer and Information Sciences 34, no. 8 (2022): 5822-5840. https://doi.org/10.1016/j.jksuci.2022.02.009.
Aminuddin, Afrig and Ernawan, Ferda "AuSR2: Image watermarking technique for authentication and self-recovery with image texture preservation." Computers and Electrical Engineering 102 (2022): 108207. https://doi.org/10.1016/j.compeleceng.2022.108207.