Ablation and Benchmark Dataset

Description

This dataset consists of; Ablation Result 25 Files.xlsx and Ablation Result Example 1 File.xlsx. 1. Ablation Result 25 sample csv files This spreadsheet contains comprehensive ablation study and benchmarking results across 25 distinct digital image files. It details 275 evaluation records comparing various configuration scenarios, capturing execution time, memory usage, security metrics (Entropy, NPCR, UACI, Correlation, Chi-Square), and visual quality metrics (MSE, PSNR, SSIM, BER). 2. Ablation Result 1 sample csv file This spreadsheet provides a focused subset of the ablation study, detailing the evaluation metrics for a single sample image file. It contains 11 records demonstrating the isolated performance impact of enabling or disabling specific processing components and comparing them against baseline models.

Steps to reproduce

1. Prepare the Data: Select a targeted set of digital image files (e.g., 25 files for the comprehensive test, or a single file for the isolated test). 2. Define Test Scenarios: Establish multiple testing configurations by systematically isolating, enabling, or disabling specific processing components, and include alternative baseline algorithms for direct comparison. 3. Execute Processing: Process the selected image files through each defined configuration scenario. 4. Calculate Metrics: Measure and record resource overhead (Execution Time, Peak RAM), security parameters (Entropy, NPCR, UACI, Correlation, Chi-Square), and image fidelity (MSE, PSNR, SSIM, BER) for every scenario. 5. Compile Results: Aggregate the recorded metrics into structured spreadsheets to facilitate comparative performance analysis.

Institutions

• Universitas Ahmad Dahlan

  Yogyakarta, Yogyakarta

Categories

Licence