Self-Supervised Continuous Colormap Recovery from a 2D Scalar Field Visualization without a Legend.

Area of Science:

• Computer Vision
• Image Processing
• Color Science

Background:

• Reproducibility in scientific research is crucial.
• Comprehensive experimental details are often scattered or incomplete.
• Accessible code and data facilitate validation and further development.

Purpose of the Study:

• To provide complete experimental details and results complementing a main research paper.
• To offer insights into the motivation, methodology, and findings of the presented work.
• To make implementation code, datasets, and trained models publicly available.

Main Methods:

• Detailed description of the network architecture, including cubic B-spline interpolation.
• Explorative ablation studies using intermediate data and visualizations.
• Perceptual study design and stimuli explanation.

Main Results:

• Qualitative results presented in the LAB color space.
• Quantitative results analyzed with varying parameters and control points.
• Computation performance comparison, including training and testing times.

Conclusions:

• The supplementary material offers a thorough resource for understanding and replicating the research.
• Open access to datasets, code, and models promotes transparency and collaboration.
• The presented interactive visualization tool enhances user interaction and understanding.