Subpixel-Based Image Scaling for Grid-like Subpixel Arrangements: A Generalized Continuous-Domain Analysis Model.

Area of Science:

• Image processing
• Computer vision
• Human visual system modeling

Background:

• Subpixel-based image scaling enhances apparent resolution by controlling individual subpixels.
• Improved luminance resolution often leads to chrominance distortion, necessitating color error suppression.
• Existing schemes face challenges in adapting to diverse subpixel arrangements and scaling factors.

Purpose of the Study:

• To propose a generalized continuous-domain analysis model for subpixel-based image scaling.
• To develop a scheme that minimizes color error while preserving image sharpness.
• To address limitations of current methods regarding subpixel arrangements and scaling factors.

Main Methods:

• Modeling the perceived signal by the human visual system (HVS) as a 2D continuous image.
• Developing a continuous-domain analysis for subpixel-based scaling (CASS) by minimizing perceived image differences.
• Introducing an improved CASS scheme incorporating Laplacian-of-Gaussian filtering to mitigate ringing artifacts.

Main Results:

• The proposed CASS methods produce sharp images with minimal color fringing artifacts.
• CASS demonstrates comparable performance to state-of-the-art methods on RGB stripe arrangements.
• The improved CASS scheme outperforms existing methods on non-RGB stripe subpixel arrangements.

Conclusions:

• Continuous-domain analysis provides an effective framework for subpixel-based image scaling.
• The proposed CASS methods successfully balance sharpness and color fidelity.
• The developed techniques offer improved performance and broader applicability for subpixel scaling.