Automated Calibration Method for Eye-Tracked Autostereoscopic Display.

Area of Science:

• Computer Vision
• Display Technology
• Human-Computer Interaction

Background:

• Autostereoscopic displays (ASDs) offer glasses-free 3D viewing but often suffer from calibration complexities and visual artifacts.
• Accurate calibration is crucial for optimal performance, yet traditional methods are time-consuming and require individual device tuning.
• Eye-tracking integration in ASDs (ETADs) necessitates precise calibration to adapt to individual viewer positions.

Purpose of the Study:

• To develop and validate an automated, simultaneous calibration system for eye-tracked autostereoscopic displays (ETADs).
• To enhance visual quality metrics such as crosstalk, peak-signal-to-noise ratio (PSNR), and structural similarity (SSIM).
• To reduce the overall calibration time and eliminate the need for manual intervention.

Main Methods:

• Established a physical model of the ETAD to identify and classify all calibration parameters.
• Developed a rendering method based on viewer eye position.
• Proposed a novel, two-image-based simultaneous calibration method automated with a custom calibration module.
• Utilized onboard and external cameras for image capture and analysis.

Main Results:

• Achieved a significant reduction in crosstalk to 8.32%.
• Improved visual quality, with PSNR increasing by 30.44% and SSIM by 40.14%.
• Completed the entire calibration process in under 1.5 seconds without external manipulation.

Conclusions:

• The proposed automated calibration system effectively enhances the visual performance of ETADs.
• Simultaneous calibration of all parameters significantly improves display quality and user experience.
• The rapid and automated nature of the system makes it practical for real-world applications.