Comprehensive model for predicting perceptual image quality of smart mobile devices

Area of Science:

• Computer Vision
• Human-Computer Interaction
• Image Processing

Background:

• Assessing image quality on smart mobile devices is crucial for user experience.
• Existing models may not fully capture the nuances of mobile display technologies and diverse usage scenarios.

Purpose of the Study:

• To develop a robust image quality model for smart mobile devices (smartphones and tablets).
• To establish mathematical relationships between visual image quality attributes and device/image factors.

Main Methods:

• Conducted psychophysical experiments evaluating attributes like naturalness, colorfulness, brightness, contrast, sharpness, and clearness.
• Utilized categorical judgment method under varying lighting conditions for diverse image types.
• Applied Pearson correlation coefficients and factor analysis for attribute importance.
• Developed multiple linear regression models to predict overall image quality.

Main Results:

• Overall image quality can be effectively predicted by two key constituent attributes.
• Mathematical expressions were derived linking visual attributes to physical device parameters and image appearance.
• The developed model demonstrated applicability across different devices, lighting, and image types.

Conclusions:

• The proposed image quality model provides a reliable method for assessing mobile image performance.
• The model's flexibility makes it suitable for diverse smart mobile devices and imaging applications.
• This work contributes to objective image quality assessment in mobile contexts.