Perceptual Depth Quality in Distorted Stereoscopic Images.

Area of Science:

• Stereoscopic 3D imaging
• Visual perception
• Image quality assessment

Background:

• Assessing depth quality in stereoscopic images is crucial but challenging.
• Traditional methods are confounded by image content and spatial quality.
• A novel approach is needed to isolate the impact of stereoscopic cues on depth perception.

Purpose of the Study:

• To develop a reliable method for measuring depth perception in stereoscopic images.
• To investigate the effects of different distortions on depth perception.
• To create a computational model for predicting depth perception difficulty.

Main Methods:

• A novel subjective study was conducted, synthesizing depth effects before applying distortions.
• Participants identified depth polarizations, leading to the development of a depth perception difficulty index (DPDI).
• A computational model was proposed to predict DPDI without explicit distortion identification.

Main Results:

• The new subjective method effectively quantifies depth perception from stereo cues.
• Significant effects of image content, distortion type, and symmetry were observed.
• The proposed computational model achieved promising DPDI prediction performance.

Conclusions:

• The developed DPDI offers a more accurate measure of depth perception in stereoscopic images.
• The findings provide insights into factors affecting 3D image quality.
• The computational model advances the prediction of 3D quality-of-experience.