What Makes for a Good Stereoscopic Image?

2024-12-30

AI Summary

Existing stereoscopic image quality evaluation (SQoE) methods primarily focus on local distortions and visual comfort, failing to capture holistic subjective experience; moreover, large-scale, VR-oriented annotated datasets are lacking. To address these limitations, we introduce SCOPE—the first VR-headset-collected stereoscopic content preference dataset—featuring both real and synthetic images with diverse perceptual distortions. We further propose iSQoE, an end-to-end preference learning model that uniquely integrates computational stereopsis modeling with deep neural networks, overcoming the constraints of single-dimensional distortion modeling. Experiments demonstrate that iSQoE achieves a 0.21 improvement in Spearman rank-order correlation coefficient (SROCC) over the best baseline on mono-to-stereo conversion quality ranking. Additionally, it attains cross-device preference annotation consistency of 0.89, significantly enhancing human perceptual alignment and generalizability of quality assessment.

Technology Category

• Computer Vision: 3D Computer Vision
• Knowledge Representation and Reasoning: Preferences
• Machine Learning: Learning Preferences or Rankings

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Security and Privacy: Large-scale security measurements

Abstract

With rapid advancements in virtual reality (VR) headsets, effectively measuring stereoscopic quality of experience (SQoE) has become essential for delivering immersive and comfortable 3D experiences. However, most existing stereo metrics focus on isolated aspects of the viewing experience such as visual discomfort or image quality, and have traditionally faced data limitations. To address these gaps, we present SCOPE (Stereoscopic COntent Preference Evaluation), a new dataset comprised of real and synthetic stereoscopic images featuring a wide range of common perceptual distortions and artifacts. The dataset is labeled with preference annotations collected on a VR headset, with our findings indicating a notable degree of consistency in user preferences across different headsets. Additionally, we present iSQoE, a new model for stereo quality of experience assessment trained on our dataset. We show that iSQoE aligns better with human preferences than existing methods when comparing mono-to-stereo conversion methods.

Problem

• Stereoscopic Image Quality Assessment
• Human Perception
• 2D to 3D Conversion

Innovation

• 3D Image Quality Assessment
• SCOPE Stereo Image Dataset
• iSQoE Model