ChipQA: No-Reference Video Quality Prediction via Space-Time Chips

Area of Science:

• Computer Vision
• Signal Processing
• Multimedia Engineering

Background:

• No-reference video quality assessment (VQA) is crucial for evaluating user experience.
• Existing methods often require reference videos or complex motion computations.
• Developing efficient and accurate VQA models remains an active research area.

Purpose of the Study:

• To propose a novel no-reference video quality assessment model.
• To introduce Space-Time Chips (ST Chips) for localized video data analysis.
• To achieve state-of-the-art VQA performance without reference videos or explicit motion computation.

Main Methods:

• Utilized highly-localized space-time (ST) slices termed Space-Time Chips (ST Chips).
• Applied perceptually-motivated bandpass and normalization models for video processing.
• Selected oriented ST Chips based on their fit to parametric models of natural video statistics.

Main Results:

• Demonstrated that parameters describing natural video statistics reliably predict video quality.
• Showcased the model's ability to implicitly capture ST video naturalness and deviations.
• Achieved state-of-the-art performance on multiple large VQA databases.

Conclusions:

• The proposed ST Chip-based model offers effective no-reference VQA.
• The method provides high performance at reduced computational cost.
• This approach advances VQA by bypassing the need for reference videos and motion estimation.