An Underwater Color Image Quality Evaluation Metric

Area of Science:

• Computer Vision
• Image Processing
• Oceanography

Background:

• Underwater image quality assessment is crucial for retrieval and intelligent processing.
• Existing natural image quality metrics are unsuitable for underwater environments due to light absorption and scattering.
• No specific metric existed for underwater color image quality evaluation (UCIQE).

Purpose of the Study:

• To propose a novel metric for underwater color image quality evaluation (UCIQE).
• To address the challenges posed by water's optical properties on image quality.
• To provide a reliable tool for assessing underwater image enhancement.

Main Methods:

• Organized subjective testing to gather human perception of underwater image quality.
• Analyzed pixel distribution in CIELab color space to identify key quality factors.
• Developed UCIQE as a linear combination of chroma, saturation, and contrast.

Main Results:

• UCIQE correlates well with subjective image quality perception, specifically sharpness and colorfulness.
• The proposed UCIQE metric demonstrates comparable performance to leading natural and grayscale image quality metrics.
• UCIQE accurately predicts degradation in underwater images and shows high correlation with subjective mean opinion scores.

Conclusions:

• UCIQE is an effective and simple metric for evaluating underwater color image quality.
• The metric is suitable for real-time underwater video processing.
• UCIQE provides a valuable tool for underwater engineering, monitoring, and image enhancement assessment.