Using a human visual system model to optimize soft-copy mammography display: influence of veiling glare

Area of Science:

• Medical Imaging
• Radiology
• Human Factors Engineering

Background:

• Digital displays are increasingly used for mammographic interpretation.
• Veiling glare, a display artifact, can potentially degrade image quality and affect diagnostic accuracy.
• Understanding the impact of veiling glare is crucial for optimizing soft-copy mammography viewing conditions.

Purpose of the Study:

• To evaluate the influence of digital display veiling glare on observer performance in mammographic mass detection.
• To compare human observer performance with a human visual system model (JNDmetrix) under varying veiling glare conditions.

Main Methods:

• 160 mammographic images (50% with masses) were processed to simulate four veiling glare levels.
• Six observers interpreted images on a low-glare LCD, with randomized presentation order.
• The JNDmetrix model was used to analyze image performance.

Main Results:

• Observer performance significantly declined with increasing veiling glare, particularly at quadruple the CRT level.
• The JNDmetrix model showed a modest correlation (0.588) with human observer data and did not fully capture performance reductions.
• Extreme veiling glare levels demonstrably impacted diagnostic accuracy in mammogram interpretation.

Conclusions:

• Soft-copy display veiling glare can negatively influence observer performance in mammography, particularly at high levels.
• The impact of veiling glare may be more pronounced for less experienced readers.
• Further research into display technologies and visual system modeling is warranted.