"It's Not a Tumor": A Framework for Capitalizing on Individual Diversity to Boost Target Detection

Area of Science:

• Cognitive psychology
• Human factors engineering
• Machine vision

Background:

• Experts frequently miss rare targets due to visual system limitations.
• Improving human performance in critical diagnostic tasks remains a challenge.

Purpose of the Study:

• To develop a novel framework for enhancing target detection by pairing individual estimates.
• To leverage the wisdom-of-crowds for improved visual search and diagnostic accuracy.

Main Methods:

• Applied a wisdom-of-crowds approach, averaging continuous estimates from two non-interacting individuals.
• Tested the framework across various tasks involving briefly presented images with prespecified targets.
• Investigated pairing strategies to maximize decorrelated performance patterns.

Main Results:

• Averaging estimates significantly boosted target detection rates across diverse tasks.
• Pairing individuals' estimates to maximize performance decorrelation optimized results.
• Demonstrated a practical method for enhancing observer performance in target detection.

Conclusions:

• Individual estimates can be effectively combined to overcome limitations in detecting infrequent targets.
• This framework offers a straightforward method to improve diagnostic accuracy and visual search.
• The findings have implications for fields requiring high-stakes visual inspection and detection.