Facial Expression Ambiguity and Face Image Quality Affect Differently on Expression Interpretation Bias

Area of Science:

• Cognitive Psychology
• Neuroscience
• Computer Vision

Background:

• Recognition of prototypical facial expressions (e.g., happiness, anger) is generally invariant across viewing conditions.
• Limited understanding exists regarding the invariance of ambiguous facial expression categorization and interpretation bias under degraded viewing settings.

Purpose of the Study:

• To investigate how facial expression ambiguity and image quality influence expression categorization, perceived intensity, and gaze patterns.
• To determine the extent to which interpretation bias for ambiguous expressions is affected by viewing conditions.

Main Methods:

• Utilized an eye-tracking study with systematically manipulated facial expression ambiguity (morphing happy/angry expressions) and image clarity (resolution).
• Measured participants' expression categorization accuracy, perceived expression intensity, and gaze distribution on faces.

Main Results:

• Increased ambiguity and decreased image quality led to opposing interpretation biases (negativity vs. positivity bias).
• Both factors similarly degraded the accuracy of perceived expression intensity ratings.
• Viewing conditions qualitatively altered gaze allocation, impacting eye and mouth focus.

Conclusions:

• The visual system exhibits reduced perceptual tolerance for ambiguous facial expressions compared to prototypical ones.
• Interpretation bias for ambiguous expressions is highly dependent on viewing conditions, unlike recognition of clear expressions.