Humanoid Robot Facial Expression Recognition Study Data

A user study with 40 participants evaluated emotion recognition across human faces, virtual avatars, and humanoid robots. The dataset includes recognition rates for emotions like surprise and fear, assessed via an eight-item questionnaire on a 7-point Likert scale. It was created by Patricia Schwarz to support research on cross-domain facial expression mapping for patient simulation.

Use Cases

• Analyze emotion recognition rates for surprise (H: 87.5%; A: 57.5%; R: 65%) and fear across human, avatar, and robot stimulus categories.
• Model the impact of dynamic expressions combined with speech on recognition rates using Likert scale questionnaire data.
• Assess perceived realism and uncanny valley effects from the eight-item questionnaire rated on a 7-point Likert scale.
• Compare recognition performance for subtle emotions like anger and disgust between humanoid robots and virtual avatars.

Strengths

• Data from a structured user study with 40 participants, providing a controlled experimental basis.
• Includes specific recognition percentages for key emotions like surprise, fear, anger, and disgust across three stimulus types.
• Emotion assessment uses an eight-item questionnaire with a standardized 7-point Likert scale for consistency.

Limitations

• The dataset is small in scale at 413.3 KB, derived from a single study with 40 participants, limiting statistical power.
• Findings are constrained by the specific mechanical expressivity of the humanoid robot used, which may not generalize.
• Data is presented in a static PDF format, requiring extraction for computational analysis.

Provenance

Source

Patricia Schwarz via figshare

Collection Method

Data collected from a controlled user study evaluating emotion recognition from static images and dynamic sequences.

Time Range

null

Freshness

null

Geography

null