To catch a thief with a recognition test: the model and some empirical results

Area of Science:

• Cognitive Psychology
• Human Memory Research
• Facial Recognition Studies

Background:

• Human memory and information processing are key to understanding facial recognition.
• Existing methods for identifying previously seen faces can be improved with novel techniques.

Purpose of the Study:

• Introduce a new technique and mathematical model, the "Catch model," for identifying target faces.
• Develop a method based on information processing principles of human memory.

Main Methods:

• Subjects identified the face most similar to a target face from pairs of test faces.
• Differentiating facial values (e.g., long nose, blue eyes) of chosen test faces were recorded.
• Target faces were reconstructed using the most frequent differentiating values across trials.

Main Results:

• The Catch model successfully reconstructs target faces by selecting high-frequency differentiating values.
• Mathematical derivations indicate the technique's effectiveness across variations.
• Three experiments validated the model's predictive power and accuracy.

Conclusions:

• The Catch model provides a viable method for facial identification based on memory principles.
• This approach demonstrates the utility of information processing models in cognitive tasks.
• The study validates the effectiveness of frequency-based feature selection for face reconstruction.