Visual re-identification of individual objects: a core problem for organisms and AI

Area of Science:

• Cognitive Science
• Robotics
• Perception

Background:

• Visual re-identification of objects over time presents two key challenges: determining necessary prior information and understanding the role of epistemic feelings.
• Epistemic feelings, like familiarity, influence object recognition and the assessment of associated opportunities and risks.

Purpose of the Study:

• To review open questions in visual object re-identification.
• To examine the interplay between human experiments and robotic systems in understanding object identity.
• To explore the contribution of epistemic feelings to object re-identification and affordance perception.

Main Methods:

• Review of existing research on object re-identification.
• Examination of experimental approaches using human subjects.
• Analysis of experiments utilizing robotic systems and virtual reality simulations.

Main Results:

• The study highlights the need for a priori information for robust object re-identification.
• It emphasizes the functional role of epistemic feelings in object recognition and risk assessment.
• Developmental robotics is proposed as a key methodology.

Conclusions:

• Developmental robotics, including virtual reality, may be crucial for understanding epistemic feelings and object identification.
• This approach can illuminate how humans implement and utilize familiarity for recognizing persistent objects.