On the inverse problem of binocular 3D motion perception

Area of Science:

• Visual neuroscience
• Computational vision
• 3D motion perception

Background:

• Current models for 3D motion perception, including interocular velocity difference and changing disparity, do not fully address the inverse optics problem of local binocular 3D motion.
• A general solution requires integrating multiple depth cues and motion information.

Purpose of the Study:

• To investigate a more flexible framework for solving the inverse optics problem of local binocular 3D motion.
• To compare two plausible default strategies for 3D motion processing: vector normal preferring slow 3D motion versus cyclopean average based on slow 2D motion.

Main Methods:

• Derivation of theoretical predictions based on the aperture problem and two default strategies.
• Analysis of local velocity constraints, binocular disparity, and other depth cues.
• Utilizing ambiguous line motion to distinguish between processing strategies.

Main Results:

• Existing processing schemes are insufficient for a general solution to local binocular 3D motion perception.
• A framework combining local velocity constraints and binocular disparity provides a more flexible approach.
• Theoretical results highlight the necessity of velocity constraints and feature-tracked disparity for solving the inverse problem.

Conclusions:

• Local velocity constraints and binocular disparity are crucial for solving the inverse problem in 3D motion perception.
• Motion and disparity information are likely processed in parallel and integrated late in the visual hierarchy.
• The proposed framework offers a more robust solution for understanding 3D motion perception.