Stereo vision requires an explicit encoding of vertical disparity

Area of Science:

• Neuroscience
• Computational Vision
• Human Perception

Background:

• Vertical disparities are crucial for 3D depth perception.
• Neuronal mechanisms for vertical disparity processing remain largely unknown.
• A proposed 1D model suggested vertical disparity information could be derived from horizontal disparity sensors.

Purpose of the Study:

• To investigate whether the human visual system uses a 1D or explicit 2D encoding for vertical disparity.
• To test the validity of the 1D model proposed by Read and Cumming (2006).

Main Methods:

• Utilized a modified induced effect stimulus simulating viewing geometry at infinity.
• Compared induced effects under infinite viewing conditions to those on a frontoparallel screen at finite distance.
• Analyzed psychophysical responses to vertical disparity stimuli.

Main Results:

• Induced effects were indistinguishable between the standard and modified stimulus conditions.
• The 1D model's predictions were not supported by the experimental data.
• The results indicate that the visual system does not rely solely on horizontal disparity cues.

Conclusions:

• The 1D model for vertical disparity encoding is refuted.
• Vertical disparity, including its sign, must be explicitly encoded across the visual field.
• This explicit encoding is essential for accurate 3D depth perception.