Abstract
Recently, we proposed a unified model for binocular fusion and depth perception and tested it with frontoparallel stereograms (Ding & Levi, Vision Research 2021). At each location, the model consists of an array of disparity detectors, each with different preferred position and phase disparities. The phase-disparity detectors compute interocular misalignment and provide phase-disparity energy (binocular fusion energy) to shift the readout of disparity detectors along position disparity space until the misalignment is eliminated; sensory fusion is achieved locally. After sensory fusion, the combination of position and possible residual phase disparity energies is calculated for depth perception. Binocular fusion occurs at multiple scales following a coarse-to-fine process. At a given location, the apparent depth is the weighted sum of sensory readout shifts combined with residual phase disparity in all spatial-frequency channels, and the weights depend on stimulus spatial frequency and contrast. To test the model with more complex depth profiles, we performed experiments using dynamic band-pass noise stereograms (dBNS, central spatial-frequency = 0.75, 1.5, 3 or 6 cpd and one octave bandwidth) with disparity corrugations of 0.094, 0.188, 0.375, 0.75, or 1.5 cpd. Our results, consistent with previous studies using broadband dRDS (Kane et al. 2014, Peterzell et al. 2017), show that stereovision has poor spatial resolution and limited disparity variance; the Dmin corrugation amplitude has a bandpass property and Dmax has a lowpass property. Using narrow-band dBNS, we also revealed that the Dmin amplitude depends on corrugation/carrier spatial-frequency ratio; best performance occurs when corrugation frequency is 4-8 times lower than the carrier frequency. Model simulations show that the unified model without late-stage filters has a lowpass property for depth corrugation perception. After including late-stage bandpass filters, the model predicts band-pass properties for Dmin corrugation amplitude and low-pass properties for Dmax amplitude, consistent with our experimental data.