Johannes Burge

The paper investigates motion illusions caused by monovision in presbyopic and non-presbyopic populations, finding that interocular blur differences significantly affect processing speed and create visible illusions in subsets of participants.

The paper investigates how natural stimulus variability affects human stereo-depth discrimination performance using psychophysical experiments and analytical methods.

The paper measures temporal dynamics of human color processing using a tracking task and finds similar elliptical response contours for detection and tracking, but reduced sensitivity to S-cone signals in tracking.

The paper shows that feature-specific divisive normalization improves the encoding of binocular disparity in natural images, enhancing depth perception accuracy.

This paper develops a continuous target-tracking psychophysics paradigm to measure millisecond-scale visual processing delays and shows it aligns with traditional methods.

The paper develops ideal observer models to estimate 3D motion from natural binocular stimuli, showing optimal computations that match human performance patterns.

This study shows that visual processing delays decrease with increasing eccentricity, affecting motion and 3D perception.

The paper investigates how overall light-level affects the Classic and Reverse Pulfrich effects, finding both increase logarithmically with decreasing light-level.

The paper measures how background reflectance variation affects lightness discrimination thresholds and finds that external noise has a modest effect compared to internal noise.

The paper investigates the anomalous Pulfrich effect, showing that interocular differences in temporal integration duration, not delay, cause misperceptions of motion depth.

The paper demonstrates that contact lenses can induce the reverse Pulfrich effect and that anti-Pulfrich corrections can eliminate it, replicating findings from trial lenses.

The paper shows humans use natural scene statistics to pool spatial information for estimating surface tilt, outperforming local models.

The paper investigates how natural image statistics at depth edges influence perceptual stability by leveraging statistical regularities for binocular fusion.

The paper develops a Bayesian ideal observer model to predict human speed perception from natural stimuli and finds that performance is limited by stimulus variability and internal noise.

The paper models neuron responses to natural images by analyzing how receptive fields, normalization, and encoding noise affect response statistics and signal-to-noise ratios.

The paper investigates how differential blur in monovision corrections affects motion perception, revealing a reverse Pulfrich effect that challenges existing understanding of visual processing.

The paper investigates how scene variability affects the estimation of object luminance reflectance using supervised statistical learning on simulated cone photoreceptor responses from naturalistic images.

This study investigates how depth variation in natural scenes affects performance in disparity-related perception tasks using computational analysis of stereo-images.

The paper investigates human tilt estimation in natural scenes and finds that estimates are imprecise and biased compared to artificial stimuli, with a normative model explaining these patterns.

The paper introduces AMA-Gauss, a faster Bayesian method for finding task-specific filters, and shows its validity for visual tasks, linking it to quadratic neural response models.