What question did this study set out to answer?

To examine the statistical regularities and spatial scaling properties of natural woodland images to identify constraints on biological vision.

August 8, 1994Physical Review Letters

Statistics of natural images: Scaling in the woods

View Full Paper

Ask AI

Bookmark

Authors

DRDaniel RudermanUniversity of Southern California WBWilliam BialekThe Graduate Center, CUNY

Discussion

Member takes

Implication

Computational image analysis reveals scale-invariant contrast distributions in woodland environments, highlighting the necessity of gain-control mechanisms in visual processing.

Key Points

To examine the statistical regularities and spatial scaling properties of natural woodland images to identify constraints on biological vision.
Gathered and evaluated an ensemble of natural photographs captured within wooded environments.
Measured probability distributions of local contrast across multiple spatial scales and computed spatial power spectra.
Distributions of local image contrast displayed scale-invariant behavior characterized by nearly exponential tails.
Power spectra followed power-law scaling characterized by a nontrivial exponent across spatial frequencies.
Calculated statistical boundaries restrict the information content of natural scenes, underscoring the requirement for adaptive gain control in visual processing systems.

Cite This Study

Ruderman et al. (1994) studied this question.

synapsesocial.com/papers/69d8377a8c03fbaff8bee14a https://doi.org/10.1103/physrevlett.73.814

View Full Paper

Ask AI

Bookmark

Implication

Computational image analysis reveals scale-invariant contrast distributions in woodland environments, highlighting the necessity of gain-control mechanisms in visual processing.

Statistics of natural images: Scaling in the woods

Discussion

Member takes

Implication

Key Points

Cite This Study

Statistics of natural images: Scaling in the woods

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider