Natural image statistics and neural representation.

By Eero P. Simoncelli, Barbara Olshausen. 2001

2,572 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Cognitive Neuroscience, Representation & Self-Supervised Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Cognitive Neuroscience

In Representation & Self-Supervised Learning

Citations in the corpus

Group Normalization (2018, background): “In fact, a well-accepted computational model in neuroscience is to normalize across the cell responses [21, 52, 55, 5], “with various receptive-field centers (covering the visual field) and with various spatiotemporal frequency tunings” (p183, [21]); this can happen not only in the primary visual co”
A Learned Representation For Artistic Style (2016, background): “Furthermore, the degree to which an artistic styling model might generalize across painting styles would directly measure our ability to build systems that parsimoniously capture the higher level features and statistics of photographs and images (Simoncelli & Olshausen, 2001).”
Hollywood in Homes: Crowdsourcing Data Collection for Activity Understanding (2016, background): “The natural world generally follows a long-tailed distribution [30,31], but we can see that the distribution of words in the dataset is relatively even.”
Geometric Deep Learning: Going beyond Euclidean data (2016, background): “the data such as stationarity and compositionality through local statistics, which are present in natural images, video, and speech [14], [15].”
Evaluating the Visualization of What a Deep Neural Network Has Learned (2015, background): “In this paper we refrain from mimicking the complex human heatmap evaluation process which includes attention, interest point models and perception models of saliency [36], [37], [38], [39] for the reason that we are interested in finding those regions that are relevant for a given classifier.”
How does the brain solve visual object recognition? (2012, neutral): “As V1 takes up the task, the number of output neurons, and hence the total dimensionality of the V1 representation, increases approximately 30-fold (Stevens, 2001); Figure 3B).”
Mixtures of Conditional Gaussian Scale Mixtures Applied to Multiscale Image Representations (2011, neutral): “In computational neuroscience, they are used as a means to understand the structure of the input to which biological vision systems have adapted and as a basis for normative theories of how those inputs are optimally processed [1,2].”
A Kronecker-factored approximate Fisher matrix for convolution layers (2016, background): “Images have a well-characterized 1/fp power spectrum with p ≈ 2 (Simoncelli & Olshausen, 2001), which implies that the curvature may be much larger in directions corresponding to low-frequency Fourier components than in directions corresponding to high-frequency components.”
ConViT: improving vision transformers with soft convolutional inductive biases (2021, neutral): “models more sample-eﬃcient and parameter-eﬃcient (Simoncelli and Olshausen, 2001, Ruderman and Bialek 1994).”
Improving Robustness Without Sacrificing Accuracy with Patch Gaussian Augmentation (2019, background): “Understanding the impact of data distributions and noise on representations has been well-studied in neuroscience [3, 42, 30].”
Image information and visual quality (2006, methodology): “Researchers believe that the visual stimulus emanating from the natural environment drove the evolution of the HVS, and that modeling natural scenes and the HVS are essentially dual problems [20].”