A quantitative theory of immediate visual recognition.

By Thomas Serre, Gabriel Kreiman, Minjoon Kouh, Charles F. Cadieu, Ulf Knoblich, Tomaso A. Poggio. 2007

301 citations recorded by the source corpus.

Human and non-human primates excel at visual recognition tasks. The primate visual system exhibits a strong degree of selectivity while at the same time being robust to changes in the input image. We have developed a quantitative theory to account for the computations performed by the feedforward path in the ventral stream of the primate visual cortex. Here we review recent predictions by a model instantiating the theory about physiological observations in higher visual areas. We also show that the model can perform recognition tasks on datasets of complex natural images at a level comparable to psychophysical measurements on human observers during rapid categorization tasks. In sum, the evidence suggests that the theory may provide a framework to explain the first 100-150 ms of visual object recognition. The model also constitutes a vivid example of how computational models can interact with experimental observations in order to advance our understanding of a complex phenomenon. We conclude by suggesting a number of open questions, predictions, and specific experiments for visual physiology and psychophysics.

Subfields: Cognitive Neuroscience

How researchers see this paper

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

In Cognitive Neuroscience

Citations in the corpus

DeepSecure: Scalable Provably-Secure Deep Learning (2017, background): “Recent theoretical and empirical advances in DL has demonstrated the importance of sparsity in training DL models [26–28].”
Shallow vs. Deep Sum-Product Networks (2011, background): “Analogies wit h the structure of the cerebral cortex (in particular the visual cortex) [31] and similarities betwee n f atures learned with some Deep Learning algorithms and those hypothesized in the visual cortex [17] further motivate investigations into deep architectures.”
Local Deep Neural Networks for gender recognition (2016, background): “The brain cortex extracts multiple levels of representation from the sensory input, doing progressively more complex processing tasks (Serre et al. (2007)).”
Hierarchical learning : theory with applications in speech and vision (2009, background): “These models quantitatively account for a host of psychophysical and physiological data, and provide human-level performance on tasks involving rapid categorization of complex imagery [95, 98, 96].”
Gearbox Fault Identification and Classification with Convolutional Neural Networks (2015, unknown): “Modern understanding of the physiology of the visual system is consistent with the processing style found in convolutional networks in the literature [40].”
Latent Hierarchical Model of Temporal Structure for Complex Activity Classification (2014, background): “Hierarchical tree-structured model is biologically inspired by the brain architecture and vision system [36], [37].”
On the Complexity of Neural Network Classifiers: A Comparison Between Shallow and Deep Architectures (2014, unknown): “For example, in neuroscience, it is well known that the neocortex of the mammalian brain is organized in a complex layered structure of neurons [7], [8].”
Recurrent Processing during Object Recognition (2013, background): “…neural response properties across these different brain areas (Fukushima, 1980, 2003; Wal-lis O’Reilly et al. feedforward-driven neural responses theventralpathwayare sufﬁcient tosolvereasonablychallenging versions of the object recognition problem (Serre et al., 2007a,b; DiCarlo et al., 2012).”
Learning Deep Architectures for AI (2007, background): “This is particularly clear in the primate visual system [173], with its sequence of processing stages: detection of edges, primitive shapes, and moving up to gradually more complex visual shapes.”
Deep Sparse Rectifier Neural Networks (2011, background): “This is particularly clear in the primate visual system (Serre et al., 2007), with its sequence of processing stages: detection of edges, primitive shapes, and moving up to gradually more complex visual shapes.”