Realistic Modeling of Simple and Complex Cell Tuning in the HMAX Model, and Implications for Invariant Object Recognition in Cortex

By Thomas Serre, Maximilian Riesenhuber. 2004

150 citations recorded by the source corpus.

Abstract : Riesenhuber & Poggio recently proposed a model of object recognition in cortex which, beyond integrating general beliefs about the visual system in a quantitative framework, made testable predictions about visual processing. In particular, they showed that invariant object representation could be obtained with a selective pooling mechanism over properly chosen afferents through a MAX operation: For instance, at the complex cells level, pooling over a group of simple cells at the same preferred orientation and position in space but at slightly different spatial frequency would provide scale tolerance, while pooling over a group of simple cells at the same preferred orientation and spatial frequency but at slightly different position in space would provide position tolerance. Indirect support for such mechanisms in the visual system comes from the ability of the architecture at the top level to replicate shape tuning as well as shift and size invariance properties of "view-tuned cells" (VTUs) found in inferotemporal cortex (IT), the highest area in the ventral visual stream, thought to be crucial in mediating object recognition in cortex. There is also now good physiological evidence that a MAX operation is performed at various levels along the ventral stream. However, in the original paper by Riesenhuber & Poggio, tuning and pooling parameters of model units in early and intermediate areas were only qualitatively inspired by physiological data. Many studies have investigated the tuning properties of simple and complex cells in primary visual cortex, V1. We show that units in the early levels of HMAX can be tuned to produce realistic simple and complex cell-like tuning, and that the earlier findings on the invariance properties of model VTUs still hold in this more realistic version of the model.

Subfields: Cognitive Neuroscience, Computer Vision and Pattern Recognition

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

In Computer Vision and Pattern Recognition

Citations in the corpus

Scattering Networks for Hybrid Representation Learning (2018, methodology): “In these approaches local neighborhoods of an image were encoded using method such as SIFT descriptors [33], HOG [15], and wavelet transforms [47].”
Scaling the Scattering Transform: Deep Hybrid Networks (2017, neutral): “In these approaches local neighborhoods of an image were encoded using method such as SIFT descriptors [23], HOG [9], and wavelet transforms [32].”
Robust Object Recognition with Cortex-Like Mechanisms (2007, methodology): “Again, parameters (see Table 1) governing this pooling operation were adjusted such that the tuning of the C 1 units match the tuning of complex cells as measured experimentally (see [41] for details).”
Physiologically Inspired Model for the Visual Recognition of Transitive Hand Actions (2013, unknown): “Theparame-tersofthemodelneuronswere,asfaraspossible,constrainedbyphysi-ologicalparameters,partiallytakingoverresultsfromrelatedmodelsin theliterature(SerreandRiesenhuber,2004;Serreetal.,2007b).”
A Theory of Object Recognition: Computations and Circuits in the Feedforward Path of the Ventral Stream in Primate Visual Cortex (2005, methodology): “S1 and C1 layers: In [Serre and Riesenhuber, 2004] we found that the S1 and C1 units in the original model were too broadly tuned to orientation and spatial frequency and revised these units accordingly.”
A quantitative theory of immediate visual recognition. (2007, background): “For instance , at the level of V1, model units agree with the tuning properties of cortical cells in terms of frequency and orientation bandwidth, as well as peak frequency selectivity and receptive field sizes (see Serre and Riesenhuber, 2004).”
Object recognition with features inspired by visual cortex (2005, methodology): “Again pooling parameters were set so that C1 units match the tuning properties of complex cells as measured experimentally (see Table 1 and [19] for a full description of how those filters were obtained).”