A Selective Sparse Coding Model with Embedded Attention Mechanism

2011 ◽  
pp. 78-90
Author(s):  
Qingyong Li ◽  
Zhiping Shi ◽  
Zhongzhi Shi
Keyword(s):  
Visual Cortex ◽  
Primary Visual Cortex ◽  
Sparse Coding ◽  
Coding Theory ◽  
Neural Representation ◽  
Natural Scenes ◽  
Processing Capacity ◽  
Uniform Sampling ◽  
Coding Efficiency ◽  
Non Uniform Sampling

Sparse coding theory demonstrates that the neurons in the primary visual cortex form a sparse representation of natural scenes in the viewpoint of statistics, but a typical scene contains many different patterns (corresponding to neurons in cortex) competing for neural representation because of the limited processing capacity of the visual system. We propose an attention-guided sparse coding model. This model includes two modules: the non-uniform sampling module simulating the process of retina and a data-driven attention module based on the response saliency. Our experiment results show that the model notably decreases the number of coefficients which may be activated, and retains the main vision information at the same time. It provides a way to improve the coding efficiency for sparse coding model and to achieve good performance in both population sparseness and lifetime sparseness.

Neural representation of the bottom-up saliency map of natural scenes in human primary visual cortex

2013 ◽  
Vol 13 (9) ◽  
pp. 233-233
Author(s):  
C. Chen ◽  
X. Zhang ◽  
T. Zhou ◽  
Y. Wang ◽  
F. Fang
Keyword(s):  
Visual Cortex ◽  
Primary Visual Cortex ◽  
Saliency Map ◽  
Neural Representation ◽  
Natural Scenes ◽  
Bottom Up

Neural Representation of the Luminance and Brightness of a Uniform Surface in the Macaque Primary Visual Cortex

2001 ◽  
Vol 86 (5) ◽  
pp. 2559-2570 ◽  
Author(s):  
Masaharu Kinoshita ◽  
Hidehiko Komatsu
Keyword(s):  
Visual Cortex ◽  
Receptive Field ◽  
Primary Visual Cortex ◽  
Neural Activity ◽  
Dynamic Range ◽  
Neural Representation ◽  
Homogeneous Surface ◽  
V1 Neurons ◽  
Classical Receptive Field ◽  
Multiple Unit

The perceived brightness of a surface is determined not only by the luminance of the surface (local information), but also by the luminance of its surround (global information). To better understand the neural representation of surface brightness, we investigated the effects of local and global luminance on the activity of neurons in the primary visual cortex (V1) of awake macaque monkeys. Single- and multiple-unit recordings were made from V1 while the monkeys were performing a visual fixation task. The classical receptive field of each neuron was identified as a region responding to a spot stimulus. Neural responses were assessed using homogeneous surfaces at least three times as large as the receptive field as stimuli. We first examined the sensitivity of neurons to variation in local surface luminance, while the luminance of the surround was held constant. The activity of a large majority of surface-responsive neurons (106/115) varied monotonically with changes in surface luminance; in some the dynamic range was over 3 log units. This monotonic relation between surface luminance and neural activity was more evident later in the stimulus period than early on. The effect of the global luminance on neural activity was then assessed in 81 of the surface-responsive neurons by varying the luminance of the surround while holding the luminance of the surface constant. The activity of one group of neurons (25/81) was unaffected by the luminance of the surround; these neurons appear to encode the physical luminance of a surface covering the receptive field. The responses of the other neurons were affected by the luminance of the surround. The effects of the luminances of the surface and the surround on the activities of 26 of these neurons were in the same direction (either increased or decreased), while the effects on the remaining 25 neurons were in opposite directions. The activities of the latter group of neurons seemed to parallel the perceived brightness of the surface, whereas the former seemed to encode the level of illumination. There were differences across different types of neurons with regard to the layer distribution. These findings indicate that global luminance information significantly modulates the activity of surface-responsive V1 neurons and that not only physical luminance, but also perceived brightness, of a homogeneous surface is represented in V1.

What Is the Goal of Sensory Coding?

1994 ◽  
Vol 6 (4) ◽  
pp. 559-601 ◽  
Author(s):  
David J. Field
Keyword(s):  
Visual Cortex ◽  
Receptive Field ◽  
Sparse Coding ◽  
Wavelet Transforms ◽  
Response Probability ◽  
General Information ◽  
Natural Scenes ◽  
Fourth Moment ◽  
Sensory Coding ◽  
Response Distribution

A number of recent attempts have been made to describe early sensory coding in terms of a general information processing strategy. In this paper, two strategies are contrasted. Both strategies take advantage of the redundancy in the environment to produce more effective representations. The first is described as a “compact” coding scheme. A compact code performs a transform that allows the input to be represented with a reduced number of vectors (cells) with minimal RMS error. This approach has recently become popular in the neural network literature and is related to a process called Principal Components Analysis (PCA). A number of recent papers have suggested that the optimal “compact” code for representing natural scenes will have units with receptive field profiles much like those found in the retina and primary visual cortex. However, in this paper, it is proposed that compact coding schemes are insufficient to account for the receptive field properties of cells in the mammalian visual pathway. In contrast, it is proposed that the visual system is near to optimal in representing natural scenes only if optimality is defined in terms of “sparse distributed” coding. In a sparse distributed code, all cells in the code have an equal response probability across the class of images but have a low response probability for any single image. In such a code, the dimensionality is not reduced. Rather, the redundancy of the input is transformed into the redundancy of the firing pattern of cells. It is proposed that the signature for a sparse code is found in the fourth moment of the response distribution (i.e., the kurtosis). In measurements with 55 calibrated natural scenes, the kurtosis was found to peak when the bandwidths of the visual code matched those of cells in the mammalian visual cortex. Codes resembling “wavelet transforms” are proposed to be effective because the response histograms of such codes are sparse (i.e., show high kurtosis) when presented with natural scenes. It is proposed that the structure of the image that allows sparse coding is found in the phase spectrum of the image. It is suggested that natural scenes, to a first approximation, can be considered as a sum of self-similar local functions (the inverse of a wavelet). Possible reasons for why sensory systems would evolve toward sparse coding are presented.

scholarly journals Orientation tuning depends on spatial frequency in mouse visual cortex

2016 ◽  
Author(s):  
Inbal Ayzenshtat ◽  
Jesse Jackson ◽  
Rafael Yuste
Keyword(s):  
Visual Cortex ◽  
Spatial Frequency ◽  
Visual System ◽  
Primary Visual Cortex ◽  
Receptive Fields ◽  
Orientation Selectivity ◽  
Natural Scenes ◽  
Response Properties ◽  
Layer 2 ◽  
Mouse Visual Cortex

AbstractThe response properties of neurons to sensory stimuli have been used to identify their receptive fields and functionally map sensory systems. In primary visual cortex, most neurons are selective to a particular orientation and spatial frequency of the visual stimulus. Using two-photon calcium imaging of neuronal populations from the primary visual cortex of mice, we have characterized the response properties of neurons to various orientations and spatial frequencies. Surprisingly, we found that the orientation selectivity of neurons actually depends on the spatial frequency of the stimulus. This dependence can be easily explained if one assumed spatially asymmetric Gabor-type receptive fields. We propose that receptive fields of neurons in layer 2/3 of visual cortex are indeed spatially asymmetric, and that this asymmetry could be used effectively by the visual system to encode natural scenes.Significance StatementIn this manuscript we demonstrate that the orientation selectivity of neurons in primary visual cortex of mouse is highly dependent on the stimulus SF. This dependence is realized quantitatively in a decrease in the selectivity strength of cells in non-optimum SF, and more importantly, it is also evident qualitatively in a shift in the preferred orientation of cells in non-optimum SF. We show that a receptive-field model of a 2D asymmetric Gabor, rather than a symmetric one, can explain this surprising observation. Therefore, we propose that the receptive fields of neurons in layer 2/3 of mouse visual cortex are spatially asymmetric and this asymmetry could be used effectively by the visual system to encode natural scenes.Highlights–Orientation selectivity is dependent on spatial frequency.–Asymmetric Gabor model can explain this dependence.

scholarly journals Coding of Natural Scenes in Primary Visual Cortex

Neuron ◽  
2003 ◽  
Vol 37 (4) ◽  
pp. 703-718 ◽  
Author(s):  
Michael Weliky ◽  
József Fiser ◽  
Ruskin H Hunt ◽  
David N Wagner
Keyword(s):  
Visual Cortex ◽  
Primary Visual Cortex ◽  
Natural Scenes

Synchronization of Neuronal Responses in Primary Visual Cortex of Monkeys Viewing Natural Images

2008 ◽  
Vol 100 (3) ◽  
pp. 1523-1532 ◽  
Author(s):  
Pedro Maldonado ◽  
Cecilia Babul ◽  
Wolf Singer ◽  
Eugenio Rodriguez ◽  
Denise Berger ◽  
...  
Keyword(s):  
Visual Cortex ◽  
Eye Movements ◽  
Primary Visual Cortex ◽  
Discharge Rate ◽  
Spike Timing ◽  
Feature Binding ◽  
Natural Scenes ◽  
Event Analysis ◽  
Scene Segmentation ◽  
Discharge Rates

When inspecting visual scenes, primates perform on average four saccadic eye movements per second, which implies that scene segmentation, feature binding, and identification of image components is accomplished in <200 ms. Thus individual neurons can contribute only a small number of discharges for these complex computations, suggesting that information is encoded not only in the discharge rate but also in the timing of action potentials. While monkeys inspected natural scenes we registered, with multielectrodes from primary visual cortex, the discharges of simultaneously recorded neurons. Relating these signals to eye movements revealed that discharge rates peaked around 90 ms after fixation onset and then decreased to near baseline levels within 200 ms. Unitary event analysis revealed that preceding this increase in firing there was an episode of enhanced response synchronization during which discharges of spatially distributed cells coincided within 5-ms windows significantly more often than predicted by the discharge rates. This episode started 30 ms after fixation onset and ended by the time discharge rates had reached their maximum. When the animals scanned a blank screen a small change in firing rate, but no excess synchronization, was observed. The short latency of the stimulation-related synchronization phenomena suggests a fast-acting mechanism for the coordination of spike timing that may contribute to the basic operations of scene segmentation.

scholarly journals Role of Homeostasis in Learning Sparse Representations

2010 ◽  
Vol 22 (7) ◽  
pp. 1812-1836 ◽  
Author(s):  
Laurent U. Perrinet
Keyword(s):  
Sparse Coding ◽  
Hebbian Learning ◽  
Receptive Fields ◽  
Sparse Representations ◽  
Natural Images ◽  
Neural Representation ◽  
Natural Scenes ◽  
Efficient Coding ◽  
Sensory Data

Neurons in the input layer of primary visual cortex in primates develop edge-like receptive fields. One approach to understanding the emergence of this response is to state that neural activity has to efficiently represent sensory data with respect to the statistics of natural scenes. Furthermore, it is believed that such an efficient coding is achieved using a competition across neurons so as to generate a sparse representation, that is, where a relatively small number of neurons are simultaneously active. Indeed, different models of sparse coding, coupled with Hebbian learning and homeostasis, have been proposed that successfully match the observed emergent response. However, the specific role of homeostasis in learning such sparse representations is still largely unknown. By quantitatively assessing the efficiency of the neural representation during learning, we derive a cooperative homeostasis mechanism that optimally tunes the competition between neurons within the sparse coding algorithm. We apply this homeostasis while learning small patches taken from natural images and compare its efficiency with state-of-the-art algorithms. Results show that while different sparse coding algorithms give similar coding results, the homeostasis provides an optimal balance for the representation of natural images within the population of neurons. Competition in sparse coding is optimized when it is fair. By contributing to optimizing statistical competition across neurons, homeostasis is crucial in providing a more efficient solution to the emergence of independent components.

Sign in / Sign up

Export Citation Format

Share Document