ArticlePDF Available

Color coding and its interaction with spatiotemporal processing in the retina

Authors:
  • University of Tuebingen, and Max Planck Institute for Biological Cybernetics

Abstract

We use the theory of early visual processing proposed in ref. [1] to deduce the color encoding strategies of the retina. The calculated retinal transfer functions display a nontrivial coupling between color and spatiotemporal processing even when the autocorrelator of natural scenes has no coupling between the chro¬matic and the space-time dimensions. This coupling in the transfer function is fundamentally due to photoreceptor noise, and where red and green cone activities are highly correlated, as they are in humans and monkeys, it leads to the spatio¬temporal-chromatic opponent ganglion cells found in primates. Ignoring the blue cones, we find two types of ganglion cells whose receptive field organization is either red center with a green surround or green center with a red surround, as found by Derrington et al. [4] in monkeys. On the other hand, when the correlation between the red and green cone outputs is small, as is the case in shallow fresh water fish, we arrive at the "double opponency" cells observed in goldfish. We also argue that adding blue cones (which are rare) leads to a third type of cell with R + G — B opponency.
... Here we provide a simple and quantitative model that optimally incorporates redundancy in a neural population under a wide range of settings. In contrast to earlier studies [24][25][26][27]56,60], the proposed model allows for an arbitrary number of neurons in a population, providing previously unavailable insights and predictions: the degree to and the mechanisms by which the error can be minimized with different input-to-output cell ratios ( Figure 6); the conditions in which the redundancy reduction model is nearoptimal ( Figure 5); the degree of adaptation of receptive fields at different eccentricities to different light levels ( Figure 8). We observed that the optimal receptive fields are non-unique, as in other models [8,25,[59][60][61], and found that the additional constraint of spatial locality of the computation [25], but not previously examined constraints such as sparse weights [41] or sparse responses [7,8], yielded receptive fields similar to those found in the retina (Figure 7). ...
... In contrast to earlier studies [24][25][26][27]56,60], the proposed model allows for an arbitrary number of neurons in a population, providing previously unavailable insights and predictions: the degree to and the mechanisms by which the error can be minimized with different input-to-output cell ratios ( Figure 6); the conditions in which the redundancy reduction model is nearoptimal ( Figure 5); the degree of adaptation of receptive fields at different eccentricities to different light levels ( Figure 8). We observed that the optimal receptive fields are non-unique, as in other models [8,25,[59][60][61], and found that the additional constraint of spatial locality of the computation [25], but not previously examined constraints such as sparse weights [41] or sparse responses [7,8], yielded receptive fields similar to those found in the retina (Figure 7). ...
Article
Full-text available
A fundamental task of a sensory system is to infer information about the environment. It has long been suggested that an important goal of the first stage of this process is to encode the raw sensory signal efficiently by reducing its redundancy in the neural representation. Some redundancy, however, would be expected because it can provide robustness to noise inherent in the system. Encoding the raw sensory signal itself is also problematic, because it contains distortion and noise. The optimal solution would be constrained further by limited biological resources. Here, we analyze a simple theoretical model that incorporates these key aspects of sensory coding, and apply it to conditions in the retina. The model specifies the optimal way to incorporate redundancy in a population of noisy neurons, while also optimally compensating for sensory distortion and noise. Importantly, it allows an arbitrary input-to-output cell ratio between sensory units (photoreceptors) and encoding units (retinal ganglion cells), providing predictions of retinal codes at different eccentricities. Compared to earlier models based on redundancy reduction, the proposed model conveys more information about the original signal. Interestingly, redundancy reduction can be near-optimal when the number of encoding units is limited, such as in the peripheral retina. We show that there exist multiple, equally-optimal solutions whose receptive field structure and organization vary significantly. Among these, the one which maximizes the spatial locality of the computation, but not the sparsity of either synaptic weights or neural responses, is consistent with known basic properties of retinal receptive fields. The model further predicts that receptive field structure changes less with light adaptation at higher input-to-output cell ratios, such as in the periphery.
... These predictions approximate the structure of visual receptive fields, including surround inhibition. Moreover, receptive fields at different light levels can be predicted by considering the effect of the signal-to-noise ratio on the optimal balance between redundancy reduction and noise reduction (Atick et al., 1990;van Hateren, 1992). ...
Article
Visual scenes in the natural world are highly correlated. To efficiently encode such an environment with a limited dynamic range, the retina ought to reduce correlations to maximize information. On the other hand, some redundancy is needed to combat the effects of noise. Here we ask how the degree of redundancy in retinal output depends on the stimulus ensemble. We find that retinal output preserves correlations in a spatially correlated stimulus but adaptively reduces changes in spatio-temporal input correlations. The latter effect can be explained by stimulus-dependent changes in receptive fields. We also find evidence that horizontal cells in the outer retina enhance changes in output correlations. GABAergic amacrine cells in the inner retina also enhance differences in correlation, albeit to a lesser degree, while gylcinergic amacrine cells have little effect on output correlation. These results suggest that the early visual system is capable of adapting to stimulus correlations to balance the challenges of redundancy and noise.
... Jameson M me si ces mod les font une bonnecorrespondance entre les sensibilit s des r cepteurs et la sensation des couleurs per ues, ils n'expliquent pas la raison de ce codage. Plusieurs tudes se sont interress es cette question en justi ant une compression de l'information de couleur ( 18,7]). En e et les r cepteurs L et M sont tr s coupl s, leurs courbes de sensibilit se recouvrent beaucoup, ils portent quasiment la m me information et sont redondants. ...
Article
The knowledge of principles and mechanisms of color human perception makes possible to implement techniques of industrial reproduction and processing of color. In the case of the equivalence of colors the studies of human perception played a significant role and made it possible to establish standards by the CIE. But the appearance of colors under unspecified conditions of observation remains uncontrolled. In this study, we propose models of color perception based on the biology of the retina. These models concern the sampling of color by the cones of the retina, non-linearities of processing of the chromatic signal and the spatial and temporal processing of the chromatic signal by the layers of neurones in the retina. These models make it possible to understand the coding of color in opposition of colors, the formation of the thresholds of color discimination and the properties of spatial and temporal filtering of color signal within the retina. These elements make the retina an efficient and robust sensor for the analysis of natural images. The results of this study allow us to consider the realization of analog circuits copying the characteristics of processing of the retina because our models can be directly translated into electric circuits. They also allow the definition of an adaptive colorimetry, taking into account the phenomena of adaptation under the photopic conditions of vision. Finally they provide a method for the analysis of vector images, for which a vector of characteristics is defined at each pixel.
... Coefficients are used according to frequency and orientation decompositions. A factor of 2 is affected to achromatic channels compared to chromatic channels according to the color coding in the human visual system [1,4]. The same weighting factor is affected to high frequency information compared to average frequency one. ...
Article
Full-text available
This paper deals with image quality assessment that is capturing the focus of several research teams from academic and industrial parts. This field has an important role in various applications related to image from acquisition to projection. A large numbers of objective image quality metrics have been developed during the last decade. These metrics are more or less correlated to end-user feedback and can be sepa-rated in three categories: 1) Full Reference (FR) trying to evaluate the impairment in comparison to the reference image, 2) Reduced Reference (RR) using some features extracted from an image to represent it and compare it with the distorted one and 3) No Reference (NR) measures known as distortions such as blockiness, blurriness,. . . without the use of a reference. Unfortunately, the quality assessment community have not achieved a universal image quality model and only empirical models established on psychophysical experimentation are generally used. In this paper, we focus only on the third category to evaluate the quality of CRT (Cathode Ray Tube) and LCD (Liquid Crystal Display) color reproduction where a blind metric is, based on modeling a part of the human visual system behavior. The objective results are validated by single-media and cross-media subjective tests. This allows to study the ability of simulating displays on a reference one.
... off between spatial and chromatic resolutions [4]. The number of 3 for the cone color types is consistent with the results of Principal Component Analysis carried on the color spectra of natural scenes [30]. ...
Conference Paper
Full-text available
This paper presents a model of the retina with its properties with respect to sampling, spatiotemporal filtering, color-coding and non-linearity, and their consequences on the processing of visual information. It’s formalism points out the architectural and algorithmic principles of neuromorphic circuits which are known to improve compactness, consumption, robustness and efficiency, leading to direct applications in engineering science. It’s biological aspect, strongly based neural and cellular descriptions makes it suitable as an investigation tool for neurobiologists, allowing the simulation of experiences difficult to set up and answering fundamental theoretical questions.
Chapter
Most of us share the qualitative impression that the nervous system is a remarkable computational device. Part of the evidence for this belief is the tremendous difficulty in imitating neural functions using man-made hardware — we are a long way from having machines which see or hear. Yet it is easy to find tasks where even rather modest modern computers far outperform relatively sophisticated animals. What then is so special about neural computation?
Article
Starting with a biological model of the foveal and parafoveal regions of the retina, this paper shows that simple considerations about the spatiotemporal filtering processed by the retinal neural network can account for colour processing at the level of early vision. We establish that the Red, Green, Blue signal can be considered as a low-pass luminance signal plus a colour-modulated signal. The structure of the Outer- and Inner Plexiform Layers of the retina leads to spatial low- and high-pass filters which account for the achromatic and transient characteristics of the Y ganglion cells as well as for the spatiotemporal colour-opponent properties of X ganglion cells. Considering this property and the logarithmic transduction of photoreceptors, it is easy to postulate that, after the retina at the level of the Lateral Geniculate Nucleus, a simple low-pass filtering can pave the way to the well known colour constancy phenomenon.
Article
The sensory pathways of animals are well adapted to processing a special class of signals, namely stimuli from the animal's environment. An important fact about natural stimuli is that they are typically very redundant and hence the sampled representation of these signals formed by the array of sensory cells is inefficient. One could argue for some animals and pathways, as we do in this review, that efficiency of information representation in the nervous system has several evolutionary advantages. Consequently, one might expect that much of the processing in the early levels of these sensory pathways could be dedicated towards recoding incoming signals into a more efficient form. In this review, we explore the principle of efficiency of information representation as a design principle for sensory processing. We give a preliminary discussion on how this principle could be applied in general to predict neural processing and then discuss concretely some neural systems where it recently has been shown to be successful. In particular, we examine the fly's LMC coding strategy and the mammalian retinal coding in the spatial, temporal and chromatic domains.
Article
Full-text available
Robust coding has been proposed as a solution to the problem of minimizing decoding error in the presence of neural noise. Many real-world problems, however, have degradation in the input signal, not just in neural representations. This generalized problem is more relevant to biological sensory coding where internal noise arises from limited neural precision and external noise from distortion of sensory signal such as blurring and phototransduction noise. In this note, we show that the optimal linear encoder for this problem can be decomposed exactly into two serial processes that can be optimized separately. One is Wiener filtering, which optimally compensates for input degradation. The other is robust coding, which best uses the available representational capacity for signal transmission with a noisy population of linear neurons. We also present spectral analysis of the decomposition that characterizes how the reconstruction error is minimized under different input signal spectra, types and amounts of degradation, degrees of neural precision, and neural population sizes.
ResearchGate has not been able to resolve any references for this publication.