Natural Scene Categorization Research Articles

Perceptual Categorization Is Impaired in Huntington’s Disease: An Electrophysiological Study

Recent evidence raised the possibility that the neostriatum and the corticostriatal circuits could play an important role in semantic categorization. In this study, we examined the electrophysiological correlates of natural scene categorization in Huntington’s disease (HD) patients and their asymptomatic relatives who were Huntington’s disease mutation carriers (HDC). Event-related potentials were recorded in HD patients, HDC subjects, and age-matched control subjects using a natural scene categorization task. The subjects had to decide whether a briefly presented image contained animals or no animals. Concerning the N1 component (150–250 ms), the mean amplitudes were more negative for nonanimal scenes as compared with stimuli containing animals at all electrode sites in the control group and at all but the lateral temporal electrode sites (T3, T4) in the HD group. Between-group comparison demonstrated that the N1 amplitudes were significantly smaller for both kinds of stimuli in the HD group in spite of a normal primary occipital component (P100). The HDC subjects were not significantly different from the controls concerning the N1 amplitudes. These results suggest that perceptual (N1) processes related to the categorization of natural scenes are specifically impaired in HD. The findings are in agreement with the hypothesis emphasizing the importance of neostriatal mechanisms in human categorization functions.

An improved algorithm for learning long-term dependency problems in adaptive processing of data structures

Many researchers have explored the use of neural-network representations for the adaptive processing of data structures. One of the most popular learning formulations of data structure processing is backpropagation through structure (BPTS). The BPTS algorithm has been successful applied to a number of learning tasks that involve structural patterns such as logo and natural scene classification. The main limitations of the BPTS algorithm are attributed to slow convergence speed and the long-term dependency problem for the adaptive processing of data structures. In this paper, an improved algorithm is proposed to solve these problems. The idea of this algorithm is to optimize the free learning parameters of the neural network in the node representation by using least-squares-based optimization methods in a layer-by-layer fashion. Not only can fast convergence speed be achieved, but the long-term dependency problem can also be overcome since the vanishing of gradient information is avoided when our approach is applied to very deep tree structures.

Rapid natural scene categorization in the near absence of attention.

What can we see when we do not pay attention? It is well known that we can be "blind" even to major aspects of natural scenes when we attend elsewhere. The only tasks that do not need attention appear to be carried out in the early stages of the visual system. Contrary to this common belief, we report that subjects can rapidly detect animals or vehicles in briefly presented novel natural scenes while simultaneously performing another attentionally demanding task. By comparison, they are unable to discriminate large T's from L's, or bisected two-color disks from their mirror images under the same conditions. We conclude that some visual tasks associated with "high-level" cortical areas may proceed in the near absence of attention.

Electrophysiological correlates of visual categorization: evidence for cognitive dysfunctions in early Parkinson’s disease

Our aim was to examine the electrophysiological correlates of perceptual categorization in Parkinson’s disease (PD). We recorded visual event-related potentials (ERPs) in a natural scene categorization task in drug naive idiopathic PD patients and healthy control subjects. In the control group, there was a significant early difference (150–250 ms poststimulus) between ERPs elicited by pictures containing animals and scenes without animals. In spite of relatively preserved basic-level visual functions, this was not observable in the PD group. These results raise the possibility for striatal contributions to visual categorization and may provide a novel protocol for further clinical studies.

Corticostriatal circuitry mediates fast-track visual categorization

Previous studies have shown that briefly presented natural scenes containing non-animals elicited more negative potentials than images with animals even at 150 ms after stimulus onset (dN150). Cognitive models suggest that both feed-forward and feature weighting processes are involved in the rapid categorization of complex natural scenes. Here we examined the possible neuronal substrates of this model. Patients with Alzheimer’s disease (AD) exhibited a delayed dN150, but in their case non-animals evoked more negative potentials similarly to the controls (presence of dN150). In contrast, in patients with Parkinson’s disease (PD) animal and non-animal stimuli elicited nearly identical early responses (absence of dN150). The results indicate that when cortico–cortical pathways mediating feed-forward mechanisms are impaired (as in the case of AD), dN150 appears later, while in the case of corticostriatal dysfunctions (as in the case of PD) no differential response is present. This supports the hypothesis that corticostriatal circuits mediate perceptual feature weighting and integration in complex situations requiring categorical judgements.

Event-related potentials from a visual categorization task

Investigation of the neural bases of human perceptual categorization is a continuously growing territory of visual sciences. Recently, several papers appeared on the cortical bases of human categorization performance, using functional brain imaging techniques such as positron emission tomography (PET) or functional magnetic resonance imaging (fMRI). However, the cheaper commercially available laboratory tests are not always the best methods to study multiple aspects of visual categorization. In this study, we describe the electrophysiological correlates of natural scene categorization in humans. The subject’s task was to decide whether briefly presented natural scenes contained animal or non-animal items. Analysis of event-related potentials (ERPs) recorded during the categorization task, revealed more negative potentials for non-animal stimuli (NAS) in the time windows of 150–250 ms (N1) and 350–500 ms (N2), and more positive potentials for animal stimuli (AS) in the time window of 250–350 ms (P2). Our work provides an inexpensive noninvasive method to study both early perceptual and late phases of semantic information processing by recording ERPs during the categorization of natural scenes.

A limit to the speed of processing in ultra-rapid visual categorization of novel natural scenes.

The processing required to decide whether a briefly flashed natural scene contains an animal can be achieved in 150 msec (Thorpe, Fize, & Marlot, 1996). Here we report that extensive training with a subset of photographs over a 3-week period failed to increase the speed of the processing underlying such Rapid Visual Categorizations: Completely novel scenes could be categorized just as fast as highly familiar ones. Such data imply that the visual system processes new stimuli at a speed and with a number of stages that cannot be compressed. This rapid processing mode was seen with a wide range of visual complex images, challenging the idea that short reaction times can only be seen with simple visual stimuli and implying that highly automatic feed-forward mechanisms underlie a far greater proportion of the sophisticated image analysis needed for everyday vision than is generally assumed.

Early and late components of visual categorization: an event-related potential study

We examined the characteristics of early and late components of event-related potentials (ERPs) accompanying the visual categorization of natural scenes. In the first experiment, ERPs were recorded in an animal–non-animal categorization task (CT), whereas the second experiment included a spatial frequency discrimination task (DT). In the CT, there were more negative potentials for non-animals in the time windows of 150–250 ms (N1) and 350–500 ms (N2), and a more positive potential for animals in the time window of 250–350 ms (P2). In the DT, spatial frequency gratings evoked only a short-duration difference N1 at the frontal sites. Our results suggest that N1 may be related to higher-level categorization processes.

On the classification of classes with nearly equal spectral response in remote sensing hyperspectral image data

It is well known that high-dimensional image data allows for the separation of classes that are spectrally very similar, i.e., possess nearly equal first-order statistics, provided that their second-order statistics differ significantly. The aim of this study is to contribute to a better understanding, from a more geometrically oriented point of view, of the role played by the second-order statistics in remote sensing digital image classification of natural scenes when the classes of interest are spectrally very similar and high dimensional multispectral image data is available. A number of the investigations that have been developed in this area deal with the fact that as the data dimensionality increases, so does the difficulty in obtaining a reasonably accurate estimate of the within-class covariance matrices from the number of available labeled samples, which is usually limited. Several approaches have been proposed to deal with this problem. This study aims toward a complementary goal. Assuming that reasonably accurate estimates for the within-class covariance matrices have been obtained, we seek to better understand what kind of geometrically-oriented interpretation can be given to them as the data dimensionality increases and also to understand how this knowledge can help the design of a classifier. In order to achieve this goal, the covariance matrix is decomposed into a number of parameters that are then analyzed separately with respect to their ability to separate the classes. Methods for image classification based on these parameters are investigated. Results of tests using data provided by the sensor system AVIRIS are presented and discussed.

Fast algorithms for texture analysis using co-occurrence matrices

Texture analysis may be of great importance for the problem of image classification and recognition. Co-occurrence matrices are quite effective for discriminating different textures but have the disadvantage of a high computational cost. In the paper a fast algorithm for calculating parameters of co-occurrence matrices is presented. This method has been applied to the problem of classification and segmentation of artificial and natural scenes: the classification, based on co-occurrence matrix parameters, is implemented pixel-by-pixel by using supervised learning and maximum likelihood estimates. The problem of texture boundary recognition has also been considered and a classification scheme based on more than one window for each pixel is presented. Experimental results show the improvements of classification rates that can be achieved by using this method when compared to a single-window classification.