How Can Selection of Biologically Inspired Features Improve the Performance of a Robust Object Recognition Model?

Humans can effectively and swiftly recognize objects in complex natural scenes. This outstanding ability has motivated many computational object recognition models. Most of these models try to emulate the behavior of this remarkable system. The human visual system hierarchically recognizes objects in several processing stages. Along these stages a set of features with increasing complexity is extracted by different parts of visual system. Elementary features like bars and edges are processed in earlier levels of visual pathway and as far as one goes upper in this pathway more complex features will be spotted. It is an important interrogation in the field of visual processing to see which features of an object are selected and represented by the visual cortex. To address this issue, we extended a hierarchical model, which is motivated by biology, for different object recognition tasks. In this model, a set of object parts, named patches, extracted in the intermediate stages. These object parts are used for training procedure in the model and have an important role in object recognition. These patches are selected indiscriminately from different positions of an image and this can lead to the extraction of non-discriminating patches which eventually may reduce the performance. In the proposed model we used an evolutionary algorithm approach to select a set of informative patches. Our reported results indicate that these patches are more informative than usual random patches. We demonstrate the strength of the proposed model on a range of object recognition tasks. The proposed model outperforms the original model in diverse object recognition tasks. It can be seen from the experiments that selected features are generally particular parts of target images. Our results suggest that selected features which are parts of target objects provide an efficient set for robust object recognition.     

Language,Reading, and Math Learning Profiles in an Epidemiological Sample of School Age Children

Dyscalculia, dyslexia, and specific language impairment (SLI) are relatively specific developmental learning disabilities in math, reading, and oral language, respectively, that occur in the context of average intellectual capacity and adequate environmental opportunities. Past research has been dominated by studies focused on single impairments despite the widespread recognition that overlapping and comorbid deficits are common. The present study took an epidemiological approach to study the learning profiles of a large school age sample in language, reading, and math. Both general learning profiles reflecting good or poor performance across measures and specific learning profiles involving either weak language, weak reading, weak math, or weak math and reading were observed. These latter four profiles characterized 70% of children with some evidence of a learning disability. Low scores in phonological short-term memory characterized clusters with a language-based weakness whereas low or variable phonological awareness was associated with the reading (but not language-based) weaknesses. The low math only group did not show these phonological deficits. These findings may suggest different etiologies for language-based deficits in language, reading, and math, reading-related impairments in reading and math, and isolated math disabilities.     

Neural representation of objects in space: a dual coding account.

I present evidence on the nature of object coding in the brain and discuss the implications of this coding for models of visual selective attention. Neuropsychological studies of task-based constraints on: (i) visual neglect; and (ii) reading and counting, reveal the existence of parallel forms of spatial representation for objects: within-object representations, where elements are coded as parts of objects, and between-object representations, where elements are coded as independent objects. Aside from these spatial codes for objects, however, the coding of visual space is limited. We are extremely poor at remembering small spatial displacements across eye movements, indicating (at best) impoverished coding of spatial position per se. Also, effects of element separation on spatial extinction can be eliminated by filling the space with an occluding object, indicating that spatial effects on visual selection are moderated by object coding. Overall, there are separate limits on visual processing reflecting: (i) the competition to code parts within objects; (ii) the small number of independent objects that can be coded in parallel; and (iii) task-based selection of whether within- or between-object codes determine behaviour. Between-object coding may be linked to the dorsal visual system while parallel coding of parts within objects takes place in the ventral system, although there may additionally be some dorsal involvement either when attention must be shifted within objects or when explicit spatial coding of parts is necessary for object identification.     

"Brain reading": achievements, perspectives and ethical problems

The new trend in brain research designated as brain reading is considered. This research deals with decoding the informational content of the brain processing via its physiological parameters. Such studies are based on rather complicated methods of mathematical analysis. Single records rather than averaged data are used to reveal their content. Three main streams of studies are distinguished, i.e. the object classification, the emotion recognition and brainotyping. Particularly, the studies directed to recognizing the type of thinking via EEG spectra, carried out in the author's laboratory, are reviewed. The possible outcome of the brain reading technique is considered. Finally it is argued that in the future, the broad application of this technique needs to be controlled with some ethical rules.     

Agnosia.

Object recognition can break down in a variety of ways after brain damage. The resulting different forms of agnosia provide us with useful constraints on theories of normal object recognition. Recent studies suggest a division of labor for the recognition of different types of stimuli (common objects, words, faces, direction of eye gaze, spatial relations among parts of the human body), a high degree of interactivity in the processes underlying object recognition, and the possibility that recognition and awareness of recognition may be neurally distinct.     

Neuropsychological studies of object recognition

It is well established that disorders of visual perception are associated with lesions in the right hemisphere. Performances on tasks as disparate as the identification of objects from unusual views of objects drawn so as to overlap, of fragmented letters, of familiar faces, and of anomalous features in drawings, have been shown to be impaired in patients with focal right posterior lesions. A series of investigations are reviewed, directed towards analysing the basis of these deficits. Explanations in terms of primary visual impairment can be rejected, as can an account in terms of faulty figure-ground organization. It is argued that a wide variety of such perceptual deficits--all of which are concerned with meaningful visual stimuli--can be encompassed by the notion of faulty perceptual categorization at an early post-sensory stage of object recognition. Moreover, there is evidence suggesting that some of these various perceptual deficits can be mutually dissociated. The concept of perceptual categorization is discussed in the wider context of tentative model of object recognition.     

A hierarchical probabilistic model for rapid object categorization in natural scenes

Humans can categorize objects in complex natural scenes within 100-150 ms. This amazing ability of rapid categorization has motivated many computational models. Most of these models require extensive training to obtain a decision boundary in a very high dimensional (e.g., ～6,000 in a leading model) feature space and often categorize objects in natural scenes by categorizing the context that co-occurs with objects when objects do not occupy large portions of the scenes. It is thus unclear how humans achieve rapid scene categorization.To address this issue, we developed a hierarchical probabilistic model for rapid object categorization in natural scenes. In this model, a natural object category is represented by a coarse hierarchical probability distribution (PD), which includes PDs of object geometry and spatial configuration of object parts. Object parts are encoded by PDs of a set of natural object structures, each of which is a concatenation of local object features. Rapid categorization is performed as statistical inference. Since the model uses a very small number (～100) of structures for even complex object categories such as animals and cars, it requires little training and is robust in the presence of large variations within object categories and in their occurrences in natural scenes. Remarkably, we found that the model categorized animals in natural scenes and cars in street scenes with a near human-level performance. We also found that the model located animals and cars in natural scenes, thus overcoming a flaw in many other models which is to categorize objects in natural context by categorizing contextual features. These results suggest that coarse PDs of object categories based on natural object structures and statistical operations on these PDs may underlie the human ability to rapidly categorize scenes.     

Can Attention Be Confined to Just Part of a Moving Object? Revisiting Target-Distractor Merging in Multiple Object Tracking

While it was initially thought that attention was space-based, more recent work has shown that attention can also be object-based, in that observers find it easier to attend to different parts of the same object than to different parts of different objects. Such studies have shown that attention more easily spreads throughout an object than between objects. However, it is not known to what extent attention can be confined to just part of an object and to what extent attending to part of an object necessarily causes the entire object to be attended. We have investigated this question in the context of the multiple object tracking paradigm in which subjects are shown a scene containing a number of identical moving objects and asked to mentally track a subset of them, the targets, while not tracking the remainder, the distractors. Previous work has shown that joining each target to a distractor by a solid connector so that each target-distractor pair forms a single physical object, a technique known as target-distractor merging, makes it hard to track the targets, suggesting that attention cannot be restricted to just parts of objects. However, in that study the target-distractor pairs continuously changed length, which in itself would have made tracking difficult. Here we show that it remains difficult to track the targets even when the target-distractor pairs do not change length and even when the targets can be differentiated from the connectors that join them to the distractors. Our experiments suggest that it is hard to confine attention to just parts of objects, at least in the case of moving objects.     

Context effects on the neural correlates of recognition memory: an electrophysiological study

Event-related potentials (ERPs) were recorded during a recognition memory test for previously studied visual objects. Some studied objects were paired with the same context (landscape scenes) as at study, some were superimposed on a different studied context, and some were paired with new contexts. Unstudied objects were paired with either a studied or a new context. Three ERP memory effects were observed: an early effect elicited by all stimuli containing at least one studied component; a second effect elicited only by stimuli in which both object and context had been studied; and a third effect elicited by stimuli containing a studied object. Thus, test stimuli engaged three distinct kinds of memory-related neural activity which differed in their specificity for task-relevant features.     

Visual, haptic and cross-modal recognition of objects and scenes.

In this article we review current literature on cross-modal recognition and present new findings from our studies on object and scene recognition. Specifically, we address the questions of what is the nature of the representation underlying each sensory system that facilitates convergence across the senses and how perception is modified by the interaction of the senses. In the first set of our experiments, the recognition of unfamiliar objects within and across the visual and haptic modalities was investigated under conditions of changes in orientation (0 degrees or 180 degrees ). An orientation change increased recognition errors within each modality but this effect was reduced across modalities. Our results suggest that cross-modal object representations of objects are mediated by surface-dependent representations. In a second series of experiments, we investigated how spatial information is integrated across modalities and viewpoint using scenes of familiar, 3D objects as stimuli. We found that scene recognition performance was less efficient when there was either a change in modality, or in orientation, between learning and test. Furthermore, haptic learning was selectively disrupted by a verbal interpolation task. Our findings are discussed with reference to separate spatial encoding of visual and haptic scenes. We conclude by discussing a number of constraints under which cross-modal integration is optimal for object recognition. These constraints include the nature of the task, and the amount of spatial and temporal congruency of information across the modalities.     

Innate visual object recognition in vertebrates: some proposed pathways and mechanisms

Almost all vertebrates are capable of recognizing biologically relevant stimuli at or shortly after birth, and in some phylogenetically ancient species visual object recognition is exclusively innate. Extensive and detailed studies of the anuran visual system have resulted in the determination of the neural structures and pathways involved in innate prey and predator recognition in these species [Behav. Brain Sci. 10 (1987) 337; Comp. Biochem. Physiol. A 128 (2001) 417]. The structures involved include the optic tectum, pretectal nuclei and an area within the mesencephalic tegmentum. Here we investigate the structures and pathways involved in innate stimulus recognition in avian, rodent and primate species. We discuss innate stimulus preferences in maternal imprinting in chicks and argue that these preferences are due to innate visual recognition of conspecifics, entirely mediated by subtelencephalic structures. In rodent species, brainstem structures largely homologous to the components of the anuran subcortical visual system mediate innate visual object recognition. The primary components of the mammalian subcortical visual system are the superior colliculus, nucleus of the optic tract, anterior and posterior pretectal nuclei, nucleus of the posterior commissure, and an area within the mesopontine reticular formation that includes parts of the cuneiform, subcuneiform and pedunculopontine nuclei. We argue that in rodent species the innate sensory recognition systems function throughout ontogeny, acting in parallel with cortical sensory and recognition systems. In primates the structures involved in innate stimulus recognition are essentially the same as those in rodents, but overt innate recognition is only present in very early ontogeny, and after a transition period gives way to learned object recognition mediated by cortical structures. After the transition period, primate subcortical sensory systems still function to provide implicit innate stimulus recognition, and this recognition can still generate orienting, neuroendocrine and emotional responses to biologically relevant stimuli.     

Using an Improved SIFT Algorithm and Fuzzy Closed-Loop Control Strategy for Object Recognition in Cluttered Scenes

Partial occlusions, large pose variations, and extreme ambient illumination conditions generally cause the performance degradation of object recognition systems. Therefore, this paper presents a novel approach for fast and robust object recognition in cluttered scenes based on an improved scale invariant feature transform (SIFT) algorithm and a fuzzy closed-loop control method. First, a fast SIFT algorithm is proposed by classifying SIFT features into several clusters based on several attributes computed from the sub-orientation histogram (SOH), in the feature matching phase only features that share nearly the same corresponding attributes are compared. Second, a feature matching step is performed following a prioritized order based on the scale factor, which is calculated between the object image and the target object image, guaranteeing robust feature matching. Finally, a fuzzy closed-loop control strategy is applied to increase the accuracy of the object recognition and is essential for autonomous object manipulation process. Compared to the original SIFT algorithm for object recognition, the result of the proposed method shows that the number of SIFT features extracted from an object has a significant increase, and the computing speed of the object recognition processes increases by more than 40%. The experimental results confirmed that the proposed method performs effectively and accurately in cluttered scenes.     

Computational object recognition: a biologically motivated approach

We propose a conceptual framework for artificial object recognition systems based on findings from neurophysiological and neuropsychological research on the visual system in primate cortex. We identify some essential questions, which have to be addressed in the course of designing object recognition systems. As answers, we review some major aspects of biological object recognition, which are then translated into the technical field of computer vision. The key suggestions are the use of incremental and view-based approaches together with the ability of online feature selection and the interconnection of object-views to form an overall object representation. The effectiveness of the computational approach is estimated by testing a possible realization in various tasks and conditions explicitly designed to allow for a direct comparison with the biological counterpart. The results exhibit excellent performance with regard to recognition accuracy, the creation of sparse models and the selection of appropriate features.     

Small mobile sequences in bacteria display diverse structure/function motifs

Small repeat sequences in bacterial genomes, which represent non-autonomous mobile elements, have close similarities to archaeon and eukaryotic miniature inverted repeat transposable elements. These repeat elements are found in both intergenic and intragenic chromosomal regions, and contain an array of diverse motifs. These can include DNA sequences containing an integration host factor binding site and a proposed DNA methyltransferase recognition site, transcribed RNA secondary structural motifs, which are involved in mRNA regulation, and translated open reading frames found fused to other open reading frames. Some bacterial mobile element fusions are in evolutionarily conserved protein and RNA genes. Others might represent or lead to creation of new protein genes. Here we review the remarkable properties of these small bacterial mobile elements in the context of possible beneficial roles resulting from random insertions into the genome.     

Visual exploration and object recognition by lattice deformation

Mechanisms of explicit object recognition are often difficult to investigate and require stimuli with controlled features whose expression can be manipulated in a precise quantitative fashion. Here, we developed a novel method (called "Dots"), for generating visual stimuli, which is based on the progressive deformation of a regular lattice of dots, driven by local contour information from images of objects. By applying progressively larger deformation to the lattice, the latter conveys progressively more information about the target object. Stimuli generated with the presented method enable a precise control of object-related information content while preserving low-level image statistics, globally, and affecting them only little, locally. We show that such stimuli are useful for investigating object recognition under a naturalistic setting--free visual exploration--enabling a clear dissociation between object detection and explicit recognition. Using the introduced stimuli, we show that top-down modulation induced by previous exposure to target objects can greatly influence perceptual decisions, lowering perceptual thresholds not only for object recognition but also for object detection (visual hysteresis). Visual hysteresis is target-specific, its expression and magnitude depending on the identity of individual objects. Relying on the particular features of dot stimuli and on eye-tracking measurements, we further demonstrate that top-down processes guide visual exploration, controlling how visual information is integrated by successive fixations. Prior knowledge about objects can guide saccades/fixations to sample locations that are supposed to be highly informative, even when the actual information is missing from those locations in the stimulus. The duration of individual fixations is modulated by the novelty and difficulty of the stimulus, likely reflecting cognitive demand.     

The signal recognition particle receptor alpha subunit assembles co-translationally on the endoplasmic reticulum membrane during an mRNA-encoded translation pause in vitro.

Many proteins, including the alpha subunit of the signal recognition particle receptor (SR alpha), are targeted within the cell by poorly defined mechanisms. A 140 residue N-terminal domain of SR alpha targets and anchors the polypeptide to the endoplasmic reticulum membrane by a mechanism independent of the pathway involving the signal recognition particle. To investigate the mechanism of membrane anchoring, translation pause sites on the SR alpha mRNA were used to examine the targeting of translation intermediates. A strong pause site at nucleotide 507 of the mRNA open reading frame corresponded with the shortest nascent SR alpha polypeptide able to assemble on membranes. An mRNA sequence at this pause site that resembles a class of viral -1 frameshift sequences caused translation pausing when transferred into another mRNA context. Site-directed mutagenesis of the mRNA greatly reduced translation pausing without altering the polypeptide sequence, demonstrating unambiguously a role for this mRNA sequence in translation pausing. SR alpha polypeptides synthesized from the non-pausing mRNA were impaired in co-translational membrane anchoring. Furthermore, co-translational membrane assembly of SR alpha appears to anchor polysomes translating SR alpha to membranes.     

Multi-scale lines and edges in V1 and beyond: Brightness,object categorization and recognition,and consciousness

In this paper we present an improved model for line and edge detection in cortical area V1. This model is based on responses of simple and complex cells, and it is multi-scale with no free parameters. We illustrate the use of the multi-scale line/edge representation in different processes: visual reconstruction or brightness perception, automatic scale selection and object segregation. A two-level object categorization scenario is tested in which pre-categorization is based on coarse scales only and final categorization on coarse plus fine scales. We also present a multi-scale object and face recognition model. Processing schemes are discussed in the framework of a complete cortical architecture. The fact that brightness perception and object recognition may be based on the same symbolic image representation is an indication that the entire (visual) cortex is involved in consciousness.