Mechanisms Of Object Recognition Research Articles

SNN using color-opponent and attention mechanisms for object recognition

The current spiking neural network (SNN) relies on spike-timing-dependent plasticity (STDP) primarily for shape learning in object recognition tasks, overlooking the equally critical aspect of color information. To address this gap, our study introduces an unsupervised variant of STDP that incorporates principles from color-opponency mechanisms (COM) and classical receptive fields (CRF) found in the biological visual system, facilitating the integration of color information during parameter updates within the SNN architecture. Our approach initially preprocesses images into two distinct feature maps: one for shape and another for color. Then, signals derived from COM and intensity concurrently drive the STDP process, thereby updating parameters associated with both color and shape feature maps. Furthermore, we propose a channel-wise attention mechanism to enhance differentiation among objects sharing similar shapes or colors. Specifically, this mechanism utilizes convolution to generate an output spike-wave, identifying a winner based on earliest spike timing and maximal potential. The winning kernel computes attention, which is then applied via convolution to each input image feature map, generating post-feature maps. A STDP-like normalization rule compares firing times between pre- and post-feature maps, dynamically adjusting channel weights to optimize object recognition during the training phase.We assessed the proposed algorithm using SNN with both single-layer and multi-layer architectures across three datasets. Experimental findings highlight its efficacy and superiority in complex object recognition tasks compared to state-of-the-art (SOTA) algorithms. Notably, our approach achieved a significant 20% performance improvement over the SOTA on the Caltech-101 dataset. Moreover, the algorithm is well-suited for hardware implementation and energy efficiency, leveraging a winner-selection mechanism based on the earliest spike time.

Memory-related brain potentials for visual objects in early AD show impairment and compensatory mechanisms.

Impaired episodic memory is the primary feature of early Alzheimer's disease (AD), but not all memories are equally affected. Patients with AD and amnestic Mild Cognitive Impairment (aMCI) remember pictures better than words, to a greater extent than healthy elderly. We investigated neural mechanisms for visual object recognition in 30 patients (14AD, 16 aMCI) and 36 cognitively unimpaired healthy (19 in the "preclinical" stage of AD). Event-related brain potentials (ERPs) were recorded while participants performed a visual object recognition task. Hippocampal occupancy (integrity), amyloid (florbetapir) PET, and neuropsychological measures of verbal & visual memory, executive function were also collected. A right-frontal ERP recognition effect (500-700ms post-stimulus) was seen in cognitively unimpaired participants only, and significantly correlated with memory and executive function abilities. A later right-posterior negative ERP effect (700-900ms) correlated with visual memory abilities across participants with low verbal memory ability, and may reflect a compensatory mechanism. A correlation of this retrieval-related negativity with right hippocampal occupancy (r = 0.55), implicates the hippocampus in the engagement of compensatory perceptual retrieval mechanisms. Our results suggest that early AD patients are impaired in goal-directed retrieval processing, but may engage compensatory perceptual mechanisms which rely on hippocampal function.

Mechanisms of human dynamic object recognition revealed by sequential deep neural networks.

Humans can quickly recognize objects in a dynamically changing world. This ability is showcased by the fact that observers succeed at recognizing objects in rapidly changing image sequences, at up to 13 ms/image. To date, the mechanisms that govern dynamic object recognition remain poorly understood. Here, we developed deep learning models for dynamic recognition and compared different computational mechanisms, contrasting feedforward and recurrent, single-image and sequential processing as well as different forms of adaptation. We found that only models that integrate images sequentially via lateral recurrence mirrored human performance (N = 36) and were predictive of trial-by-trial responses across image durations (13-80 ms/image). Importantly, models with sequential lateral-recurrent integration also captured how human performance changes as a function of image presentation durations, with models processing images for a few time steps capturing human object recognition at shorter presentation durations and models processing images for more time steps capturing human object recognition at longer presentation durations. Furthermore, augmenting such a recurrent model with adaptation markedly improved dynamic recognition performance and accelerated its representational dynamics, thereby predicting human trial-by-trial responses using fewer processing resources. Together, these findings provide new insights into the mechanisms rendering object recognition so fast and effective in a dynamic visual world.

Using representational similarity analysis to reveal category and process specificity in visual object recognition

Cross-condition comparisons on neurodevelopmental conditions are central in neurodiversity research. In the realm of visual perception, the performance of participants with different category-specific disorders such as developmental prosopagnosia (problems with faces) and dyslexia (problems with words) have contributed to understanding of perceptual processes involved in word and face recognition. Alterations in face and word recognition are present in several neurodiverse populations, and improved knowledge about their relationship may increase our understanding of this variability of impairment. The present study investigates organizing principles of visual object processing and their implications for developmental disorders of recognition. Some accounts suggest that distinct mechanisms are responsible for recognizing objects of different categories, while others propose that categories share or even compete for cortical resources. We took an individual differences approach to estimate the relationship between abilities in recognition. Neurotypical participants (N = 97 after outlier exclusion) performed a match-to-sample task with faces, houses, and pseudowords. Either individual features or feature configurations were manipulated. To estimate the separability of visual recognition mechanisms, we used representational similarity analysis (RSA) where correlational matrices for accuracy were compared to predicted data patterns. Recognition abilities separated into face recognition on one hand and house/pseudoword recognition on the other, indicating that face recognition may rely on relatively selective mechanisms in neurotypicals. We also found evidence for a general visual object recognition mechanism, while some combinations of category (faces, houses, words) and processing type (featural, configural) likely rely on additional mechanisms. Developmental conditions may therefore reflect combinations of impaired and intact aspects of specific and general visual object recognition mechanisms, where featural and configural processes for one object category separate from the featural or configural processing of another. More generally, RSA is a promising approach for advancing understanding of neurodiversity, including shared aspects and distinctions between neurodevelopmental conditions of visual recognition.

Backscattering Analysis at ATR on Rough Surfaces by Ground-Based Polarimetric Radar Using Coherent Decomposition

This article deals with the analysis of backscattering at automatic target recognition (ATR) by ground-based radar located on rough terrain surfaces, using the properties of wave polarization. The purpose of the study is to examine and compare linear and circular polarized reflected waves, which can be described by decomposition theorems. Coherent decompositions (Pauli, Krogager, Cameron decomposition) are considered in the case of a rough terrain, for which the advantage of the Pauli decomposition has been shown. The article demonstrates an approach to the extraction of polarization signal backscattering data for two types of vehicles with different profiles. It is shown that the measurement results can be calibrated by a corner reflector that takes into account the properties of the ground surface, and further used for ATR based on supervised learning algorithms. The accuracy of object classification was 68.1% and 54.2% for the signal generated by linearly and elliptically polarized waves, respectively. Based on these results, we recommend using a linearly polarized wave as an object recognition mechanism. At the same time, any reflected depolarized wave significantly reshapes the structure due to the rotation of the object profile and the influence of a rough surface (vegetation fluctuations). This explains the low recognition accuracy in general.

Redundancy Mitigation Mechanism for Collective Perception in Connected and Autonomous Vehicles

Due to poor local range of the perception and object recognition mechanisms used by autonomous vehicles, incorrect decisions can be made, which can jeopardize a fully autonomous operation. A connected and autonomous vehicle should be able to combine its local perception with the perceptions of other vehicles to improve its capability to detect and predict obstacles. Such a collective perception system aims to expand the field of view of autonomous vehicles, augmenting their decision-making process, and as a consequence, increasing driving safety. Regardless of the benefits of a collective perception system, autonomous vehicles must intelligently select which data should be shared with who and when in order to conserve network resources and maintain the overall perception accuracy and time usefulness. In this context, the operational impact and benefits of a redundancy reduction mechanism for collective perception among connected autonomous vehicles are analyzed in this article. Therefore, we propose a reliable redundancy mitigation mechanism for collective perception services to reduce the transmission of inefficient messages, which is called VILE. Knowledge, selection, and perception are the three phases of the cooperative perception process developed in VILE. The results have shown that VILE is able to reduce it the absolute number of redundant objects of 75% and generated packets by up to 55%. Finally, we discuss possible research challenges and trends.

A Study on Fish Classification Techniques using Convolutional Neural Networks on Highly Challenged Underwater Images

Underwater Fish Species Recognition (UFSR) has attained significance because of evolving research in underwater life. Manual techniques to distinguish fish can be tricky and tedious. They might require enormous inspecting endeavours, but they can be costly. It results in limited data and a lack of human resources, which may cause incorrect object identification. Automating the fish species detection and recognition utilizing technology would assist sea life science to evolve further. UFSR in wild natural habitats is difficult because the images open natural habitat, complex background, and low luminance. Species Visualization can assist us with deep knowledge of the movements of the species underwater. Automation systems can help to classify the fish accurately and consistently. Image classification has been emerging research with the advancement of deep learning systems. The reason is that the convolutional neural networks (CNNs) don't require explicit feature extraction methods. The vast majority of the current object detection and recognition mechanisms are based on images in the outdoor environment. This paper mainly reviews the strategies proposed in the past years for underwater fish detection and classification. Further, the paper also presents the classification of three different underwater datasets using CNN with evaluation metrics.

Causal neural mechanisms of context-based object recognition.

Objects can be recognized based on their intrinsic features, including shape, color, and texture. In daily life, however, such features are often not clearly visible, for example when objects appear in the periphery, in clutter, or at a distance. Interestingly, object recognition can still be highly accurate under these conditions when objects are seen within their typical scene context. What are the neural mechanisms of context-based object recognition? According to parallel processing accounts, context-based object recognition is supported by the parallel processing of object and scene information in separate pathways. Output of these pathways is then combined in downstream regions, leading to contextual benefits in object recognition. Alternatively, according to feedback accounts, context-based object recognition is supported by (direct or indirect) feedback from scene-selective to object-selective regions. Here, in three pre-registered transcranial magnetic stimulation (TMS) experiments, we tested a key prediction of the feedback hypothesis: that scene-selective cortex causally and selectively supports context-based object recognition before object-selective cortex does. Early visual cortex (EVC), object-selective lateral occipital cortex (LOC), and scene-selective occipital place area (OPA) were stimulated at three time points relative to stimulus onset while participants categorized degraded objects in scenes and intact objects in isolation, in different trials. Results confirmed our predictions: relative to isolated object recognition, context-based object recognition was selectively and causally supported by OPA at 160-200 ms after onset, followed by LOC at 260-300 ms after onset. These results indicate that context-based expectations facilitate object recognition by disambiguating object representations in the visual cortex.

Does Neuronal Recycling Result in Destructive Competition? The Influence of Learning to Read on the Recognition of Faces

Written language, a human cultural invention, is far too recent a development for dedicated neural infrastructure to have evolved in its service. Newly acquired cultural skills, such as reading, thus recycle evolutionarily older circuits that originally evolved for different, but similar, functions (e.g., visual object recognition). The destructive-competition hypothesis predicts that this neuronal recycling has detrimental behavioral effects on the cognitive functions for which a cortical network originally evolved. In a study with 97 literate, low-literate, and illiterate participants from the same socioeconomic background, we found that even after adjusting for cognitive ability and test-taking familiarity, learning to read was associated with an increase, rather than a decrease, in object-recognition abilities. These results are incompatible with the claim that neuronal recycling results in destructive competition and are consistent with the possibility that learning to read instead fine-tunes general object-recognition mechanisms, a hypothesis that needs further neuroscientific investigation.

Offline transcranial direct current stimulation improves the ability to perceive crowded targets

The deleterious effect of nearby flankers on target identification in the periphery is known as visual crowding. Studying visual crowding can advance our understanding of the mechanisms of visual awareness and object recognition. Alleviating visual crowding is one of the major ways to improve peripheral vision. The aim of the current study was to examine whether transcranial direct current stimulation (tDCS) was capable of alleviating visual crowding at different visual eccentricities and with different visual tasks. In the present single-blind sham-controlled study, subjects were instructed to perform an orientation discrimination task or a letter identification task with isolated and crowded targets in the periphery, before and after applying 20 minutes of 2 mA anodal tDCS to visual cortex of the hemisphere contralateral or ipsilateral to visual stimuli. Contralateral tDCS significantly alleviated the orientation crowding effect at two different eccentricities and the letter crowding effect. This alleviation was absent after sham or ipsilateral stimulation and could not be fully explained by the performance improvement with the isolated targets. These findings demonstrated that offline tDCS was effective in alleviating visual crowding across different visual eccentricities and tasks, therefore providing a promising way to improve spatial vision rapidly in crowded scenes.

Converging Evidence That Neural Plasticity Underlies Transcranial Direct-Current Stimulation.

It is not definitely known how direct-current stimulation causes its long-lasting effects. Here, we tested the hypothesis that the long time course of transcranial direct-current stimulation (tDCS) is because of the electrical field increasing the plasticity of the brain tissue. If this is the case, then we should see tDCS effects when humans need to encode information into long-term memory, but not at other times. We tested this hypothesis by delivering tDCS to the ventral visual stream of human participants during different tasks (i.e., recognition memory vs. visual search) and at different times during a memory task. We found that tDCS improved memory encoding, and the neural correlates thereof, but not retrieval. We also found that tDCS did not change the efficiency of information processing during visual search for a certain target object, a task that does not require the formation of new connections in the brain but instead relies on attention and object recognition mechanisms. Thus, our findings support the hypothesis that direct-current stimulation modulates brain activity by changing the underlying plasticity of the tissue.

Offline transcranial direct current stimulation (tDCS) can improve the ability to perceive crowded targets

The deleterious influence of nearby flankers on target identification in the periphery is often referred to as visual crowding. Studying visual crowding can advance our understanding of the mechanisms of visual awareness and object recognition. Alleviating visual crowding (e.g., through perceptual learning) is one of the major ways to improve peripheral vision. Although there is a rapidly growing interest in using tDCS to modulate visual perception in humans, it remains unknown whether tDCS can alleviate visual crowding effects. We performed three experiments to investigate this issue. In Experiment 1, subjects were asked to perform an orientation discrimination task with the isolated and crowded targets in the periphery, before and after applying 20 minutes of 2 mA anodal tDCS to early visual cortex (P1 or P2) of the hemisphere contralateral or ipsilateral to the visual stimuli. We found that, electrical stimulation of the hemisphere contralateral to the visual stimuli could significantly reduce the crowding effect. This reduction was absent after the sham stimulation and could not be explained by the performance improvement with the isolated target. In Experiment 2, using the same behavioral task and the same tDCS protocol, we found that the contralateral DC stimulation remained effective in alleviating crowding at a smaller eccentricity. In Experiment 3, we adopted a letter recognition task and found that the alleviation of the letter crowding effect still existed after tDCS. In all the three experiments, no reduction was observed when tDCS was applied to the hemisphere ipsilateral to the visual stimuli. Taken together, we conclude that offline tDCS is effective in alleviating visual crowding across different visual eccentricities and perceptual tasks, which sheds new light on the mechanisms of visual crowding and possible practical applications.

Multi-Scale Semantic Segmentation and Spatial Relationship Recognition of Remote Sensing Images Based on an Attention Model

A comprehensive interpretation of remote sensing images involves not only remote sensing object recognition but also the recognition of spatial relations between objects. Especially in the case of different objects with the same spectrum, the spatial relationship can help interpret remote sensing objects more accurately. Compared with traditional remote sensing object recognition methods, deep learning has the advantages of high accuracy and strong generalizability regarding scene classification and semantic segmentation. However, it is difficult to simultaneously recognize remote sensing objects and their spatial relationship from end-to-end only relying on present deep learning networks. To address this problem, we propose a multi-scale remote sensing image interpretation network, called the MSRIN. The architecture of the MSRIN is a parallel deep neural network based on a fully convolutional network (FCN), a U-Net, and a long short-term memory network (LSTM). The MSRIN recognizes remote sensing objects and their spatial relationship through three processes. First, the MSRIN defines a multi-scale remote sensing image caption strategy and simultaneously segments the same image using the FCN and U-Net on different spatial scales so that a two-scale hierarchy is formed. The output of the FCN and U-Net are masked to obtain the location and boundaries of remote sensing objects. Second, using an attention-based LSTM, the remote sensing image captions include the remote sensing objects (nouns) and their spatial relationships described with natural language. Finally, we designed a remote sensing object recognition and correction mechanism to build the relationship between nouns in captions and object mask graphs using an attention weight matrix to transfer the spatial relationship from captions to objects mask graphs. In other words, the MSRIN simultaneously realizes the semantic segmentation of the remote sensing objects and their spatial relationship identification end-to-end. Experimental results demonstrated that the matching rate between samples and the mask graph increased by 67.37 percentage points, and the matching rate between nouns and the mask graph increased by 41.78 percentage points compared to before correction. The proposed MSRIN has achieved remarkable results.

Detection of recognition from ambiguous visual information using FMRI

Background: The functional neuroimaging has allowed the understanding of neural mechanisms of object recognition and language. However, the process of how we guess things from ambiguous information remains uncertain.

V4 shape features for contour representation and object detection

Cortical area V4 lies in the middle of the visual pathway involved with object recognition. Neurons in V4 selectively respond to different curve fragments along the object contour. In this paper, we propose a computational model that captures the shape features extracted by V4 neurons. The computational model emulated the information processing mechanism in the visual cortex. It extracted curve segments that V4 neurons respond to and quantitatively represented features of the curve segments. The proposed V4 shape features could describe object contours accurately and efficiently. With quantitative evaluation using the MPEG7 shape dataset, we showed that complex shapes could be represented with a very limited number of V4 shape features. Based on V4 features, we further developed a self-organizing map neural network to learn object shape models. The shape model was defined by a group of V4 features with constraints on their spatial relationships. The model was evaluated in object detection experiments using ETHZ objects and INRIA horses datasets. The proposed model could learn to recognize objects by shapes and accurately outline the object contour in the images. Thus, this model provides insight into the neural mechanisms of shape-based object recognition.

General object recognition is specific: Evidence from novel and familiar objects

In tests of object recognition, individual differences typically correlate modestly but nontrivially across familiar categories (e.g. cars, faces, shoes, birds, mushrooms). In theory, these correlations could reflect either global, non-specific mechanisms, such as general intelligence (IQ), or more specific mechanisms. Here, we introduce two separate methods for effectively capturing category-general performance variation, one that uses novel objects and one that uses familiar objects. In each case, we show that category-general performance variance is unrelated to IQ, thereby implicating more specific mechanisms. The first approach examines three newly developed novel object memory tests (NOMTs). We predicted that NOMTs would exhibit more shared, category-general variance than familiar object memory tests (FOMTs) because novel objects, unlike familiar objects, lack category-specific environmental influences (e.g. exposure to car magazines or botany classes). This prediction held, and remarkably, virtually none of the substantial shared variance among NOMTs was explained by IQ. Also, while NOMTs correlated nontrivially with two FOMTs (faces, cars), these correlations were smaller than among NOMTs and no larger than between the face and car tests themselves, suggesting that the category-general variance captured by NOMTs is specific not only relative to IQ, but also, to some degree, relative to both face and car recognition. The second approach averaged performance across multiple FOMTs, which we predicted would increase category-general variance by averaging out category-specific factors. This prediction held, and as with NOMTs, virtually none of the shared variance among FOMTs was explained by IQ. Overall, these results support the existence of object recognition mechanisms that, though category-general, are specific relative to IQ and substantially separable from face and car recognition. They also add sensitive, well-normed NOMTs to the tools available to study object recognition.

The role of the lateral occipital cortex in aesthetic appreciation of representational and abstract paintings: A TMS study

Neuroimaging studies of aesthetic appreciation have shown that activity in the lateral occipital area (LO)—a key node in the object recognition pathway—is modulated by the extent to which visual artworks are liked or found beautiful. However, the available evidence is only correlational. Here we used transcranial magnetic stimulation (TMS) to investigate the putative causal role of LO in the aesthetic appreciation of paintings. In our first experiment, we found that interfering with LO activity during aesthetic appreciation selectively reduced evaluation of representational paintings, leaving appreciation of abstract paintings unaffected. A second experiment demonstrated that, although the perceived clearness of the images overall positively correlated with liking, the detrimental effect of LO TMS on aesthetic appreciation does not owe to TMS reducing perceived clearness. Taken together, our findings suggest that object-recognition mechanisms mediated by LO play a causal role in aesthetic appreciation of representational art.

Mechanisms of object recognition: what we have learned from pigeons.

Behavioral studies of object recognition in pigeons have been conducted for 50 years, yielding a large body of data. Recent work has been directed toward synthesizing this evidence and understanding the visual, associative, and cognitive mechanisms that are involved. The outcome is that pigeons are likely to be the non-primate species for which the computational mechanisms of object recognition are best understood. Here, we review this research and suggest that a core set of mechanisms for object recognition might be present in all vertebrates, including pigeons and people, making pigeons an excellent candidate model to study the neural mechanisms of object recognition. Behavioral and computational evidence suggests that error-driven learning participates in object category learning by pigeons and people, and recent neuroscientific research suggests that the basal ganglia, which are homologous in these species, may implement error-driven learning of stimulus-response associations. Furthermore, learning of abstract category representations can be observed in pigeons and other vertebrates. Finally, there is evidence that feedforward visual processing, a central mechanism in models of object recognition in the primate ventral stream, plays a role in object recognition by pigeons. We also highlight differences between pigeons and people in object recognition abilities, and propose candidate adaptive specializations which may explain them, such as holistic face processing and rule-based category learning in primates. From a modern comparative perspective, such specializations are to be expected regardless of the model species under study. The fact that we have a good idea of which aspects of object recognition differ in people and pigeons should be seen as an advantage over other animal models. From this perspective, we suggest that there is much to learn about human object recognition from studying the “simple” brains of pigeons.

Shape from shading in starlings (Sturnus vulgaris).

Birds behave as if they quickly and accurately perceive an object-filled visual world. Beyond the extensive research with pigeons, however, there is a large and important gap in our knowledge about the mechanisms of object perception and recognition in other avian visual systems. The pattern of shading reflected from the surfaces of objects is one important optical feature that provides fundamental information about shape. To better understand how surface and object shading is processed by a passerine species, 5 starlings were tested with differentially illuminated convex and concave curved surfaces in 3 experiments using a simultaneous visual discrimination procedure. Starlings rapidly learned this shape-from-shading discrimination independent of varied lighting direction, surface color, and camera perspective. Variations in the pattern of lighting through experimental manipulations of camera perspective, surface height, contrast, material specularity, and surface shape were consistent with the hypothesis that the starlings perceived these illuminated surfaces as having 3-dimensional shape, similar to results previously collected with pigeons. These similarities across different orders of birds indicate that the relative shading of objects in a visual scene is a highly salient feature for shape processing in birds and is likely a highly conserved visual process that is widely distributed within this class of animal.

Facilitation of naming in aphasia with auditory repetition: An investigation of neurocognitive mechanisms

Prior phonological processing can enhance subsequent picture naming performance in individuals with aphasia, yet the neurocognitive mechanisms underlying this effect and its longevity are unknown. This study used functional magnetic resonance imaging to examine the short-term (within minutes) and long-term (within days) facilitation effects from a phonological task in both participants with aphasia and age-matched controls. Results for control participants suggested that long-term facilitation of subsequent picture naming may be driven by a strengthening of semantic–phonological connections, while semantic and object recognition mechanisms underlie more short-term effects. All participants with aphasia significantly improved in naming accuracy following both short- and long-term facilitation. A descriptive comparison of the neuroimaging results identified different patterns of activation for each individual with aphasia. The exclusive engagement of a left hemisphere phonological network underlying facilitation was not revealed. The findings suggest that improved naming in aphasia with phonological tasks may be supported by changes in right hemisphere activity in some individuals and reveal the potential contribution of the cerebellum to improved naming following phonological facilitation. Conclusions must be interpreted with caution, however, due to the comparison of corrected group control results to that of individual participants with aphasia, which were not corrected for multiple comparisons.