Feature Cues Research Articles

Cue responsiveness as a measure of emerging language ability in aphasia

ABSTRACT Background Prior research suggests that initial aphasia severity, lesion size, and lesion location are the most salient factors in predicting recovery outcomes. While these factors provide important prognostic information, information that is individualized and readily available to clinicians is limited. Deficits in naming are common to all aphasia types and are routinely targeted in aphasia assessment and treatment, with cues provided to facilitate lexical retrieval. Objectives In this study, we examine aphasia recovery factors that are readily available to clinicians, examining whether a person’s ability to improve naming with cues, indicating “stimulability,” will be predictive of future word retrieval. Methods Ten participants with aphasia following a left-hemisphere stroke participated in initial assessment, seven of whom met criteria for longitudinal assessment. Stroke and early clinical recovery data were collected for all participants. At four timepoints over one year we evaluated longitudinal participants’ naming ability and measured the proportion of successful lexical retrieval with the presentation of phonemic, feature, and sentence cues. Results For all participants, multiple descriptive factors regarding recovery, including lesion information, information from the acute inpatient timeframe, and communication opportunities, were examined. For individuals followed longitudinally, naming stimulability did not consistently predict naming accuracy at the subsequent assessment timepoint. Individuals’ attempts at naming emerged as a metric related to future naming performance warranting further evaluation. Conclusions Multiple factors related to recovery must be considered when providing prognostic information. Naming stimulability and attempts at naming provide some information regarding future performance, but are not consistently reliable across timepoints.

Do Older and Young Adults Learn to Integrate Geometry While Navigating in an Environment of a Serious Game?

We evaluated the outcomes of an intervention using a serious game designed to be played on iPads for improving spatial reorientation by training users to integrate geometry of the environment, instead of relying solely on featural cues. Using data logged online through a clinical study of using this game, the effect of training among 16 older adults (69.3 ± 6.4 years, 4 males), who played the game repeatedly (self-administered) over a period of 8 weeks, was investigated. The game contains a hexagonal room with 3 objects, textured walls, and grids on the floor, which are removed one by one as the participant played the game. In each level, the room also rotates such that the viewpoint of the user is different from that of the previous level. Participants cannot play a higher level unless they make no mistake during the trials of the lower test level. In addition to data of older adults available from that clinical trial, we recruited 16 young adults (27.3 ± 5.6 years, 4 males) to play the game for 5 sessions and compared their results with those of the older adults. We evaluated the error type made in each test level and the scores for each session among older adults. Further, we compared the frequency of each error type between young and older adults during the test levels that a landmark adjacent to the target was removed over the first 5 sessions. The results of older adults’ performance suggest they learned to make fewer mistakes over the sessions. Also, both young and older adults learned to integrate the geometrical cues rather than relying on the landmark cue adjacent to the target to find the target. Overall, the results indicate the designed hexagonal room game can enhance spatial cognition among all age groups of adults.

Detection and Localization of Mask Occluded Faces by transfer learning using Faster RCNN

Here in this work we utilized the novel transfer learning based technique, where a pretrained network model weights are used to train our Faster Region Convolutional Neural Network (FRCNN). With the help of captured mask-face dataset, we further propose a masked face detection and recognition work. It incorporate three major network training modules. Proposed network module cascade two pre-trained CNNs to extract target facial features and strong region of interest (ROI) from the input dataset image for transformed domain represention with higher spatial descriptors. By similar method, omitted facial feature cues are sufficiently reconstructed and the distortion introduced by noise pixs cues by masks are reduced. Lastly, the pre-trained CNN module is used to detect and identify masked face regions of the image.The network estimates mask detection accuracy by putting bounding boxs around all the occluded mask face for performing the classification and line regression work. Experiments conducted on proposed dataset shows good accuracy and reduced running time overfew state-of-the-arts by 11.46 %.

Impaired mechanism of visual focal attention in posterior cortical atrophy.

Objective: Simultanagnosia, a deficit in holistic visual perception, is among the most prominent features of posterior cortical atrophy (PCA). Deficits in visuoperceptual and attentional mechanisms could contribute to simultanagnosia. In the present study, we explored the impaired visual perception of global configuration with two main hypotheses: (a) It is due to a deficit in processing low-spatial frequency stimuli, and (b) it arises from deficits in adjusting attentional focus. Method: The visuoperceptual mechanism was explored by asking participants (5 PCA patients and 20 age- and education-matched healthy controls) to report the local and global elements of incongruent hierarchical letters. Stimuli were unbiased (black letters/white background) and parvocellular biased (red letters/green background). A cued T-detection task, where the stimulus onset asynchrony and the cues' features varied, was used to explore focal attention. Results: PCA patients systematically failed in reporting the global but not the local element. The parvocellular-biased condition partially improved the performance in only 1 patient. In the T-detection task, controls responded faster to targets cued by red dots and small cues as compared to no cues. Conversely, the cue's features did not affect patients' performance. Conclusions: Results only partially support the hypothesis according to which simultanagnosia is driven by an impairment in processing low-spatial frequencies. Data indicate a deficit in the flexibility of focal attention that prevents PCA patients from adapting the attentional window to the stimulus features. Simultanagnosia in PCA can be conceptualized as a complex result of a deficit involving visuoperceptual and exogenous attentional mechanisms. (PsycInfo Database Record (c) 2020 APA, all rights reserved).

Detection of replay spoof speech using teager energy feature cues

The vulnerability of Automatic Speaker Verification (ASV) systems to spoofing or presentation attacks is still an open security issue. In this context, replay spoofing attacks pose a great threat to an ASV system since they can be easily performed (using a playback device, and without needing any technical skill). In this paper, we analyze replay speech signals in terms of reverberation that may occur during recording of the speech signal. Such reverberation introduces delay and changes in amplitude, producing close copies of speech signals, which significantly influences the replay components. To that effect, we propose to exploit the capabilities of the Teager Energy Operator (TEO) to compute a running estimate of subband energies for replay vs. genuine signals. We have used a linearly-spaced Gabor filterbank to obtain a narrowband filtered signal. The TEO has the ability to track the instantaneous changes of a signal. Experiments are performed on the ASVspoof 2017 Challenge version 2.0 database using a Gaussian Mixture Model (GMM) as pattern classifier. Furthermore, we compared our results with state-of-the-art feature sets, namely, Constant Q Cepstral Coefficients (CQCC), Linear Frequency Cepstral Coefficients (LFCC), Mel Frequency Cepstral Coefficients (MFCC), and used their score-level fusion with the proposed feature sets, i.e., Teager Energy Cepstral Coefficients (TECC), in order to obtain possible complementary information that further reduces the Equal Error Rate (EER). Relatively low EERs are obtained with score-level fusion of CQCC, MFCC, LFCC, and TECC feature sets, resulted in 6.68% and 10.45% on development and evaluation sets, respectively. Moreover, for the evaluation dataset, we also studied the performance of the TECC feature set on different Replay Configurations (RC), namely, for acoustic environments, playback, and recording devices. For all the levels of threat conditions (i.e., low, medium, and high-level) to an ASV system, the proposed feature set performed better compared to existing state-of-the-art feature sets. In addition to the ASVspoof 2017 Challenge database, we also performed experiments on other spoofing databases, namely, the ASVspoof 2015 Challenge, BTAS 2016, and ASVspoof 2019 Challenge databases. For all the spoofing databases used in this study, the proposed TECC feature set perform significantly better than the other feature sets.

Distinct and combined responses to environmental geometry and features in a working-memory reorientation task in rats and chicks

The original provocative formulation of the ‘geometric module’ hypothesis was based on a working-memory task in rats which suggested that spontaneous reorientation behavior is based solely on the environmental geometry and is impervious to featural cues. Here, we retested that claim by returning to a spontaneous navigation task with rats and domestic chicks, using a single prominent featural cue (a striped wall) within a rectangular arena. Experiments 1 and 2 tested the influence of geometry and features separately. In Experiment 1, we found that both rats and chicks used environmental geometry to compute locations in a plain rectangular arena. In Experiment 2, while chicks failed to spontaneously use a striped wall in a square arena, rats showed a modest influence of the featural cue as a local marker to the goal. The critical third experiment tested the striped wall inside the rectangular arena. We found that although chicks solely relied on geometry, rats navigated based on both environmental geometry and the featural cue. While our findings with rats are contrary to classic claims of an impervious geometric module, they are consistent with the hypothesis that navigation by boundaries and features may involve distinct underlying cognitive computations. We conclude by discussing the similarities and differences in feature-use across tasks and species.

Social motivation and the use of distal, but not local, featural cues are related to ranging behavior in free-range chickens (Gallus gallus domesticus).

Animals can navigate an environment relying on different sources of information, such as geometrical or featural cues. The favoring of one type of information over another depends on multiple factors, such as inter-individual differences in behavior and cognition. Free-range chickens present different range use patterns, which may be explained by behavioral and cognitive differences. However, how behavior, cognition, and range use intercorrelate is still poorly understood. In this work, we aimed to further understand possible differences in behavior and cognition between two groups of free-range broiler chickens: those who frequently explore their range ('high rangers') and those who prefer to stay in or near the barn ('low rangers'). Prior to range access, individual behavior was measured in open field-, emergence-, and social motivation tests. To investigate cognitive differences, we analyzed whether exploratory behavior was linked to different performances in the use of distal and local spatial cues during an orientation task. During the social motivation test, low rangers showed a higher inclination to be near conspecifics than did high rangers. Our orientation tests show that chickens preferred to orientate themselves using the local cues over the distal cues. Individual differences were only found for distal, but not for local, cue use suggesting that demanding tasks are more efficient in revealing individual cognitive differences. Our results suggest that considering variation in social motivation may allow a more comprehensive understanding of chicken range use. Our results also support the importance of incorporating multiple aspects of individual differences to understand individual reactions to its environment.

Limited interactions between space- and feature-based attention in visually sparse displays.

Top-down visual attention selectively filters sensory input so relevant information receives preferential processing. Feature-based attention (FBA) enhances the representation of relevant low-level features, whereas space-based attention (SBA) enhances information at relevant location(s). The present study investigates whether the unique influences of SBA and FBA combine to facilitate behavior in a perceptually demanding discrimination task. We first demonstrated that, independently, both color and location pre-cues could effectively direct attention to facilitate perceptual decision making of a target. We then examined the combined effects of SBA and FBA in the same design by deploying a predictive color arrow pre-cue. Only SBA effects were observed in performance accuracy and reaction time. However, we detected a reaction time cost when a valid spatial cue was paired with a feature cue. A computational perceptual decision-making model largely provided converging evidence that contributions from FBA were restricted to facilitating the speed with which the relevant item was identified. Our results suggest that both selection mechanisms can be used in isolation to resolve a perceptually challenging target in a sparse display, but with little additive perceptual benefit when cued simultaneously. We conclude that there is at least some higher order interdependence between space-based and feature-based selection during decision making under specific conditions.

DV-Net: Dual-view network for 3D reconstruction by fusing multiple sets of gated control point clouds

Deep learning for 3D reconstruction have just shown some promising advantanges, where 3D shapes can be predicted from a single RGB image. However, such works are often limited by single feature cue, which does not capture the 3D shape of objects well. To address this problem, an end-to-end 3D reconstruction approach that predicts 3D point cloud from dual-view RGB images is proposed in this paper. It consists of several processing parts. A dual-view 3D reconstruction network is proposed for 3D reconstruction, which predicts object’s point clouds by exploiting two RGB images with different views, and avoids the limitation of single feature cue. Another structure feature learning network is performed to extract the structure features with stronger representation ability from point clouds. A gated control network for data fusion is proposed to gather point clouds. It takes two sets of point clouds with different views as input and fuses them. The proposed approach is thoroughly evaluated with extensive experiments on the widely-used ShapeNet dataset. Both the qualitative results and quantitative analysis demonstrate that this method not only captures the detailed geometric structures of 3D shapes for different object categories with complex topologies, but also achieves state-of-the-art performance.

Underwater Target Tracking Using Forward-Looking Sonar for Autonomous Underwater Vehicles.

In the scenario where autonomous underwater vehicles (AUVs) carry out tasks, it is necessary to reliably estimate underwater-moving-target positioning. While cameras often give low-precision visibility in a limited field of view, the forward-looking sonar is still an attractive method for underwater sensing, which is especially effective for long-range tracking. This paper describes an online processing framework based on forward-looking-sonar (FLS) images, and presents a novel tracking approach based on a Gaussian particle filter (GPF) to resolve persistent multiple-target tracking in cluttered environments. First, the character of acoustic-vision images is considered, and methods of median filtering and region-growing segmentation were modified to improve image-processing results. Second, a generalized regression neural network was adopted to evaluate multiple features of target regions, and a representation of feature subsets was created to improve tracking performance. Thus, an adaptive fusion strategy is introduced to integrate feature cues into the observation model, and the complete procedure of underwater target tracking based on GPF is displayed. Results obtained on a real acoustic-vision AUV platform during sea trials are shown and discussed. These showed that the proposed method is feasible and effective in tracking targets in complex underwater environments.

Neural Switch Asymmetry in Feature-Based Auditory Attention Tasks.

Active listening involves dynamically switching attention between competing talkers and is essential to following conversations in everyday environments. Previous investigations in human listeners have examined the neural mechanisms that support switching auditory attention within the acoustic featural cues of pitch and auditory space. Here, we explored the cortical circuitry underlying endogenous switching of auditory attention between pitch and spatial cues necessary to discern target from masker words. Because these tasks are of unequal difficulty, we expected an asymmetry in behavioral switch costs for hard-to-easy versus easy-to-hard switches, mirroring prior evidence from vision-based cognitive task-switching paradigms. We investigated the neural correlates of this behavioral switch asymmetry and associated cognitive control operations in the present auditory paradigm. Behaviorally, we observed no switch-cost asymmetry, i.e., no performance difference for switching from the more difficult attend-pitch to the easier attend-space condition (P→S) versus switching from easy-to-hard (S→P). However, left lateral prefrontal cortex activity, correlated with improved performance, was observed during a silent gap period when listeners switched attention from P→S, relative to switching within pitch cues. No such differential activity was seen for the analogous easy-to-hard switch. We hypothesize that this neural switch asymmetry reflects proactive cognitive control mechanisms that successfully reconfigured neurally-specified task parameters and resolved competition from other such "task sets," thereby obviating the expected behavioral switch-cost asymmetry. The neural switch activity observed was generally consistent with that seen in cognitive paradigms, suggesting that established cognitive models of attention switching may be productively applied to better understand similar processes in audition.

Robust Visual Tracking Using Multi-Frame Multi-Feature Joint Modeling

It remains a huge challenge to design effective and efficient trackers under complex scenarios, including occlusions, illumination changes and pose variations. To cope with this problem, a promising solution is to integrate the temporal consistency across consecutive frames and multiple feature cues in a unified model. Motivated by this idea, we propose a novel correlation filter-based tracker in this work, in which the temporal relatedness is reconciled under a multi-task learning framework and the multiple feature cues are modeled using a multi-view learning approach. We demonstrate the resulting regression model can be efficiently learned by exploiting the structure of blockwise diagonal matrix. A fast blockwise diagonal matrix inversion algorithm is developed thereafter for efficient online tracking. Meanwhile, we incorporate an adaptive scale estimation mechanism to strengthen the stability of scale variation tracking. We implement our tracker using two types of features and test it on two benchmark datasets. Experimental results demonstrate the superiority of our proposed approach when compared with other state-of-the-art trackers. project homepage http://bmal.hust.edu.cn/project/KMF2JMTtracking.html

Functional specialization for feature-based and symmetry-based groupings in multiple object tracking

Our vision systems utilize attention to process dynamic visual information everyday. Multiple object tracking (MOT) is widely used for studying sustained attention in dynamic environments. Previous research reported inter-target grouping based on feature similarity (e.g., targets sharing the same color, size, or shape) can facilitate tracking. A recent study also found that symmetric relation among targets in MOT tasks can automatically improve tracking performance, and a greater grouping effect based on symmetry and feature was observed. However, grouping based on multiple feature cues didn't produce a greater grouping effect than grouping based on one feature cue. Why were groupings based on symmetry and feature additive, but two feature-based groupings were not? How stimulus-driven inter-target groupings based on color or symmetry improve the goal-directed target tracking process remains unknown. We hypothesized that symmetry-based and feature-based groupings would involve different cortical and subcortical areas, and color-based and color-shape-based groupings would be manifested in significant activity in the same regions. This study used functional magnetic resonance imaging (fMRI) to identify cortical and subcortical areas involved in feature-based and symmetry-based groupings respectively during tracking. The results of our experiment showed that lateral occipital (LO) cortex is involved in symmetry-based grouping during tracking. Bilateral putamen, temporal parietal junction (TPJ), and frontal eye field (FEF) are involved in the feature-based grouping. Our findings likely indicate that stimulus-driven inter-target symmetry-based grouping and feature-based grouping improve goal-directed target tracking in different ways.

Associative mechanisms underlying the function of satiety cues in the control of energy intake and appetitive behavior

While previous research has identified a number of metabolic, neural, and hormonal events that could serve as potential satiety signals, the mechanisms that enable satiety signals to suppress food-seeking and eating behavior remain poorly specified. Here we investigate the idea that the inhibitory power of satiety signals is derived, at least in part, from their ability to signal that foods and food-related stimuli will not be followed by reinforcing postingestive consequences. Viewed in this way, the signaling relationship in which satiety cues are embedded defines what is known in Pavlovian conditioning as a "serial feature negative" (sFN) discrimination problem. In this problem a "negative feature" cue precedes the presentation of a "target" cue on trials without reinforcement. In contrast, the target is reinforced on trials when the negative feature cue is not presented. Satiety cues can be seen as paralleling the function of negative feature cues in that they signal when food-related target cues will be nonreinforced. We conducted two experiments with rats that assessed if satiety signals functioned like negative feature stimuli. Experiment 1 explicitly pretrained satiety cues as negative feature stimuli, irrelevant stimuli, or under conditions where their ability to serve as negative feature stimuli would be attenuated. Control by satiety cues was highly sensitive to these experimental contingencies, with the best performance exhibited by rats given sFN pretraining. This sFN pretraining also transferred to enhance performance during subsequent training on another sFN problem with both external and internal negative feature cues. We also found that discriminative control by satiety cues blocked the development of that control by external cues. Experiment 2 evaluated whether a manipulation known to impair sFN performance with external negative feature cues (i.e., maintenance on a western diet) would also impair sFN performance when satiety cues were trained as negative feature stimuli. The results showed that compared to standard chow, WD intake impaired sFN performance similarly with both types of stimuli. These experiments provide evidence that an associative mechanism, like that underlying sFN performance, is involved with the control of appetitive behavior by satiety cues.

Experience with featural-cue reliability influences featural- and geometric-cue use by mice (Mus musculus).

Orienting is a critical skill for all mobile animals. Two commonly studied visual components used to guide orientation in an environment are geometric (e.g., distance or direction) and featural cues (e.g., color or texture). Previous research has shown that visual-cue use and cue weighing can depend on the navigator's previous experience, the nature and reliability of the cues, and genetic factors. Accordingly, the domestic mouse (Mus musculus) is a species of increasing interest because of its potential as a model for human neurological disorders with associated spatial disorientation, as is seen in Alzheimer's disease. In the present study, adult C57BL/6 mice were trained to search for a hidden food reward in one corner of a rectangular environment with featural information displayed continuously along the walls. After training, one group of mice was given a block of testing in which the featural information was removed, followed by a second block of testing in which the featural information was put in conflict with the learned configuration of featural and geometric cues. A second group of mice was given the same set of tests, but in the reverse order. Our results show that the mice incidentally encoded the geometry of the environment if they had experience with featural cues being unreliable prior to tests, during which featural cues were completely removed (unstable). Furthermore, we found when featural and geometric cues provide conflicting spatial information, this unreliability of featural cues over the course of the study may influence cue weighing. (PsycINFO Database Record

The effect of feature-based attention on flanker interference processing: An fMRI-constrained source analysis

The present study examined whether feature-based cueing affects early or late stages of flanker conflict processing using EEG and fMRI. Feature cues either directed participants’ attention to the upcoming colour of the target or were neutral. Validity-specific modulations during interference processing were investigated using the N200 event-related potential (ERP) component and BOLD signal differences. Additionally, both data sets were integrated using an fMRI-constrained source analysis. Finally, the results were compared with a previous study in which spatial instead of feature-based cueing was applied to an otherwise identical flanker task. Feature-based and spatial attention recruited a common fronto-parietal network during conflict processing. Irrespective of attention type (feature-based; spatial), this network responded to focussed attention (valid cueing) as well as context updating (invalid cueing), hinting at domain-general mechanisms. However, spatially and non-spatially directed attention also demonstrated domain-specific activation patterns for conflict processing that were observable in distinct EEG and fMRI data patterns as well as in the respective source analyses. Conflict-specific activity in visual brain regions was comparable between both attention types. We assume that the distinction between spatially and non-spatially directed attention types primarily applies to temporal differences (domain-specific dynamics) between signals originating in the same brain regions (domain-general localization).

Image Super-Resolution Reconstruction Based on Disparity Map and CNN

This paper presents a new method for registration of digital micro-mirror device (DMD) cameras through image reconstruction. The method called dual-channel super-resolution reconstruction (DCSR) uses a dual-channel input convolutional neural network method to iteratively obtain high resolution images. The originality of this study relies on the input color weight disparity map that enables extracting edges cues and feature cues effectively during the super-resolution. The iterative initial disparity map is obtained by a cost aggregation method based on the color weights, and the model parameters are updated during the iterative process. To validate performance, a series of the comparison experiments is developed on different network structures and parameter settings for the diverse deconv-pooling layers, convolution kernel sizes, and activation function. Experimental results show that the DCSR method can solve the sub-pixel registration problem of DMD cameras and achieve a good image reconstruction effect. The DCSR performs higher accuracy and an efficient process compared with traditional reconstruction made by single-channel input convolution neural network. The spatial resolution of the reconstructed image is significantly improved, and the detail resolution is enhanced.

Sex differences and the effect of instruction on reorientation abilities by humans.

This study examined whether differences in the amount of information provided to men and women, in the form of verbal instruction, influenced their encoding during a reorientation task. When a navigator needs to orient, featural (e.g., colour or texture) and geometry (e.g., metric information) are used to determine which direction to begin traveling. The current study used a spatial reorientation task to examine how men and women use featural and geometric cues and whether the content of the task's instructions influenced how these cues were used. Participants were trained to find a target location in a rectangular room with distinctive objects situated at each corner. Once the participants were accurately locating the target, various tests manipulating the spatial information were conducted. We found both men and women encoded the featural cues, and even though the features provided reliable information, participants generally showed an encoding of geometry. However, when participants were not provided with any information about the spatial aspects of the task in the instructions, they failed to encode geometry. We also found that women used distant featural cues as landmarks when the featural cue closest to the target was removed, whereas men did not. Yet, when the two types of cues were placed in conflict, both sexes weighed featural cues more heavily than geometric cues. The content of the task instructions also influenced how cues were relied upon in this conflict situation. Our results have important implications for our understanding of how spatial cues are used for reorientation.

Orientation in Virtual Reality Does Not Fully Measure Up to the Real-World

Adult participants learned to reorient to a specific corner inside either a real or virtual rectangular room containing a distinct featural object in each corner. Participants in the virtual-reality (VR) condition experienced an immersive virtual version of the physical room using a head-mounted display (HMD) and customized manual wheelchair to provide self-movement. Following a disorientation procedure, people could reorient by using either the geometry of the room and/or the distinct features in the corners. Test trials in which the different spatial cues were manipulated revealed participants encoded features and geometry in both the real and VR rooms. However, participants in the VR room showed less facility with using geometry. Our results suggest caution must be taken when interpreting the nuances of spatial cue use in virtual environments. Reduced reliability of geometric cues in VR environments may result in greater reliance on feature cues than would normally be expected under similar real-world conditions.

First Direct Evidence of Cue Integration in Reorientation: A New Paradigm.

There are several models of the use of geometric and feature cues in reorientation (Cheng, Huttenlocher, & Newcombe, ). The adaptive combination approach posits that people integrate cues with weights that depend on cue salience and learning, or, when discrepancies are large, they choose between cues based on these variables (Cheng, Shettleworth, Huttenlocher, & Rieser, ; Newcombe & Huttenlocher, ). In a new paradigm designed to evaluate integration and choice, disoriented participants attempted to return to a heading direction, in a trapezoidal enclosure in which feature and geometric cues both unambiguously specified a heading, but later the feature was moved. With discrepancies greater than 90 degrees, participants choose geometry. With smaller discrepancies, integration appeared in three of five situations; otherwise, participants used geometry alone. Variation depended on direction of feature movement and whether the nearest corner was acute or obtuse. The results have implications for contrasting adaptive combination and modularity theory, and for future research, offering a new paradigm for reorientation research, and for testing cue integration more broadly.