Semantic-level Visual Content Descriptor combining visual and textual cues as collaterals

In this paper, we introduce a novel high-level visual content descriptor which is devised for performing semantic-based image classification and retrieval. The work can be treated as an attempt to bridge the so called “semantic gap”. The proposed image feature vector model is fundamentally underpinned by the image labelling framework, called Collaterally Confirmed Labelling (CCL), which incorporates the collateral knowledge extracted from the collateral texts of the images with the state-of-the-art low-level image processing and visual feature extraction techniques for automatically assigning linguistic keywords to image regions. Two different high-level image feature vector models are developed based on the CCL labelling of results for the purposes of image data clustering and retrieval respectively. A subset of the Corel image collection has been used for evaluating our proposed method. The experimental results to-date already indicates that our proposed semantic-based visual content descriptors outperform both traditional visual and textual image feature models.

Motion detection and motion verbs: language affects low-level visual perception

Recent theories propose that semantic representation and sensorimotor processing have a common substrate via simulation. We tested the prediction that comprehension interacts with perception, using a standard psychophysics methodology.While passively listening to verbs that referred to upward or downward motion, and to control verbs that did not refer to motion, 20 subjects performed a motion-detection task, indicating whether or not they saw motion in visual stimuli containing threshold levels of coherent vertical motion. A signal detection analysis revealed that when verbs were directionally incongruent with the motion signal, perceptual sensitivity was impaired. Word comprehension also affected decision criteria and reaction times, but in different ways. The results are discussed with reference to existing explanations of embodied processing and the potential of psychophysical methods for assessing interactions between language and perception.

Collaterally cued labelling framework underpinning semantic-level visual content descriptor

In this paper, we introduce a novel high-level visual content descriptor devised for performing semantic-based image classification and retrieval. The work can be treated as an attempt for bridging the so called "semantic gap". The proposed image feature vector model is fundamentally underpinned by an automatic image labelling framework, called Collaterally Cued Labelling (CCL), which incorporates the collateral knowledge extracted from the collateral texts accompanying the images with the state-of-the-art low-level visual feature extraction techniques for automatically assigning textual keywords to image regions. A subset of the Corel image collection was used for evaluating the proposed method. The experimental results indicate that our semantic-level visual content descriptors outperform both conventional visual and textual image feature models.

The impact of novel labels on visual processing during infancy

The impact of novel labels on visual processing was investigated across two experiments with infants aged between 9 and 21 months. Infants viewed pairs of images across a series of preferential looking trials. On each trial, one image was novel, and the other image had previously been viewed by the infant. Some infants viewed images in silence; other infants viewed images accompanied by novel labels. The pattern of fixations both across and within trials revealed that infants in the labelling condition took longer to develop a novelty preference than infants in the silent condition. Our findings contrast with prior research by Robinson and Sloutsky (e.g., Robinson & Sloutsky, 2007a; Sloutsky & Robinson, 2008) who found that novel labels did not disrupt visual processing for infants aged over a year. Provided that overall task demands are sufficiently high, it appears that labels can disrupt visual processing for infants during the developmental period of establishing a lexicon. The results suggest that when infants are processing labels and objects, attentional resources are shared across modalities.

Manual response preparation and saccade programming are linked to attention shifts: ERP evidence for covert attentional orienting and spatially specific modulations of visual processing

The premotor theory of attention claims that attentional shifts are triggered during response programming, regardless of which response modality is involved. To investigate this claim, event-related brain potentials (ERPs) were recorded while participants covertly prepared a left or right response, as indicated by a precue presented at the beginning of each trial. Cues signalled a left or right eye movement in the saccade task, and a left or right manual response in the manual task. The cued response had to be executed or withheld following the presentation of a Go/Nogo stimulus. Although there were systematic differences between ERPs triggered during covert manual and saccade preparation, lateralised ERP components sensitive to the direction of a cued response were very similar for both tasks, and also similar to the components previously found during cued shifts of endogenous spatial attention. This is consistent with the claim that the control of attention and of covert response preparation are closely linked. N1 components triggered by task-irrelevant visual probes presented during the covert response preparation interval were enhanced when these probes were presented close to cued response hand in the manual task, and at the saccade target location in the saccade task. This demonstrates that both manual and saccade preparation result in spatially specific modulations of visual processing, in line with the predictions of the premotor theory.

Modeling data with structural and temporal correlation using lower level and higher level multilevel models

Novel imaging techniques are playing an increasingly important role in drug development, providing insight into the mechanism of action of new chemical entities. The data sets obtained by these methods can be large with complex inter-relationships, but the most appropriate statistical analysis for handling this data is often uncertain - precisely because of the exploratory nature of the way the data are collected. We present an example from a clinical trial using magnetic resonance imaging to assess changes in atherosclerotic plaques following treatment with a tool compound with established clinical benefit. We compared two specific approaches to handle the correlations due to physical location and repeated measurements: two-level and four-level multilevel models. The two methods identified similar structural variables, but higher level multilevel models had the advantage of explaining a greater proportion of variation, and the modeling assumptions appeared to be better satisfied.

Microsaccades distinguish between global and local visual processing

Much is known about the functional mechanisms involved in visual search. Yet, the fundamental question of whether the visual system can perform different types of visual analysis at different spatial resolutions still remains unsettled. In the visual-attention literature, the distinction between different spatial scales of visual processing corresponds to the distinction between distributed and focused attention. Some authors have argued that singleton detection can be performed in distributed attention, whereas others suggest that even such a simple visual operation involves focused attention. Here we showed that microsaccades were spatially biased during singleton discrimination but not during singleton detection. The results provide support to the hypothesis that some coarse visual analysis can be performed in a distributed attention mode.

Visual processing of optic flow and motor control in the human posterior cingulate sulcus

Previous studies have shown that the human posterior cingulate contains a visual processing area selective for optic flow (CSv). However, other studies performed in both humans and monkeys have identified a somatotopic motor region at the same location (CMA). Taken together, these findings suggested the possibility that the posterior cingulate contains a single visuomotor integration region. To test this idea we used fMRI to identify both visual and motor areas of the posterior cingulate in the same brains and to test the activity of those regions during a visuomotor task. Results indicated that rather than a single visuomotor region the posterior cingulate contains adjacent but separate motor and visual regions. CSv lies in the fundus of the cingulate sulcus, while CMA lies in the dorsal bank of the sulcus, slightly superior in terms of stereotaxic coordinates. A surprising and novel finding was that activity in CSv was suppressed during the visuomotor task, despite the visual stimulus being identical to that used to localize the region. This may provide an important clue to the specific role played by this region in the utilization of optic flow to control self-motion.

Semantic-associative visual content labelling and retrieval: A multimodal approach

A novel framework referred to as collaterally confirmed labelling (CCL) is proposed, aiming at localising the visual semantics to regions of interest in images with textual keywords. Both the primary image and collateral textual modalities are exploited in a mutually co-referencing and complementary fashion. The collateral content and context-based knowledge is used to bias the mapping from the low-level region-based visual primitives to the high-level visual concepts defined in a visual vocabulary. We introduce the notion of collateral context, which is represented as a co-occurrence matrix of the visual keywords. A collaborative mapping scheme is devised using statistical methods like Gaussian distribution or Euclidean distance together with collateral content and context-driven inference mechanism. We introduce a novel high-level visual content descriptor that is devised for performing semantic-based image classification and retrieval. The proposed image feature vector model is fundamentally underpinned by the CCL framework. Two different high-level image feature vector models are developed based on the CCL labelling of results for the purposes of image data clustering and retrieval, respectively. A subset of the Corel image collection has been used for evaluating our proposed method. The experimental results to-date already indicate that the proposed semantic-based visual content descriptors outperform both traditional visual and textual image feature models. (C) 2007 Elsevier B.V. All rights reserved.

Visual motion interferes with lexical decision on motion words

Embodied theories of cognition propose that neural substrates used in experiencing the referent of a word, for example perceiving upward motion, should be engaged in weaker form when that word, for example ‘rise’, is comprehended. Motivated by the finding that the perception of irrelevant background motion at near-threshold, but not supra-threshold, levels interferes with task execution, we assessed whether interference from near-threshold background motion was modulated by its congruence with the meaning of words (semantic content) when participants completed a lexical decision task (deciding if a string of letters is a real word or not). Reaction times for motion words, such as ‘rise’ or ‘fall’, were slower when the direction of visual motion and the ‘motion’ of the word were incongruent — but only when the visual motion was at nearthreshold levels. When motion was supra-threshold, the distribution of error rates, not reaction times, implicated low-level motion processing in the semantic processing of motion words. As the perception of near-threshold signals is not likely to be influenced by strategies, our results support a close contact between semantic information and perceptual systems.

N1 enhancement in synesthesia during visual and audio-visual perception in semantic cross-modal conflict situations: an ERP study

Synesthesia entails a special kind of sensory perception, where stimulation in one sensory modality leads to an internally generated perceptual experience of another, not stimulated sensory modality. This phenomenon can be viewed as an abnormal multisensory integration process as here the synesthetic percept is aberrantly fused with the stimulated modality. Indeed, recent synesthesia research has focused on multimodal processing even outside of the specific synesthesia-inducing context and has revealed changed multimodal integration, thus suggesting perceptual alterations at a global level. Here, we focused on audio-visual processing in synesthesia using a semantic classification task in combination with visually or auditory-visually presented animated and in animated objects in an audio-visual congruent and incongruent manner. Fourteen subjects with auditory-visual and/or grapheme-color synesthesia and 14 control subjects participated in the experiment. During presentation of the stimuli, event-related potentials were recorded from 32 electrodes. The analysis of reaction times and error rates revealed no group differences with best performance for audio-visually congruent stimulation indicating the well-known multimodal facilitation effect. We found enhanced amplitude of the N1 component over occipital electrode sites for synesthetes compared to controls. The differences occurred irrespective of the experimental condition and therefore suggest a global influence on early sensory processing in synesthetes.

Faces and awareness: low-level, not emotional factors determine perceptual dominance

Threat-relevant stimuli such as fear faces are prioritized by the human visual system. Recent research suggests that this prioritization begins during unconscious processing: A specialized (possibly subcortical) pathway evaluates the threat relevance of visual input, resulting in preferential access to awareness for threat stimuli. Our data challenge this claim. We used a continuous flash suppression (CFS) paradigm to present emotional face stimuli outside of awareness. It has been shown using CFS that salient (e.g., high contrast) and recognizable stimuli (faces, words) become visible more quickly than less salient or less recognizable stimuli. We found that although fearful faces emerge from suppression faster than other faces, this was wholly explained by their low-level visual properties, rather than their emotional content. We conclude that, in the competition for visual awareness, the visual system prefers and promotes unconscious stimuli that are more “face-like,” but the emotional content of a face has no effect on stimulus salience.

Instantaneous threat detection based on a semantic representation of activities, zones and trajectories

Threat detection is a challenging problem, because threats appear in many variations and differences to normal behaviour can be very subtle. In this paper, we consider threats on a parking lot, where theft of a truck’s cargo occurs. The threats range from explicit, e.g. a person attacking the truck driver, to implicit, e.g. somebody loitering and then fiddling with the exterior of the truck in order to open it. Our goal is a system that is able to recognize a threat instantaneously as they develop. Typical observables of the threats are a person’s activity, presence in a particular zone and the trajectory. The novelty of this paper is an encoding of these threat observables in a semantic, intermediate-level representation, based on low-level visual features that have no intrinsic semantic meaning themselves. The aim of this representation was to bridge the semantic gap between the low-level tracks and motion and the higher-level notion of threats. In our experiments, we demonstrate that our semantic representation is more descriptive for threat detection than directly using low-level features. We find that a person’s activities are the most important elements of this semantic representation, followed by the person’s trajectory. The proposed threat detection system is very accurate: 96.6 % of the tracks are correctly interpreted, when considering the temporal context.