Visual Task Research Articles

Associations between attention and threat in dental settings

Dental fear still causes health risks and high costs for health insurances due to the common avoidance of dental visits by dentally fearful patients. Selective attention may be a potential trigger of dental fear. The aim of the present studies was to investigate the influence of threatening dental stimuli in image combinations on the performance in a visual search task. In three studies, we compared the influence of dental stimuli with pictures of weapons or accidents and different distractors, i.e., cutlery, animals, landscapes. In the first two studies, participants had to decide if a target is available in a 3 × 3 matrix or not. In the third experiment, participants had to decide if a threatening or a neutral stimulus was the target. In all three studies we could replicate the finding that there is a threat-superiority-effect as well as a disengagement-effect for dental stimuli in visual search tasks. Dental stimuli were not only detected faster than deciding that no dental stimulus is available with different distractors, but dental stimuli also decreased search performance within the 3 × 3 matrices when presented as distractors. This is the first paper that demonstrates that dental stimuli that are associated with a negative threatening valence attract attention very quickly in the context of non-threatening stimuli. Future studies should investigate if a distraction from these threatening stimuli or a habituation to these threatening stimuli could reduce dental fear.

Boundary points guided 3D object detection for point clouds

3D object detection in LiDAR point clouds is crucial for computer vision tasks such as autonomous driving. The two-stage approach with point cloud completion achieves remarkable performance by generating semantic surface points for foreground objects or learning bird’s eye view shape heat map labels. However, these methods require additional completion datasets, leading to substantial computation and memory demands. In this context, we propose a Boundary Points Guided 3D (BPG3D) object detection method that complements point cloud boundary information without the need for additional data. Specifically, we generate Region of Interest (RoI) boundary points to aggregate the neighbor voxel information at the RoI boundary during the refinement stage to complement the missing boundary information. Meanwhile, we design a Dual Feature Selection (DFS) module to adaptively fuse RoI grid point features and RoI boundary point features for bounding box refinement with negligible computational cost. Additionally, inspired by tensor decomposition theory, we use low-rank tensors to reconstruct high-rank tensors in the point cloud feature encoder to enhance contextual semantic information. The proposed method achieves 65.81% mAP on KITTI Test Set, obtaining a good trade-off between accuracy and efficiency.

Patient teacher can impart locality to improve lightweight vision transformer on small dataset

Vision Transformer (ViT) has achieved unprecedented success in vision tasks with the assistance of abundant data. However, the lack of inductive bias in lightweight ViT makes learning locality challenging on small datasets, leading to poor performance. This limitation impedes the application of lightweight ViT in scenarios with limited datasets and computational power. Knowledge Distillation (KD) allows student models to benefit from the teacher model. However, in the progressive learning stage, traditional single-stage KD methods are usually suboptimal for delivering fixed knowledge to the growing student model. To address these issues, we propose a simple yet effective two-stage KD method called Curriculum Information Knowledge Distillation (CIKD) for the first time. Specifically, we incorporate a curriculum learning framework, progressing from easy to difficult, in the KD curriculum. At the first stage, i.e., Attention Locality Imitation (ALI), the student model learns locality from the low-level semantic features of the teacher model through self-attention distillation. Afterward, at the second stage, i.e., Logit Mimicking (LM), the student model learns label information and high-level semantic logit from the teacher model. Without bells and whistles, our approach achieves state-of-the-art results on 8 small-scale datasets with ViT-Tiny (5.0M). Our code and model weights are available at: https://github.com/newLLing/CIKD.

Soft computing-driven infrared and visible image fusion network for security application service

The high-level visual tasks (HLVT)-specific infrared and visible image fusion networks aim to guide the fusion network to focus on specific feature regions. The resulting fused images enhance the performance of HLVT in adverse conditions. However, cross-layer visual task fusion may diminish the fusion network’s adaptability to diverse data. This deficiency results in a lack of robustness and security in the fusion network. Consequently, the fusion network is rendered insufficient to fulfill the requirements of HLVT security applications, including object detection and security systems. To address these challenges, we propose a soft computing-driven infrared and visible image fusion network based on layer separation and dual-stream adversarial learning, referred to as LsDSGFuse. Firstly, we employ a layer segmentation network based on a high-resolution network to create salient object segmentation masks for the original image. This helps mitigate the impact of task-specific guidance on the fusion network learning process, preventing overfitting to HLVT. Secondly, we use the original image after information separation as the input of the generator. The separated network inputs clearly define foreground objects and background environments, aiding the generator in better fusing foreground and background. Lastly, we adopt a no-loading training strategy to balance adversarial learning, prevent discriminator overfitting, and reduce resource waste. Experimental results demonstrate the outstanding performance of our proposed image fusion method in terms of fusion quality, surpassing several representative approaches. Furthermore, experiments in HLVT indicate that our method facilitates HLVT in better responding to potential threats and unknown variables. This improvement significantly enhances HLVT’s security performance.

Auditory cortex encodes lipreading information through spatially distributed activity

Watching a speaker's face improves speech perception accuracy. This benefit is enabled, in part, by implicit lipreading abilities present in the general population. While it is established that lipreading can alter the perception of a heard word, it is unknown how these visual signals are represented in the auditory system or how they interact with auditory speech representations. One influential, but untested, hypothesis is that visual speech modulates the population-coded representations of phonetic and phonemic features in the auditory system. This model is largely supported by data showing that silent lipreading evokes activity in the auditory cortex, but these activations could alternatively reflect general effects of arousal or attention or the encoding of non-linguistic features such as visual timing information. This gap limits our understanding of how vision supports speech perception. To test the hypothesis that the auditory system encodes visual speech information, we acquired functional magnetic resonance imaging (fMRI) data from healthy adults and intracranial recordings from electrodes implanted in patients with epilepsy during auditory and visual speech perception tasks. Across both datasets, linear classifiers successfully decoded the identity of silently lipread words using the spatial pattern of auditory cortex responses. Examining the time course of classification using intracranial recordings, lipread words were classified at earlier time points relative to heard words, suggesting a predictive mechanism for facilitating speech. These results support a model in which the auditory system combines the joint neural distributions evoked by heard and lipread words to generate a more precise estimate of what was said.

Moving spiders do not boost visual search in spider fear

Previous research on attention to fear-relevant stimuli has largely focused on static pictures or drawings, and thus did not consider the potential effect of natural motion. Here, we aimed to investigate the effect of motion on attentional capture in spider-fearful and non-fearful participants by using point-light stimuli and naturalistic videos. Point-light stimuli consist of moving dots representing joints and thereby visualizing biological motion (e.g. of a walking human or cat) without needing a visible body. Spider-fearful (n = 30) and non-spider-fearful (n = 31) participants completed a visual search task with moving targets (point-light/naturalistic videos) and static distractors (images), static targets and moving distractors, or static targets and static distractors. Participants searched for a specified animal type (snakes, spiders, cats, or doves) as quickly as possible. We replicated previous findings with static stimuli: snakes were detected faster and increased distraction, while spiders just increased distraction. However, contrary to our hypotheses, spider targets did not speed up responses, neither in the group of control nor in the group of spider-fearful participants. Interestingly, stimuli-specific effects were toned down, abolished, or even changed direction when motion was introduced. Also, we demonstrated that point-light stimuli were of similar efficiency as naturalistic videos, indicating that for testing effects of motion in visual search, “pure” motion stimuli might be sufficient. As we do show a substantial modulation of visual search phenomena by biological motion, we advocate for future studies to use moving stimuli, equivalent to our dynamic environment, to increase ecological validity.

Infrared and Harsh Light Visible Image Fusion Using an Environmental Light Perception Network.

The complementary combination of emphasizing target objects in infrared images and rich texture details in visible images can effectively enhance the information entropy of fused images, thereby providing substantial assistance for downstream composite high-level vision tasks, such as nighttime vehicle intelligent driving. However, mainstream fusion algorithms lack specific research on the contradiction between the low information entropy and high pixel intensity of visible images under harsh light nighttime road environments. As a result, fusion algorithms that perform well in normal conditions can only produce low information entropy fusion images similar to the information distribution of visible images under harsh light interference. In response to these problems, we designed an image fusion network resilient to harsh light environment interference, incorporating entropy and information theory principles to enhance robustness and information retention. Specifically, an edge feature extraction module was designed to extract key edge features of salient targets to optimize fusion information entropy. Additionally, a harsh light environment aware (HLEA) module was proposed to avoid the decrease in fusion image quality caused by the contradiction between low information entropy and high pixel intensity based on the information distribution characteristics of harsh light visible images. Finally, an edge-guided hierarchical fusion (EGHF) module was designed to achieve robust feature fusion, minimizing irrelevant noise entropy and maximizing useful information entropy. Extensive experiments demonstrate that, compared to other advanced algorithms, the method proposed fusion results contain more useful information and have significant advantages in high-level vision tasks under harsh nighttime lighting conditions.

Refined Land Use Classification for Urban Core Area from Remote Sensing Imagery by the EfficientNetV2 Model

In the context of accelerated urbanization, assessing the quality of the existing built environment plays a crucial role in urban renewal. In the existing research and use of deep learning models, most categories are urban construction areas, forest land, farmland, and other categories. These categories are not conducive to a more accurate analysis of the spatial distribution characteristics of urban green space, parking space, blue space, and square. A small sample of refined land use classification data for urban built-up areas was produced using remote sensing images. The large-scale remote sensing images were classified using deep learning models, with the objective of inferring the fine land category of each tile image. In this study, satellite remote sensing images of four cities, Handan, Shijiazhuang, Xingtai, and Tangshan, were acquired by Google Class 19 RGB three-channel satellite remote sensing images to establish a data set containing fourteen urban land use classifications. The convolutional neural network model EfficientNetV2 is used to train and validate the network framework that performs well on computer vision tasks and enables intelligent image classification of urban remote sensing images. The model classification effect is compared and analyzed through accuracy, precision, recall, and F1-score. The results show that the EfficientNetV2 model has a classification recognition accuracy of 84.56% on the constructed data set. The testing set accuracy increases sequentially after transfer learning. This paper verifies that the proposed research framework has good practicality and that the results of the land use classification are conducive to the fine-grained quantitative analysis of built-up environmental quality.

The link between executive skills and neural dynamics during encoding, inhibition, and retrieval of visual information in the elderly.

During aging the inter-individual variability in both the neural and behavioral functions is likely to be emphasized. Decreased competence particularly in working memory and general executive control compromises many aspects of the quality of life also within the nonclinical population. We aimed, first, to clarify the brain basis of visual working memory and inhibition during multi-stage natural-like task performance, and second, to identify associations between variation in task-related neural activity and relevant cognitive skills, namely inhibition and general working memory capacity. We recorded, using magnetoencephalography (MEG), the neural modulations associated with encoding, maintenance, and retrieval, as well as interference suppression during a visual working memory task in older adults. We quantified the neural correlates of these cognitive processes through two complementary approaches: evoked responses and oscillatory activity. Neural activity during memory retrieval and interference suppression were correlated with behavioral measures of task switching and general executive functions. Our results show that general inhibitory control induced frontocentral neural modulation across a broad range of frequencies whereas domain-specific inhibition was limited to right posterior areas. Our findings also suggest that modulations particularly in phase-locked evoked neural activity can be reliably associated with explicit measures of cognitive skills, with better inhibitory control linked with an early neural effect of distractor inhibition during retrieval. In general, we show that exploiting the inherent inter-individual variability in neural measures and behavioral markers of cognition in aging populations can help establish reliable links between specific brain functions and their behavioral manifestations.

Group efficiency based on the termination rule in the multiple-targets visual search task.

In practical visual search fields, observers often encounter errors that result from an unknown number of targets, which may induce reduced accuracy and speed. Our current study addresses the potential enhancement of collaborative search efficiency as a dyad to mitigate such incurred search costs. Utilizing the capacity coefficient, we evaluated search efficiency and explored the interplay of task difficulty and termination rule in collaborative visual search. Our prediction that collaborative benefits increased with elevated task difficulty was not supported in Experiment 1, where participants were tasked with judging the presence of any target. In contrast, Experiment 2 demonstrated that dyads exhibited greater search efficiency during exhaustive searches for multiple targets with elevated task difficulty. Notably, our findings indicated an advantage in dyad searches compared to baseline predictions from individual searches. Our results underscored the significance of task difficulty and termination rules in leveraging human resources for improved collaborative visual search performance.

Deciphering the impairment of perimenopausal insomnia on visual search from a neurocognitive processing perspective.

Perimenopausal insomnia (PMI) is associated with observable performance impairments in visual search tasks. This study examines how various cognitive processing stages contribute to search performance delays in PMI compared to healthy controls (HCs). We recruited 76 participants diagnosed with PMI and 63 HCs. Event-related potentials (ERPs) were recorded as participants engaged in a visual search task, reporting the orientation of a color popout target within an array of ellipses. We analyzed group differences in behavioral performance and ERP components across cognitive processing stages. Compared to HCs, PMI patients exhibited behavioral response delays, although accuracy was not different between groups. Electrophysiological analyses revealed group differences across several ERP components. Firstly, the N1 component's amplitude increased bilaterally, suggesting enhanced visual sensory processing. Secondly, a slower and smaller N2pc indicated reduced attentional orienting. Thirdly, a decreased SPCN amplitude pointed to deficits in target discrimination. Fourthly, an increased amplitude of the stimulus-locked LRP, with unchanged latency, suggested heightened neural inputs for maintaining motor initiation speed. Fifthly, prolonged response-locked LRP latency indicated slower motor execution. Finally, these changes in ERP components, along with significant correlations between LRP components and insomnia symptoms, suggest potential neural biomarkers for PMI. Our findings provide high-temporal-resolution insights into the neurocognitive disruptions associated with PMI, highlighting how sleep disturbances affect cognitive processing in visual tasks. These insights enhance our understanding of PMI and contribute to discussions on neural mechanisms driving behavioral performance in various conditions.

Geo-Hgan: Unsupervised anomaly detection in geochemical data via latent space learning

Reconstructing geochemical data for anomaly detection using Generative Adversarial Networks (GANs) has become a prevalent method in identifying geochemical anomalies. However, injecting random noise into GANs can induce model instability. To mitigate this issue, we propose a novel anomaly detection model, Geo-Hgan, which integrates a dual adversarial network architecture with a Latent Space Adversarial Module (LSAM) to learn the distribution of latent variables from arbitrary data and optimize the sample reconstruction process, thereby alleviating instability during GAN training. Additionally, an encoder guided by the LSAM-pretrained GAN is employed to extract variational features, facilitating rapid and effective sample mapping into the latent space defined by LSAM. Experimental results demonstrate that under unsupervised conditions, Geo-Hgan achieves an Area Under the Curve (AUC) score of 85% across three geochemical datasets, outperforming similar models in accuracy and reconstruction capabilities. To assess its versatility and generalization ability, we extend Geo-Hgan to anomaly detection tasks in computer vision, where it achieves an average AUC score of 98.7% on the MvtecAD dataset, setting a new state-of-the-art performance in the domain. Furthermore, we propose AnomFilter, a method for setting anomaly thresholds based on the clustering hypothesis. AnomFilter identifies high-confidence anomaly samples identified by Geo-Hgan in the source domain and iteratively transfers them to the target domain. These high-confidence anomaly samples, combined with a small number of known positive samples in the target domain, enhance the accuracy of supervised geochemical anomaly detection in the target domain, which achieved an AUC score of 94%. The utilization of anomaly detection models for sample transfer learning offers a novel perspective for future work.

Effectiveness of Noise Cancelling Earbuds in Reducing Hearing and Auditory Attention Deficits in Children with Autism.

Background/Objectives: Autism spectrum disorder (ASD) is a lifelong neurodevelopmental condition characterised by impairments in social communication, sensory abnormalities, and attentional deficits. Children with ASD often face significant challenges with speech perception and auditory attention, particularly in noisy environments. This study aimed to assess the effectiveness of noise cancelling Bluetooth earbuds (Nuheara IQbuds Boost) in improving speech perception and auditory attention in children with ASD. Methods: Thirteen children aged 6-13 years diagnosed with ASD participated. Pure tone audiometry confirmed normal hearing levels. Speech perception in noise was measured using the Consonant-Nucleus-Consonant-Word test, and auditory/visual attention was evaluated via the Integrated Visual and Auditory Continuous Performance Task. Participants completed these assessments both with and without the IQbuds in situ. A two-week device trial evaluated classroom listening and communication improvements using the Listening Inventory for Education-Revised (teacher version) questionnaire. Results: Speech perception in noise was significantly poorer for the ASD group compared to typically developing peers and did not change with the IQbuds. Auditory attention, however, significantly improved when the children were using the earbuds. Additionally, classroom listening and communication improved significantly after the two-week device trial. Conclusions: While the noise cancelling earbuds did not enhance speech perception in noise for children with ASD, they significantly improved auditory attention and classroom listening behaviours. These findings suggest that Bluetooth earbuds could be a viable alternative to remote microphone systems for enhancing auditory attention in children with ASD, offering benefits in classroom settings and potentially minimising the stigma associated with traditional assistive listening devices.

Spatial attention guided cGAN for improved salient object detection

Recent research shows that Conditional Generative Adversarial Networks (cGANs) are effective for Salient Object Detection (SOD), a challenging computer vision task that mimics the way human vision focuses on important parts of an image. However, implementing cGANs for this task has presented several complexities, including instability during training with skip connections, weak generators, and difficulty in capturing context information for challenging images. These challenges are particularly evident when dealing with input images containing small salient objects against complex backgrounds, underscoring the need for careful design and tuning of cGANs to ensure accurate segmentation and detection of salient objects. To address these issues, we propose an innovative method for SOD using a cGAN framework. Our method utilizes encoder-decoder framework as the generator component for cGAN, enhancing the feature extraction process and facilitating accurate segmentation of the salient objects. We incorporate Wasserstein-1 distance within the cGAN training process to improve the accuracy of finding the salient objects and stabilize the training process. Additionally, our enhanced model efficiently captures intricate saliency cues by leveraging the spatial attention gate with global average pooling and regularization. The introduction of global average pooling layers in the encoder and decoder paths enhances the network's global perception and fine-grained detail capture, while the channel attention mechanism, facilitated by dense layers, dynamically modulates feature maps to amplify saliency cues. The generated saliency maps are evaluated by the discriminator for authenticity and gives feedback to enhance the generator's ability to generate high-resolution saliency maps. By iteratively training the discriminator and generator networks, the model achieves improved results in finding the salient object. We trained and validated our model using large-scale benchmark datasets commonly used for salient object detection, namely DUTS, ECSSD, and DUT-OMRON. Our approach was evaluated using standard performance metrics on these datasets. Precision, recall, MAE and Fβ score metrics are used to evaluate performance. Our method achieved the lowest MAE values: 0.0292 on the ECSSD dataset, 0.033 on the DUTS-TE dataset, and 0.0439 on the challenging and complex DUT-OMRON dataset, compared to other state-of-the-art methods. Our proposed method demonstrates significant improvements in salient object detection, highlighting its potential benefits for real-life applications.

Serial visual reversal learning in captive black-handed spider monkeys, Ateles geoffroyi

Recent research suggests that socio-ecological factors such as dietary specialization and social complexity may be drivers of advanced cognitive skills among primates. Therefore, we assessed the ability of 12 black-handed spider monkeys (Ateles geoffroyi), a highly frugivorous platyrrhine primate with strong fission-fusion dynamics, to succeed in a serial visual reversal learning task. Using a two-alternative choice paradigm we first trained the animals to reliably choose a rewarded visual stimulus over a non-rewarded one. Upon reaching a pre-set learning criterion we then switched the reward values of the two stimuli and assessed if and how quickly the animals learned to reverse their choices, again to a pre-set learning criterion. This stimulus reversal procedure was then continued for a total of 80 sessions of 10 trials each. We found that the spider monkeys quickly learned to reliably discriminate between two simultaneously presented visual stimuli, that they succeeded in a visual reversal learning task, and that they displayed an increase in learning speed across consecutive reversals, suggesting that they are capable of serial reversal learning-set formation with visual cues. The fastest-learning individual completed five reversals within the 80 sessions. The spider monkeys outperformed most other primate and nonprimate mammal species tested so far on this type of cognitive task, including chimpanzees, with regard to their learning speed in both the initial learning task and in the first reversal task, suggesting a high degree of behavioral flexibility and inhibitory control. Our findings support the notion that socio-ecological factors such as dietary specialization and social complexity foster advanced cognitive skills in primates.

Relating distractor suppression to problematic drinking behavior

BackgroundImpaired cognitive control has been linked to weakened self-regulatory processes underlying compulsive substance intake. Previous research has provided evidence for impaired task performance in substance-abusing groups during Stroop and Go/No-Go tasks. Mechanisms of distractor suppression in visual search might also involve overlapping regulatory components that support goal-directed behavior by resolving the attentional competition between distractors and the target of search. However, the efficiency of learning-dependent distractor suppression has not been examined in the context of drug abuse and a direct comparison between cognitive control and distractor suppression is lacking. MethodA total of 84 participants were assigned either to the heavy drinking group (ALC, n = 42) or the control group (CTL, n = 42) based on self-reported substance use. Participants completed the Alcohol Use Disorders Identification Test (AUDIT) and Barratt Impulsiveness Scale (BIS). After that, participants completed a computerized version of the Stroop task, Go/No-go task, and a visual search task measuring learning-dependent distractor suppression. ResultsThe Stroop effect and the frequency of no-go errors did not differ between groups. However, learned distractor suppression was significantly blunted in the ALC group compared to the control group. Across participants, performance on the Stroop and Go/No-go task were correlated, while the magnitude of distractor suppression was related to neither. ConclusionsOur findings support a divergence of mechanistic processes underlying cognitive control and attentional control, and demonstrate impaired learning-dependent distractor suppression in heavy drinkers relative to a control group. Impaired distractor suppression offers new insight into why drug cues can be difficult to ignore.

Visual selective attention in individuals with age-related hearing loss

Evidence from epidemiological studies suggests that hearing loss is associated with an accelerated decline in cognitive function, but the underlying pathophysiological mechanism remains poorly understood. Studies using auditory tasks have suggested that degraded auditory input increases the cognitive load for auditory perceptual processing and thereby reduces the resources available for other cognitive tasks. Attention-related networks are among the systems overrecruited to support degraded auditory perception, but it is unclear how they function when no excessive recruitment of cognitive resources for auditory processing is needed. Here, we implemented an EEG study using a nonauditory visual attentional selection task in 30 individuals with age-related hearing loss (ARHLs, 60–73 years) and compared them with aged (N = 30, 60–70 years) and young (N = 35, 22–29 years) normal-hearing controls. Compared with their normal-hearing peers, ARHLs demonstrated a significant amplitude reduction for the posterior contralateral N2 component, which is a well-validated index of the allocation of selective visual attention, despite the comparable behavioral performance. Furthermore, the amplitudes were observed to correlate significantly with hearing acuities (pure tone audiometry thresholds) and higher-order hearing abilities (speech-in-noise thresholds) in aged individuals. The target-elicited alpha lateralization, another mechanism of visuospatial attention, demonstrated in control groups was not observed in ARHLs. Although behavioral performance is comparable, the significant decrease in N2pc amplitude in ARHLs provides neurophysiologic evidence that may suggest a visual attentional deficit in ARHLs even without extra-recruitment of cognitive resources by auditory processing. It supports the hypothesis that constant degraded auditory input in ARHLs has an adverse impact on the function of cognitive control systems, which is a possible mechanism mediating the relationship between hearing loss and cognitive decline.

Conditional learning abilities in Down syndrome and Williams syndrome

ABSTRACT We investigated whether individuals with Down (DS) or Williams (WS) syndrome are capable of conditional learning, a hippocampus-dependent process enabling flexible reasoning, using three touch-screen tasks: a visual learning task to assess unconditional learning (A > B and C > D), a 3-item conditional learning task (A > B, B > C), and a 5-item conditional learning task (A > B, B > C, C > D, D > E) to further assess transitive inference abilities. We compared the performance of individuals with DS or WS to typically developing 3–12-year-old children and 20–30-year-old adults. Most participants exhibited unconditional learning, but only 12% of individuals with DS, 15% of individuals with WS, and 42% of children exhibited conditional learning, as compared to 95% of young adults; only adults demonstrated transitive inference. Our findings indicate that conditional learning and transitive inference, which depend on brain structures impacted in these disorders, including the hippocampus and the prefrontal cortex, are beyond reach for most individuals with DS and WS.

Dynamic Feature Focusing Network for small object detection

Deep learning has driven research in object detection and achieved proud results. Despite its significant advancements in object detection, small object detection still struggles with low recognition rates and inaccurate positioning, primarily attributable to their miniature size. The location deviation of small objects induces severe feature misalignment, and the disequilibrium between classification and regression tasks hinders accurate recognition. To address these issues, we propose a Dynamic Feature Focusing Network (DFFN), which contains a duo of crucial modules: Visual Perception Enhancement Module (VPEM) and Task Association Module (TAM). Drawing upon the deformable convolution and attention mechanism, the VPEM concentrates on sparse key features and perceives the misalignment via positional offset. We aggregate multi-level features at identical spatial locations via layer average operation for learning a more discriminative representation. Incorporating class alignment and bounding box alignment parts, the TAM promotes classification ability, refines bounding box regression, and facilitates the joint learning of classification and localization. We conduct diverse experiments, and the proposed method considerably enhances the small object detection performance on four benchmark datasets of MS COCO, VisDrone, VOC, and TinyPerson. Our method has improved by 3.4 and 2.2 in mAP and APs, making solid improvements on COCO. Compared to other classic detection models, DFFN exhibits a high level of competitiveness in precision.

Prestimulus Alpha Phase Modulates Visual Temporal Integration.

When presented shortly after another, discrete pictures are naturally perceived as continuous. The neuronal mechanism underlying such continuous or discrete perception is not well understood. While continuous alpha oscillations are a candidate for orchestrating such neuronal mechanisms, recent evidence is mixed. In this study, we investigated the influence of prestimulus alpha oscillation on visual temporal perception. Specifically, we were interested in whether prestimulus alpha phase modulates neuronal and perceptual processes underlying discrete or continuous perception. Participants had to report the location of a missing object in a visual temporal integration task, while simultaneously MEG data were recorded. Using source reconstruction, we evaluated local phase effects by contrasting phase angle values between correctly and incorrectly integrated trials. Our results show a phase opposition cluster between -0.8 and -0.5 s (relative to stimulus presentation) and between 6 and 20 Hz. These momentary phase angle values were correlated with behavioral performance and event-related potential amplitude. There was no evidence that frequency defined a window of temporal integration.