Spatial Attention Model Research Articles

Foreign object detection in urban rail transit based on deep differentiation segmentation neural network

With the increasing scale of urban rail transit, foreign object intrusion has become a significant operational safety hazard in urban rail transit. Although the laser-based automatic foreign object detection system has advantages such as long-distance detection and insensitivity to light changes, it has drawbacks such as large blind spots and low visualization. In response to the problems existing in laser detection systems, we proposed a novel video-based deep differentiation segmentation neural network for foreign object detection. Firstly, the foreign object detection is transformed into a binary classification problem, and the foreign object is determined as the image's foreground using image segmentation principles. Secondly, build a deep segmentation network based on deep convolution. Finally, perform morphological operations and threshold judgment on the foreground segmentation image to filter out the final detection results. To improve the detection effect, we reduced the impact of airflow disturbance by sampling and calculating the average background image. At the same time, the channel attention model and spatial attention model are added to the deep differentiation neural network. Collecting real data on subway platforms for experiments shows that the proposed method has a detection accuracy of 95.8 %, which is superior to traditional detection methods and recent image segmentation neural networks.

DaSAM: Disease and Spatial Attention Module-Based Explainable Model for Brain Tumor Detection

Brain tumors are the result of irregular development of cells. It is a major cause of adult demise worldwide. Several deaths can be avoided with early brain tumor detection. Magnetic resonance imaging (MRI) for earlier brain tumor diagnosis may improve the chance of survival for patients. The most common method of diagnosing brain tumors is MRI. The improved visibility of malignancies in MRI makes therapy easier. The diagnosis and treatment of brain cancers depend on their identification and treatment. Numerous deep learning models are proposed over the last decade including Alexnet, VGG, Inception, ResNet, DenseNet, etc. All these models are trained on a huge dataset, ImageNet. These general models have many parameters, which become irrelevant when implementing these models for a specific problem. This study uses a custom deep-learning model for the classification of brain MRIs. The proposed Disease and Spatial Attention Model (DaSAM) has two modules; (a) the Disease Attention Module (DAM), to distinguish between disease and non-disease regions of an image, and (b) the Spatial Attention Module (SAM), to extract important features. The experiments of the proposed model are conducted on two multi-class datasets that are publicly available, the Figshare and Kaggle datasets, where it achieves precision values of 99% and 96%, respectively. The proposed model is also tested using cross-dataset validation, where it achieved 85% accuracy when trained on the Figshare dataset and validated on the Kaggle dataset. The incorporation of DAM and SAM modules enabled the functionality of feature mapping, which proved to be useful for the highlighting of important features during the decision-making process of the model.

Top-Down Priors Disambiguate Target and Distractor Features in Simulated Covert Visual Search.

Several models of visual search consider visual attention as part of a perceptual inference process, in which top-down priors disambiguate bottom-up sensory information. Many of these models have focused on gaze behavior, but there are relatively fewer models of covert spatial attention, in which attention is directed to a peripheral location in visual space without a shift in gaze direction. Here, we propose a biologically plausible model of covert attention during visual search that helps to bridge the gap between Bayesian modeling and neurophysiological modeling by using (1) top-down priors over target features that are acquired through Hebbian learning, and (2) spatial resampling of modeled cortical receptive fields to enhance local spatial resolution of image representations for downstream target classification. By training a simple generative model using a Hebbian update rule, top-down priors for target features naturally emerge without the need for hand-tuned or predetermined priors. Furthermore, the implementation of covert spatial attention in our model is based on a known neurobiological mechanism, providing a plausible process through which Bayesian priors could locally enhance the spatial resolution of image representations. We validate this model during simulated visual search for handwritten digits among nondigit distractors, demonstrating that top-down priors improve accuracy for estimation of target location and classification, relative to bottom-up signals alone. Our results support previous reports in the literature that demonstrated beneficial effects of top-down priors on visual search performance, while extending this literature to incorporate known neural mechanisms of covert spatial attention.

Revolutionizing anemia detection: integrative machine learning models and advanced attention mechanisms

This study addresses the critical issue of anemia detection using machine learning (ML) techniques. Although a widespread blood disorder with significant health implications, anemia often remains undetected. This necessitates timely and efficient diagnostic methods, as traditional approaches that rely on manual assessment are time-consuming and subjective. The present study explored the application of ML – particularly classification models, such as logistic regression, decision trees, random forest, support vector machines, Naïve Bayes, and k-nearest neighbors – in conjunction with innovative models incorporating attention modules and spatial attention to detect anemia. The proposed models demonstrated promising results, achieving high accuracy, precision, recall, and F1 scores for both textual and image datasets. In addition, an integrated approach that combines textual and image data was found to outperform the individual modalities. Specifically, the proposed AlexNet Multiple Spatial Attention model achieved an exceptional accuracy of 99.58%, emphasizing its potential to revolutionize automated anemia detection. The results of ablation studies confirm the significance of key components – including the blue-green-red, multiple, and spatial attentions – in enhancing model performance. Overall, this study presents a comprehensive and innovative framework for noninvasive anemia detection, contributing valuable insights to the field.

The disengagement deficit after right-hemisphere damage: Distinct roles of lateral frontal and parietal damage

An influential model of spatial attention postulates three main attention-orienting mechanisms: disengagement, shifting, and engagement. Early research linked disengagement deficits with superior parietal damage, regardless of hemisphere or presence of spatial neglect. Subsequent studies supported the involvement of more ventral parietal regions, especially in the right hemisphere, and linked spatial neglect to deficient disengagement from ipsilateral cues. However, previous lesion studies faced serious limitations, such as small sample sizes and the lack of brain-injured controls without neglect. Additionally, some studies employed symbolic cues or used long cue-target intervals, which may fail to reveal impaired disengagement. We here used a machine-learning approach to conduct lesion-symptom mapping (LSM) on 89 patients with focal cerebral lesions to the left (LH) or right (RH) cerebral hemisphere. A group of 54 healthy participants served as controls. The paradigm used to uncover disengagement deficits employed non-predictive cues presented in the visual periphery and at short cue-target intervals, targeting exogenous attention. The main factors of interest were group (healthy participants, LH, RH), target position (left, right hemifield) and cue validity (valid, invalid). LSM-analyses were performed on two indices: the validity effect, computed as the absolute difference between reaction times (RTs) following invalid compared to valid cues, and the disengagement deficit, determined by the difference between contralesional and ipsilesional validity effects. While LH patients showed general slowing of RTs to contralesional targets, only RH patients exhibited a disengagement deficit from ipsilesional cues. LSM associated the validity effect with a right lateral frontal cluster, which additionally affected subcortical white matter of the right arcuate fasciculus, the corticothalamic pathway, and the superior longitudinal fasciculus. In contrast, the disengagement deficit was related to damage involving the right temporoparietal junction. Thus, our results support the crucial role of right inferior parietal and posterior temporal regions for attentional disengagement, but also emphasize the importance of lateral frontal regions, for the reorienting of attention.

A multiple directions turning vehicle counting method at intersections based on arbitrary-oriented detection and stack Long Short-Term Memory

The traffic flow parameters of turning vehicles at urban road intersections play an important role in the intelligent traffic system. Traffic road monitoring is difficult to cover large intersection scenes. With the application of unmanned aerial vehicles in traffic, drone-based intersection monitoring has great potential due to the large perspective. But there is rarely research on multiple turning vehicle counting at intersections based on drones. A few existing multiple turning vehicle counting methods are based on vehicle detection and tracking methods. However, the vehicles at the intersection have various orientations, the accurate information of the vehicle cannot be obtained by detection, and the tracking process is complicated and redundant for the counting task. Based on the traffic video at intersections taken by unmanned aerial vehicles, we propose a eight turnings vehicle counting model without complex tracking. Firstly, a spatial attention and channel adaptive attention model net is proposed for arbitrary-oriented vehicle detection to get the orientation and position of vehicles. Secondly, a turning spatio-temporal counting feature for different turning vehicles and its extraction method are proposed. Finally, the stack Long Short-Term Memory net based counting model is designed to process the turning spatio-temporal counting feature and counting eight different turnings vehicles at intersection. The experiments show that the proposed method can realize counting eight different turnings vehicles simultaneously in drone-based traffic video at the intersection without relying on complex multi-target tracking. The average counting accuracy has reached 98.18%.

Putting the Spotlight Back Onto the Flanker Task in Autism: Autistic Adults Show Increased Interference from Foils Compared with Non-autistic Adults.

Autistic people may have a less focused spotlight of spatial selective attention than non-autistic people, meaning that distracting stimuli are less effectively suppressed. Previous studies using the flanker task have supported this suggestion with observations of increased congruency effects in autistic participants. However, findings across studies have been mixed, mainly based on research in children and on response time measures, which may be influenced by differences in response strategy between autistic and non-autistic people rather than differences in selective attention. In this pre-registered study, 153 autistic and 147 non-autistic adults completed an online flanker task. The aims of this study were to test whether increased congruency effects replicate in autistic adults and to extend previous work by fitting a computational model of spatial selective attention on the flanker task to the data. Congruency effects were increased in the autistic group. The modelling revealed that the interference time from the foils was increased in the autistic group. This suggests that the activation of the foils was increased, meaning suppression was less effective for autistic participants. There were also differences in non-interference parameters between the groups. The estimate of response caution was increased in the autistic group and the estimate of perceptual efficiency was decreased. Together these findings suggest inefficient suppression, response strategy and perceptual processing all contribute to differences in performance on the flanker task between autistic and non-autistic people.

Efficient multiscale spatial attention 3D abdominal multiorgan segmentation model

AbstractAccurate and efficient 3D medical image segmentation models are critical for clinical applications. However, many models which performed well in recent years have consumed significant computational cost, especially those based on self‐attention. For more efficient and accurate segmentation of medical 3D images, we develop an efficient multiscale spatial attention model for 3D abdominal multiorgan segmentation. We use 1D and 2D kernels in the encoder and decoder modules to simulate 3D kernels to reduce the number of parameters and computational cost. Furthermore, we use the lightweight spatial attention and mixed pooling module to capture cross‐dimensional information, multiscale features, and long‐range contextual information. We achieved a Dice Similarity Coefficient (DSC) of 89.86 and a Normalized Surface Dice (NSD) of 78.2 on the FLARE dataset with only 1183 M GPU memory usage, 9 M parameters, and 193G floating point operations(FLOPs).

Breast Tumour Segmentation Using Advanced UNet with Saliency, Channel, and Spatial Attention Models

Cancers are getting pretty common these days and in that the second most common cancer in the world after lung cancer is breast cancer. The primary screening techniques for early diagnosis of cancerous nodules in women breast are Ultrasound, Mammography, and MRI analysis. However, accurately identifying and outlining the boundaries of tumor regions remains challenging due to their unpredictable shapes, random variations, and blurry outlines. Currently, the CAD based image analysis systems have gained huge attention in biomedical image processing systems along with Machine Learning (ML) and Deep Learning (DL) methods. The conventional machine learning methods suffer from accuracy related issues therefore deep learning-based schemes have been adopted. Specifically, UNet based approach is widely employed for biomedical image segmentation. This research focuses on development of a UNet based deep learning architecture for breast cancer image analysis and segmentation. The proposed model improves the segmentation accuracy by adding saliency model, channel and spatial attention model to consider the low- and high-level features. These modules help to achieve the clear boundary during segmentation. The performance of proposed model is evaluated by using publicly available mammogram dataset known as DDSM dataset.

Unified Spatio-Temporal Attention Models for Advanced Human Action Recognition & Detection

Abstract: This research paper introduces innovative approaches to enhance human action analytics in computer vision through the development of spatial and temporal attention models. The first model is based on Recurrent Neural Networks (RNNs) with Long Short-Term Memory (LSTM) units, aiming to extract discriminative spatial and temporal features from skeleton data for improved human action recognition and detection. The model selectively focuses on discriminative joints within each frame, utilizing a regularized cross-entropy loss for effective training. Additionally, a joint training strategy is proposed to optimize the learning process. The second model, a spatio-temporal attention (STA) network, is designed to address the limitations of existing 3D Convolutional Neural Networks (3D CNNs) in treating all video frames equally. The STA network characterizes beneficial information at both the frame and channel levels, leveraging differences in spatial and temporal dimensions to enhance the learning capability of 3D convolutions. The proposed STA method can be seamlessly integrated into state-of-the-art 3D CNN architectures for video action detection and recognition. Evaluation of diverse datasets, including SBU Kinect Interaction, NTU RGB+D, PKU-MMD, UCF-101, HMDB-51, and THUMOS 2014, demonstrates the effectiveness of both proposed models in achieving state-of-the-art performance for action recognition and detection tasks. The spatial and temporal attention models contribute significantly to capturing discriminative features, showcasing their potential in advancing the field of human action analytics in computer vision.

ResCCFusion: Infrared and visible image fusion network based on ResCC module and spatial criss-cross attention models

We proposed an infrared and visible image fusion method based on the ResCC module and spatial criss-cross attention models. The proposed method adopts an auto-encoder structure consisting of an encoder network, fusion layers, and a decoder network. The encoder network has a convolution layer and three ResCC blocks with dense connections. Each ResCC block can extract multi-scale features from source images without down-sampling operations and retain as many feature details as possible for image fusion. The fusion layer adopts spatial criss-cross attention models, which can capture contextual information in both horizontal and vertical directions. Attention in these two directions can also reduce the calculation of the attention maps. The decoder network consists of four convolution layers designed to reconstruct images from the feature map. Experiments performed on the public datasets demonstrate that the proposed method obtains better fusion performance on objective and subjective evaluations compared to other advanced fusion methods. The code is available at https://github.com/xiongzhangzzz/ResCCFusion.

Deep learning and thermographic imaging method for thermal comfort prediction in different genders

With the rising demand for electrification and intelligent functions in automobiles, the interest in passenger-centric air conditioning thermal management systems has swiftly increased. However, predicting personalized passenger comfort and applicability continues to pose challenges. The main challenge is how to use valid data to accurately predict the individual needs and preferences of different passengers. In this paper, we developed a deep learning model for a passenger-centric air conditioning system using ResNet50 (a deep residual network architecture consisting of 50 neural network layers). A spatial attention model was also incorporated (Attention-ResNet50) to enhance thermal comfort prediction accuracy. Experiments were conducted to collect basic image data under various thermal comfort and environmental conditions. Infrared images were gathered and classified into cold, comfortable, and warm categories for training, validation, and testing. The accuracy of thermal comfort prediction on testing sets is approximately 5 % higher when incorporating the spatial attention mechanism module compared to the standard ResNet50 model. Thermal comfort tests and predictions were also conducted for different genders. The female thermal comfort prediction model demonstrated significantly higher accuracy levels, nearing 100 % in situations characterized by cold or warm scenarios. The higher accuracy of female comfort prediction results, verifying that female facial features are more sensitive to temperature variation in their surroundings.

Invited Session I: 30 years of normalization in the visual system: Testing and expanding the Reynolds & Heeger Normalization Model of Attention.

I present evidence supporting key predictions of the Reynolds-Heeger Normalization Model of Attention: 1) Both exogenous and endogenous covert spatial attention, can affect performance by contrast or response gain changes, depending on the stimulus size and the relative size of the attention field. 2) Feature-based attention improves performance, consistent with a change in response gain, when the featural extent of the attention field is small (low uncertainty) or large (high uncertainty) relative to the featural extent of the stimulus. 3) An expanded model of spatial resolution and attention that distinguishes between exogenous and endogenous attention, focusing on texture-based segmentation as a model system, which has revealed a dissociation between these types of attention across eccentricity. Our model emulates sensory encoding to segment figures from their background and predict performance. To explain attentional effects, exogenous and endogenous attention require separate operating regimes across spatial frequency. The model reproduces behavioral performance across several experiments and resolves three unexplained phenomena: 1) the parafoveal advantage in segmentation, 2) the uniform improvements across eccentricity by endogenous attention, and 3) the peripheral improvements and central impairments by exogenous attention. This model provides a generalizable framework for predicting effects of endogenous and exogenous attention on perception across the visual field.

Deep Learning Neural Network-Based Detection of Wafer Marking Character Recognition in Complex Backgrounds

Wafer characters are used to record the transfer of important information in industrial production and inspection. Wafer character recognition is usually used in the traditional template matching method. However, the accuracy and robustness of the template matching method for detecting complex images are low, which affects production efficiency. An improved model based on YOLO v7-Tiny is proposed for wafer character recognition in complex backgrounds to enhance detection accuracy. In order to improve the robustness of the detection system, the images required for model training and testing are augmented by brightness, rotation, blurring, and cropping. Several improvements were adopted in the improved YOLO model, including an optimized spatial channel attention model (CBAM-L) for better feature extraction capability, improved neck structure based on BiFPN to enhance the feature fusion capability, and the addition of angle parameter to adapt to tilted character detection. The experimental results showed that the model had a value of 99.44% for mAP@0.5 and an F1 score of 0.97. In addition, the proposed model with very few parameters was suitable for embedded industrial devices with small memory, which was crucial for reducing the hardware cost. The results showed that the comprehensive performance of the improved model was better than several existing state-of-the-art detection models.

Hyperspectral Images Weakly Supervised Classification with Noisy Labels

The deep network model relies on sufficient training samples to achieve superior processing performance, which limits its application in hyperspectral image (HSI) classification. In order to perform HSI classification with noisy labels, a robust weakly supervised feature learning (WSFL) architecture combined with multi-model attention is proposed. Specifically, the input noisy labeled data are first subjected to multiple groups of residual spectral attention models and multi-granularity residual spatial attention models, enabling WSFL to refine and optimize the extracted spectral and spatial features, with a focus on extracting clean samples information and reducing the model’s dependence on labels. Finally, the fused and optimized spectral-spatial features are mapped to the multilayer perceptron (MLP) classifier to increase the constraint of the model on the noisy samples. The experimental results on public datasets, including Pavia Center, WHU-Hi LongKou, and HangZhou, show that WSFL is better at classifying noise labels than excellent models such as spectral-spatial residual network (SSRN) and dual channel residual network (DCRN). On Hangzhou dataset, the classification accuracy of WSFL is superior to DCRN by 6.02% and SSRN by 7.85%, respectively.

EDSV-Net: An efficient defect segmentation network based on visual attention and visual perception

In industrial production, surface defect detection algorithms based on convolutional neural networks have been widely studied to improve production quality. However, for practical applications, there are still many issues to be solved, such as the complexity and diversity of defect categories, the difficulty of obtaining defect samples, and the difficulty of existing algorithms in accurately segmenting defects. To solve these issues, we present an effective defect segmentation network based on visual attention and visual perception termed EDSV-Net. Specifically, we use ResNet18 as the backbone network in EDSV-Net. Then a multi-scale feature extraction (MSFE) module is introduced to enhance the scale invariance of high-level features and the diversity of contextual features. In addition, a spatial attention (SA) model combined with a channel attention (CA) model is applied to low level features and MSFE features, respectively, to extract more effective spatial and semantic information. Moreover, a depthwise separable convolution is introduced to reduce the network complexity. Finally, due to the issues of existing defect detection algorithms ignoring structural similarity and defects being difficult to obtain, we design a balanced defect and structural measure loss function. Meanwhile, we propose a structural similarity measure, which combines the pixel similarity for evaluation. EDSV-Net only requires no more than 60 random abnormal samples to obtain accurate segmentation results and the real-time performance meets the requirements of actual industrial production. Based on three challenging real-world defect datasets, the results of the evaluation demonstrate that EDSV-Net outperforms seven state-of-the-art methods on accuracy and real-time performance.

Spatial-Semantic Transformer for Spatial Relation Recognition

Spatial relation recognition, which aims to predict a spatial relation predicate, has attracted increasing attention in the computer vision study. During tackling this problem, modeling spatial relation of the subjects and objects is of great importance. We find that only using spatial features leads to poor results in predicting the spatial relation. To overcome these challenges, we propose an effective spatial attention module to enhance spatial features using semantic features. After identifying the importance of spatial attention mechanism, we propose a spatial transformer module with encoder layers to recognize unseen spatial relation based on spatial attention mechanism. Extensive experiments on the benchmark dataset (SpatialSense) show that, by using refined spatial feature, our spatial transformer model and spatial attention model achieve state-of-the-art performance on overall accuracy.

CCS-UNet: a cross-channel spatial attention model for accurate retinal vessel segmentation.

Precise segmentation of retinal vessels plays an important role in computer-assisted diagnosis. Deep learning models have been applied to retinal vessel segmentation, but the efficacy is limited by the significant scale variation of vascular structures and the intricate background of retinal images. This paper supposes a cross-channel spatial attention U-Net (CCS-UNet) for accurate retinal vessel segmentation. In comparison to other models based on U-Net, our model employes a ResNeSt block for the encoder-decoder architecture. The block has a multi-branch structure that enables the model to extract more diverse vascular features. It facilitates weight distribution across channels through the incorporation of soft attention, which effectively aggregates contextual information in vascular images. Furthermore, we suppose an attention mechanism within the skip connection. This mechanism serves to enhance feature integration across various layers, thereby mitigating the degradation of effective information. It helps acquire cross-channel information and enhance the localization of regions of interest, ultimately leading to improved recognition of vascular structures. In addition, the feature fusion module (FFM) module is used to provide semantic information for a more refined vascular segmentation map. We evaluated CCS-UNet based on five benchmark retinal image datasets, DRIVE, CHASEDB1, STARE, IOSTAR and HRF. Our proposed method exhibits superior segmentation efficacy compared to other state-of-the-art techniques with a global accuracy of 0.9617/0.9806/0.9766/0.9786/0.9834 and AUC of 0.9863/0.9894/0.9938/0.9902/0.9855 on DRIVE, CHASEDB1, STARE, IOSTAR and HRF respectively. Ablation studies are also performed to evaluate the the relative contributions of different architectural components. Our proposed model is potential for diagnostic aid of retinal diseases.

Research on Siamese network algorithm based on parallel channel attention mechanism for target tracking

In order to effectively improve the tracking performance of the target in various complex environment in the tracking process, the reinforcement research of target features has become one of the important work. In this paper, a Siamese network target tracking algorithm based on parallel channel attention mechanism (PCAM) is proposed by combining feature cascade algorithm with visual attention. Firstly, the characteristics of SENet network and ECA network are fully analyzed. Secondly, the parallel channel attention mechanism is constructed based on ECA module, which integrates global average pooling and maximum pooling. Parallel channel attention mechanism not only solves the problem of channel correlation reduction of SENet module, but also solves the problem of target feature information enhancement. Thirdly, the output model of channel attention is used as the input of spatial attention model to realize the effective complement to channel attention mechanism. By calculating the weight value of different spatial locations, the structural relation between spatial location information is constructed, the feature expression ability of the model is enhanced. Finally, the algorithm is evaluated on standard data sets OTB100, OTB2013, OTB2015, VOT2016 and VOT2018. Experimental results show that the PCAM has stronger feature extraction performance for complex environment, higher target tracking accuracy and robustness, and has strong advantages compared with other comparative experiments.

Enhancing Fingerprint Liveness Detection Accuracy Using Deep Learning: A Comprehensive Study and Novel Approach.

Liveness detection for fingerprint impressions plays a role in the meaningful prevention of any unauthorized activity or phishing attempt. The accessibility of unique individual identification has increased the popularity of biometrics. Deep learning with computer vision has proven remarkable results in image classification, detection, and many others. The proposed methodology relies on an attention model and ResNet convolutions. Spatial attention (SA) and channel attention (CA) models were used sequentially to enhance feature learning. A three-fold sequential attention model is used along with five convolution learning layers. The method's performances have been tested across different pooling strategies, such as Max, Average, and Stochastic, over the LivDet-2021 dataset. Comparisons against different state-of-the-art variants of Convolutional Neural Networks, such as DenseNet121, VGG19, InceptionV3, and conventional ResNet50, have been carried out. In particular, tests have been aimed at assessing ResNet34 and ResNet50 models on feature extraction by further enhancing the sequential attention model. A Multilayer Perceptron (MLP) classifier used alongside a fully connected layer returns the ultimate prediction of the entire stack. Finally, the proposed method is also evaluated on feature extraction with and without attention models for ResNet and considering different pooling strategies.