SARC-UNet: A coronary artery segmentation method based on spatial attention and residual convolution.

BACKGROUND AND OBJECTIVE: Coronary artery segmentation is a pivotal field that has received increasing attention in recent years. However, this task remains challenging because of the inhomogeneous distributions of the contrast agent and dim light, resulting in noise, vascular breakages and small vessel losses in the obtained segmentation results. METHODS: To acquire better automatic blood vessel segmentation results for coronary angiography images, a UNet-based segmentation network (SARC-UNet) is constructed for coronary artery segmentation; this approach is based on residual convolution and spatial attention. First, we use the low-light image enhancement (LIME) approach to increase the contrast and clarity levels of coronary angiography images. Then, we design two residual convolution fusion modules (RCFM1 and RCFM2) that can successfully fuse the local and global information of coronary images while also capturing the characteristics of finer-grained blood vessels, hence preventing the loss of tiny blood vessels in the segmentation findings. Finally, using a cascaded waterfall structure, we create a new location-enhanced spatial attention (LESA) mechanism that can efficiently improve the long-distance dependencies between coronary vascular pixel features, eradicating vascular ruptures and noise in the segmentation results. RESULTS: This article subjectively and objectively evaluates the experimental results. This method has performed well on five general indicators. Furthermore, it outperforms the connectivity indicators proposed in this article. This method can effectively segment blood vessels and obtain higher accuracy results. CONCLUSIONS: Numerous experiments have shown that the suggested method outperforms the state-of-the-art approaches, particularly in terms of vessel connectivity and small blood vessel segmentation.

Electroencephalogram Decoding Reveals Distinct Processes for Directing Spatial Attention and Encoding Into Working Memory.

Past work reveals a tight relationship between spatial attention and storage in visual working memory. But is spatially attending an item tantamount to working memory encoding? Here, we tracked electroencephalography (EEG) signatures of spatial attention and working memory encoding while independently manipulating the number of memory items and the spatial extent of attention in two studies of adults (N = 39; N = 33). Neural measures of spatial attention tracked the position and size of the attended area independent of the number of individuated items encoded into working memory. At the same time, multivariate decoding of the number of items stored in working memory was insensitive to variations in the breadth and position of spatial attention. Finally, representational similarity analyses provided converging evidence for a pure load signal that is insensitive to the spatial extent of the stored items. Thus, although spatial attention is a persistent partner of visual working memory, it is functionally dissociable from the selection and maintenance of individuated representations in working memory.

Correlated variability and its attentional modulation depend on anatomical connectivity.

Visual cortical neurons show variability in their responses to repeated presentations of a stimulus and a portion of this variability is shared across neurons. Attention may enhance visual perception by reducing shared spiking variability. However, shared variability and its attentional modulation are not consistent within or across cortical areas, and depend on additional factors such as neuronal type. A critical factor that has not been tested is actual anatomical connectivity. We measured spike count correlations among pairs of simultaneously recorded neurons in the primary visual cortex (V1) for which anatomical connectivity was inferred from spiking cross-correlations. Neurons were recorded in monkeys performing a contrast-change discrimination task requiring covert shifts in visual spatial attention. Accordingly, spike count correlations were compared across trials in which attention was directed toward or away from the visual stimulus overlapping recorded neuronal receptive fields. Consistent with prior findings, attention did not significantly alter spike count correlations among random pairings of unconnected V1 neurons. However, V1 neurons connected via excitatory synapses showed a significant reduction in spike count correlations with attention. Interestingly, V1 neurons connected via inhibitory synapses demonstrated high spike count correlations overall that were not modulated by attention. Correlated variability in excitatory circuits also depended upon neuronal tuning for contrast, the task-relevant stimulus feature. These results indicate that shared variability depends on the type of connectivity in neuronal circuits. Also, attention significantly reduces shared variability in excitatory circuits, even when attention effects on randomly sampled neurons within the same area are weak.

Electrophysiological correlation of auditory selective spatial attention in the "cocktail party" situation.

The auditory system can selectively attend to the target source in complex environments, the phenomenon known as the "cocktail party" effect. However, the spatiotemporal dynamics of electrophysiological activity associated with auditory selective spatial attention (ASSA) remain largely unexplored. In this study, single-source and multiple-source paradigms were designed to simulate different auditory environments, and microstate analysis was introduced to reveal the electrophysiological correlates of ASSA. Furthermore, cortical source analysis was employed to reveal the neural activity regions of these microstates. The results showed that five microstates could explain the spatiotemporal dynamics of ASSA, ranging from MS1 to MS5. Notably, MS2 and MS3 showed significantly lower partial properties in multiple-source situations than in single-source situations, whereas MS4 had shorter durations and MS5 longer durations in multiple-source situations than in single-source situations. MS1 had insignificant differences between the two situations. Cortical source analysis showed that the activation regions of these microstates initially transferred from the right temporal cortex to the temporal-parietal cortex, and subsequently to the dorsofrontal cortex. Moreover, the neural activity of the single-source situations was greater than that of the multiple-source situations in MS2 and MS3, correlating with the N1 and P2 components, with the greatest differences observed in the superior temporal gyrus and inferior parietal lobule. These findings suggest that these specific microstates and their associated activation regions may serve as promising substrates for decoding ASSA in complex environments.

Fourier Ptychographic Microscopy Reconstruction Method Based on Residual Local Mixture Network.

Fourier Ptychographic Microscopy (FPM) is a microscopy imaging technique based on optical principles. It employs Fourier optics to separate and combine different optical information from a sample. However, noise introduced during the imaging process often results in poor resolution of the reconstructed image. This article has designed an approach based on a residual local mixture network to improve the quality of Fourier ptychographic reconstruction images. By incorporating channel attention and spatial attention into the FPM reconstruction process, the network enhances the efficiency of the network reconstruction and reduces the reconstruction time. Additionally, the introduction of the Gaussian diffusion model further reduces coherent artifacts and improves image reconstruction quality. Comparative experimental results indicate that this network achieves better reconstruction quality, and outperforming existing methods in both subjective observation and objective quantitative evaluation.

Time to get deep.

Asymmetries in the size of structures deep below the cortex explain how alpha oscillations in the brain respond to shifts in attention.

Modulation of alpha oscillations by attention is predicted by hemispheric asymmetry of subcortical regions.

Evidence suggests that subcortical structures play a role in high-level cognitive functions such as the allocation of spatial attention. While there is abundant evidence in humans for posterior alpha band oscillations being modulated by spatial attention, little is known about how subcortical regions contribute to these oscillatory modulations, particularly under varying conditions of cognitive challenge. In this study, we combined MEG and structural MRI data to investigate the role of subcortical structures in controlling the allocation of attentional resources by employing a cued spatial attention paradigm with varying levels of perceptual load. We asked whether hemispheric lateralization of volumetric measures of the thalamus and basal ganglia predicted the hemispheric modulation of alpha-band power. Lateral asymmetry of the globus pallidus, caudate nucleus, and thalamus predicted attention-related modulations of posterior alpha oscillations. When the perceptual load was applied to the target and the distractor was salient caudate nucleus asymmetry predicted alpha-band modulations. Globus pallidus was predictive of alpha-band modulations when either the target had a high load, or the distractor was salient, but not both. Finally, the asymmetry of the thalamus predicted alpha band modulation when neither component of the task was perceptually demanding. In addition to delivering new insight into the subcortical circuity controlling alpha oscillations with spatial attention, our finding might also have clinical applications. We provide a framework that could be followed for detecting how structural changes in subcortical regions that are associated with neurological disorders can be reflected in the modulation of oscillatory brain activity.

What is the Role of Spatial Attention in Statistical Learning During Visual Search?

Our ability to learn the regularities embedded in our environment is a fundamental aspect of our cognitive system. Does such statistical learning depend on attention? Research on this topic is scarce and has yielded mixed findings. In this preregistered study, we examined the role of spatial attention in statistical learning, and specifically in learned distractor-location suppression. This phenomenon refers to the finding that during visual search, participants are better at ignoring a salient distractor at a high-probability location than at low-probability locations - a bias persisting long after the probability imbalance has ceased. Participants searched for a shape-singleton target and a color-singleton distractor was sometimes present. During the learning phase, the color-singleton distractor was more likely to appear in the high-probability location than in the low-probability locations. Crucially, we manipulated spatial attention by having the experimental group focus their attention on the target's location in advance of the search display, using a 100%-informative spatial precue, while the control group was presented with a neutral, uninformative cue. During the subsequent test phase, the color-singleton distractor was equally likely to appear at any location and there were no cues. As expected, the results for the neutral-cue group replicated previous findings. Crucially, for the informative-cue group, interference from the distractor was minimal when attention was diverted from it (during learning) and no statistical learning was observed during test. Intertrial priming accounted for the small statistical-learning effect found during learning. These findings show that statistical learning in visual search requires attention.

Multi-Scale Spatial Attention-Based Multi-Channel 2D Convolutional Network for Soil Property Prediction.

Visible near-infrared spectroscopy (VNIR) is extensively researched for obtaining soil property information due to its rapid, cost-effective, and environmentally friendly advantages. Despite its widespread application and significant achievements in soil property analysis, current soil prediction models continue to suffer from low accuracy. To address this issue, we propose a convolutional neural network model that can achieve high-precision soil property prediction by creating 2D multi-channel inputs and applying a multi-scale spatial attention mechanism. Initially, we explored two-dimensional multi-channel inputs for seven soil properties in the public LUCAS spectral dataset using the Gramian Angular Field (GAF) method and various preprocessing techniques. Subsequently, we developed a convolutional neural network model with a multi-scale spatial attention mechanism to improve the network's extraction of relevant spatial contextual information. Our proposed model showed superior performance in a statistical comparison with current state-of-the-art techniques. The RMSE (R²) values for various soil properties were as follows: organic carbon content (OC) of 19.083 (0.955), calcium carbonate content (CaCO3) of 24.901 (0.961), nitrogen content (N) of 0.969 (0.933), cation exchange capacity (CEC) of 6.52 (0.803), pH in H2O of 0.366 (0.927), clay content of 4.845 (0.86), and sand content of 12.069 (0.789). Our proposed model can effectively extract features from visible near-infrared spectroscopy data, contributing to the precise detection of soil properties.

Attention spotlight in V1-based cortico-cortical interactions in human visual hierarchy.

Attention is often viewed as a mental spotlight, which can be scaled like a zoom lens at specific spatial locations and features a center-surround gradient. Here, we demonstrate a neural signature of attention spotlight in signal transmission along the visual hierarchy. fMRI background connectivity analysis was performed between retinotopic V1 and downstream areas to characterize the spatial distribution of inter-areal interaction under two attentional states. We found that, compared to diffused attention, focal attention sharpened the spatial gradient in the strength of the background connectivity. Dynamic causal modeling analysis further revealed the effect of attention in both the feedback and feedforward connectivity between V1 and extrastriate cortex. In a context which induced a strong effect of crowding, the effect of attention in the background connectivity profile diminished. Our findings reveal a context-dependent attention prioritization in information transmission via modulating the recurrent processing across the early stages in human visual cortex.

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.

MADR-Net: multi-level attention dilated residual neural network for segmentation of medical images.

Medical image segmentation has made a significant contribution towards delivering affordable healthcare by facilitating the automatic identification of anatomical structures and other regions of interest. Although convolution neural networks have become prominent in the field of medical image segmentation, they suffer from certain limitations. In this study, we present a reliable framework for producing performant outcomes for the segmentation of pathological structures of 2D medical images. Our framework consists of a novel deep learning architecture, called deep multi-level attention dilated residual neural network (MADR-Net), designed to improve the performance of medical image segmentation. MADR-Net uses a U-Net encoder/decoder backbone in combination with multi-level residual blocks and atrous pyramid scene parsing pooling. To improve the segmentation results, channel-spatial attention blocks were added in the skip connection to capture both the global and local features and superseded the bottleneck layer with an ASPP block. Furthermore, we introduce a hybrid loss function that has an excellent convergence property and enhances the performance of the medical image segmentation task. We extensively validated the proposed MADR-Net on four typical yet challenging medical image segmentation tasks: (1) Left ventricle, left atrium, and myocardial wall segmentation from Echocardiogram images in the CAMUS dataset, (2) Skin cancer segmentation from dermoscopy images in ISIC 2017 dataset, (3) Electron microscopy in FIB-SEM dataset, and (4) Fluid attenuated inversion recovery abnormality from MR images in LGG segmentation dataset. The proposed algorithm yielded significant results when compared to state-of-the-art architectures such as U-Net, Residual U-Net, and Attention U-Net. The proposed MADR-Net consistently outperformed the classical U-Net by 5.43%, 3.43%, and 3.92% relative improvement in terms of dice coefficient, respectively, for electron microscopy, dermoscopy, and MRI. The experimental results demonstrate superior performance on single and multi-class datasets and that the proposed MADR-Net can be utilized as a baseline for the assessment of cross-dataset and segmentation tasks.

Integrating attention mechanism and multi-scale feature extraction for fall detection.

Addressing the critical need for accurate fall event detection due to their potentially severe impacts, this paper introduces the Spatial Channel and Pooling Enhanced You Only Look Once version 5 small (SCPE-YOLOv5s) model. Fall events pose a challenge for detection due to their varying scales and subtle pose features. To address this problem, SCPE-YOLOv5s introduces spatial attention to the Efficient Channel Attention (ECA) network, which significantly enhances the model's ability to extract features from spatial pose distribution. Moreover, the model integrates average pooling layers into the Spatial Pyramid Pooling (SPP) network to support the multi-scale extraction of fall poses. Meanwhile, by incorporating the ECA network into SPP, the model effectively combines global and local features to further enhance the feature extraction. This paper validates the SCPE-YOLOv5s on a public dataset, demonstrating that it achieves a mean Average Precision of 88.29 %, outperforming the You Only Look Once version 5 small by 4.87 %. Additionally, the model achieves 57.4 frames per second. Therefore, SCPE-YOLOv5s provides a novel solution for fall event detection.

Peripheral vision contributes to implicit attentional learning: Findings from the "mouse-eye" paradigm.

The central visual field is essential for activities like reading and face recognition. However, the impact of peripheral vision loss on daily activities is profound. While the importance of central vision is well established, the contribution of peripheral vision to spatial attention is less clear. In this study, we introduced a "mouse-eye" method as an alternative to traditional gaze-contingent eye tracking. We found that even in tasks requiring central vision, peripheral vision contributes to implicit attentional learning. Participants searched for a T among Ls, with the T appearing more often in one visual quadrant. Earlier studies showed that participants' awareness of the T location probability was not essential for their ability to learn. When we limited the visible area around the mouse cursor, only participants aware of the target's location probability showed learning; those unaware did not. Adding placeholders in the periphery did not restore implicit attentional learning. A control experiment showed that when participants were allowed to see all items while searching and moving the mouse to reveal the target's color, both aware and unaware participants acquired location probability learning. Our results underscore the importance of peripheral vision in implicitly guided attention. Without peripheral vision, only explicit, but not implicit, attentional learning prevails.

Auditory vigilance task performance and cerebral hemodynamics: effects of spatial uncertainty.

The vigilance decrement, a temporal decline in detection performance, has been observed across multiple sensory modalities. Spatial uncertainty about the location of task-relevant stimuli has been demonstrated to increase the demands of vigilance and increase the severity of the vigilance decrement when attending to visual displays. The current study investigated whether spatial uncertainty also increases the severity of the vigilance decrement and task demands when an auditory display is used. Individuals monitored an auditory display to detect critical signals that were shorter in duration than non-target stimuli. These auditory stimuli were presented in either a consistent, predictable pattern that alternated sound presentation from left to right (spatial certainty) or an inconsistent, unpredictable pattern that randomly presented sounds from the left or right (spatial uncertainty). Cerebral blood flow velocity (CBFV) was measured to assess the neurophysiological demands of the task. A decline in performance and CBFV was observed in both the spatially certain and spatially uncertain conditions, suggesting that spatial auditory vigilance tasks are demanding and can result in a vigilance decrement. Spatial uncertainty resulted in a more severe vigilance decrement in correct detections compared to spatial certainty. Reduced right-hemispheric CBFV was also observed during spatial uncertainty compared to spatial certainty. Together, these results suggest that auditory spatial uncertainty hindered performance and required greater attentional demands compared to spatial certainty. These results concur with previous research showing the negative impact of spatial uncertainty in visual vigilance tasks, but the current results contrast recent research showing no effect of spatial uncertainty on tactile vigilance.

Sustained bias of spatial attention in a 3 T MRI scanner.

When lying inside a MRI scanner and even in the absence of any motion, the static magnetic field of MRI scanners induces a magneto-hydrodynamic stimulation of subjects' vestibular organ (MVS). MVS thereby not only causes a horizontal vestibular nystagmus but also induces a horizontal bias in spatial attention. In this study, we aimed to determine the time course of MVS-induced biases in both VOR and spatial attention inside a 3 T MRI-scanner as well as their respective aftereffects after participants left the scanner. Eye movements and overt spatial attention in a visual search task were assessed in healthy volunteers before, during, and after a one-hour MVS period. All participants exhibited a VOR inside the scanner, which declined over time but never vanished completely. Importantly, there was also an MVS-induced horizontal bias in spatial attention and exploration, which persisted throughout the entire hour within the scanner. Upon exiting the scanner, we observed aftereffects in the opposite direction manifested in both the VOR and in spatial attention, which were statistically no longer detectable after 7 min. Sustained MVS effects on spatial attention have important implications for the design and interpretation of fMRI-studies and for the development of therapeutic interventions counteracting spatial neglect.

An Improved Ningxia Desert Herbaceous Plant Classification Algorithm Based on YOLOv8.

Wild desert grasslands are characterized by diverse habitats, uneven plant distribution, similarities among plant class, and the presence of plant shadows. However, the existing models for detecting plant species in desert grasslands exhibit low precision, require a large number of parameters, and incur high computational cost, rendering them unsuitable for deployment in plant recognition scenarios within these environments. To address these challenges, this paper proposes a lightweight and fast plant species detection system, termed YOLOv8s-KDT, tailored for complex desert grassland environments. Firstly, the model introduces a dynamic convolutional KernelWarehouse method to reduce the dimensionality of convolutional kernels and increase their number, thus achieving a better balance between parameter efficiency and representation ability. Secondly, the model incorporates triplet attention into its feature extraction network, effectively capturing the relationship between channel and spatial position and enhancing the model's feature extraction capabilities. Finally, the introduction of a dynamic detection head tackles the issue related to target detection head and attention non-uniformity, thus improving the representation of the target detection head while reducing computational cost. The experimental results demonstrate that the upgraded YOLOv8s-KDT model can rapidly and effectively identify desert grassland plants. Compared to the original model, FLOPs decreased by 50.8%, accuracy improved by 4.5%, and mAP increased by 5.6%. Currently, the YOLOv8s-KDT model is deployed in the mobile plant identification APP of Ningxia desert grassland and the fixed-point ecological information observation platform. It facilitates the investigation of desert grassland vegetation distribution across the entire Ningxia region as well as long-term observation and tracking of plant ecological information in specific areas, such as Dashuikeng, Huangji Field, and Hongsibu in Ningxia.

The who and the where: Attention to identities and locations in groups.

While it is widely accepted that the single gaze of another person elicits shifts of attention, there is limited work on the effects of multiple gazes on attention, despite real-world social cues often occurring in groups. Further, less is known regarding the role of unequal reliability of varying social and nonsocial information on attention. We addressed these gaps by employing a variant of the gaze cueing paradigm, simultaneously presenting participants with three faces. Block-wise, we manipulated whether one face (Identity condition) or one location (Location condition) contained a gaze cue entirely predictive of target location; all other cues were uninformative. Across trials, we manipulated the number of valid cues (number of faces gazing at target). We examined whether these two types of information (Identity vs. Location) were learned at a similar rate by statistically modelling cueing effects by trial count. Preregistered analyses returned no evidence for an interaction between condition, number of valid faces, and presence of the predictive element, indicating type of information did not affect participants' ability to employ the predictive element to alter behaviour. Exploratory analyses demonstrated (i) response times (RT) decreased faster across trials for the Identity compared with Location condition, with greater decreases when the predictive element was present versus absent, (ii) RTs decreased across trials for the Location condition only when it was completed first, and (iii) social competence altered RTs across conditions and trial number. Our work demonstrates a nuanced relationship between cue utility, condition type, and social competence on group cueing.

Effects of Functional Electrical Stimulation on Attention and Brain Activity in Healthy Participants Using Near-Infrared Spectroscopy: An Interventional Study.

Background Involuntary limb activation using functional electrical stimulation (FES) can improve unilateral spatial neglect. However, the impact of FES on brain activity related to spatial attention remains unclear. Thus, in this study, we aimed to examine the effects of FES on spatial attention. Methodology In this interventional study, 13 healthy right-handed participants were asked to perform the Posner task for six minutes both before and after either FES or sham stimulation during each set, resulting in a total of two sets. FES was applied to the left forearm extensor muscles, with a frequency of 25 Hz, a pulse width of 100 µs, and the intensity adjusted to reach the motor threshold. Both the energization and pause times were set to five seconds. The Posner task was used to measure reaction time to a target appearing on a computer screen. Brain activity, indicated by oxygenated hemoglobin values, was measured using near-infrared spectroscopy with 24 probes according to the International 10-20 system method. Results In the left hemisphere, oxygenated hemoglobin values in the premotor and supplementary motor areas, primary somatosensory cortex, and somatosensory association areas were significantly higher after FES than after sham stimulation. In the right hemisphere, oxygenated hemoglobin values were significantly increased in the premotor, primary, and supplementary motor areas; in the supramarginal gyrus; and in the somatosensory association areas after FES. Reaction times in the Posner task did not differ significantly between the FES and sham conditions. Conclusions Collectively, these results suggest that FES of the upper limbs can activate the ventral pathway of the visual attention network and improve stimulus-driven attention. Activation of stimulus-driven attentional function could potentially contribute to symptom improvement in patients with unilateral spatial neglect.

Fine-grained image classification method based on hybrid attention module.

To efficiently capture feature information in tasks of fine-grained image classification, this study introduces a new network model for fine-grained image classification, which utilizes a hybrid attention approach. The model is built upon a hybrid attention module (MA), and with the assistance of the attention erasure module (EA), it can adaptively enhance the prominent areas in the image and capture more detailed image information. Specifically, for tasks involving fine-grained image classification, this study designs an attention module capable of applying the attention mechanism to both the channel and spatial dimensions. This highlights the important regions and key feature channels in the image, allowing for the extraction of distinct local features. Furthermore, this study presents an attention erasure module (EA) that can remove significant areas in the image based on the features identified; thus, shifting focus to additional feature details within the image and improving the diversity and completeness of the features. Moreover, this study enhances the pooling layer of ResNet50 to augment the perceptual region and the capability to extract features from the network's less deep layers. For the objective of fine-grained image classification, this study extracts a variety of features and merges them effectively to create the final feature representation. To assess the effectiveness of the proposed model, experiments were conducted on three publicly available fine-grained image classification datasets: Stanford Cars, FGVC-Aircraft, and CUB-200-2011. The method achieved classification accuracies of 92.8, 94.0, and 88.2% on these datasets, respectively. In comparison with existing approaches, the efficiency of this method has significantly improved, demonstrating higher accuracy and robustness.