Local Map Research Articles

MCTformer+: Multi-Class Token Transformer for Weakly Supervised Semantic Segmentation.

This paper proposes a novel transformer-based framework to generate accurate class-specific object localization maps for weakly supervised semantic segmentation (WSSS). Leveraging the insight that the attended regions of the one-class token in the standard vision transformer can generate class-agnostic localization maps, we investigate the transformer's capacity to capture class-specific attention for class-discriminative object localization by learning multiple class tokens. We present the Multi-Class Token transformer, which incorporates multiple class tokens to enable class-aware interactions with patch tokens. This is facilitated by a class-aware training strategy that establishes a one-to-one correspondence between output class tokens and ground-truth class labels. We also introduce a Contrastive-Class-Token (CCT) module to enhance the learning of discriminative class tokens, enabling the model to better capture the unique characteristics of each class. Consequently, the proposed framework effectively generates class-discriminative object localization maps from the class-to-patch attentions associated with different class tokens. To refine these localization maps, we propose the utilization of patch-level pairwise affinity derived from the patch-to-patch transformer attention. Furthermore, the proposed framework seamlessly complements the Class Activation Mapping (CAM) method, yielding significant improvements in WSSS performance on PASCAL VOC 2012 and MS COCO 2014. These results underline the importance of the class token for WSSS.

LTM-NeRF: Embedding 3D Local Tone Mapping in HDR Neural Radiance Field.

Recent advances in Neural Radiance Fields (NeRF) have provided a new geometric primitive for novel view synthesis. High Dynamic Range NeRF (HDR NeRF) can render novel views with a higher dynamic range. However, effectively displaying the scene contents of HDR NeRF on diverse devices with limited dynamic range poses a significant challenge. To address this, we present LTM-NeRF, a method designed to recover HDR NeRF and support 3D local tone mapping. LTM-NeRF allows for the synthesis of HDR views, tone-mapped views, and LDR views under different exposure settings, using only the multi-view multi-exposure LDR inputs for supervision. Specifically, we propose a differentiable Camera Response Function (CRF) module for HDR NeRF reconstruction, globally mapping the scene's HDR radiance to LDR pixels. Moreover, we introduce a Neural Exposure Field (NeEF) to represent the spatially varying exposure time of an HDR NeRF to achieve 3D local tone mapping, for compatibility with various displays. Comprehensive experiments demonstrate that our method can not only synthesize HDR views and exposure-varying LDR views accurately but also render locally tone-mapped views naturally.

Dispersion of Activation in Single-Beat Global Maps During Programmed Ventricular Stimulation Identifies Infarct-Related Ventricular Tachycardia Isthmus Sites.

Electrophysiological characterization of ventricular tachycardia (VT) isthmus sites is complex and time-consuming. We aimed at developing and validating a global mapping strategy during programmed ventricular stimulation (PVS) to reveal the underlying electrophysiological properties of the infarct-related substrate and to enable identification of highly heterogeneous activation sites associated with protected VT isthmus sites. Experimental study that included 22 pigs with established myocardial infarction undergoing invivo characterization of the anatomical and functional myocardial substrate associated with potential arrhythmogenicity. High-density sequential activation maps during ventricular pacing and VT were compared with single-beat maps using a 64-pole basket catheter positioned in the left ventricle. Further analyses were performed using a novel local activation time-dispersion score to identify regional activation time heterogeneities on both baseline drive pacing and each of the extrastimuli of the PVS protocol. Basket catheter splines covered a median of 81.2% of the endocardial surface of the left ventricle. Basket-catheter-derived single-beat activation maps (N=16) during pacing showed a linear relationship with high-density sequential activation maps. Induction of ventricular arrhythmias was associated with higher local activation time-dispersion score values on single-beat global maps during PVS (N=6, 46 arrhythmia induction attempts). Single-beat-derived local activation time-dispersion score maps during successive coupled extrastimuli of the PVS showed a progressive increase in the predictive performance to identify monomorphic VT isthmus sites within the scar region (area under the curve = 0.779 in S2, area under the curve = 0.859 in S4; N=7). Sixty-four-pole-derived single-beat local activation time-dispersion score global maps during PVS identify infarct-related endocardial regions with highly heterogeneous activation times that are associated with protected VT isthmus sites.

PLM-SLAM: Enhanced Visual SLAM for Mobile Robots in Indoor Dynamic Scenes Leveraging Point-Line Features and Manhattan World Model

This paper proposes an enhanced visual simultaneous localization and mapping (vSLAM) algorithm tailored for mobile robots operating in indoor dynamic scenes. By incorporating point-line features and leveraging the Manhattan world model, the proposed PLM-SLAM framework significantly improves localization accuracy and map consistency. This algorithm optimizes the line features detected by the Line Segment Detector (LSD) through merging and pruning strategies, ensuring real-time performance. Subsequently, dynamic point-line features are rejected based on Lucas–Kanade (LK) optical flow, geometric constraints, and depth information, minimizing the impact of dynamic objects. The Manhattan world model is then utilized to reduce rotational estimation errors and optimize pose estimation. High-precision line feature matching and loop closure detection mechanisms further enhance the robustness and accuracy of the system. Experimental results demonstrate the superior performance of PLM-SLAM, particularly in high-dynamic indoor environments, outperforming existing state-of-the-art methods.

Understanding Sociodemographic Determinants of Voluntary Cesarean Section Deliveries in India through Geospatial Modeling using Multiscale Geographically Weighted Regression (MGWR)

IntroductionOver the period of time, the increasing trend of voluntary cesarean section (CS) delivery has become a serious maternal health concern. Previous studies have shown wide geographic variations in the prevalence of voluntary CS delivery constraints by several sociodemographic determinants. However, none of them have attempted to disentangle this phenomenon by using spatially varying coefficient models to consider local factors. This study aims to identify the spatially heterogeneous relationships among voluntary CS deliveries against the backdrop of sociodemographic determinants in India.Data and MethodsThis study utilized data from the National Family Health Survey (NFHS-5), 2019-21. The ordinary least square (OLS) model was used as a base model, and multiscale geographically weighted regression (MGWR) was employed at the district level to consider the local factors associated with voluntary CS delivery. An appropriate model diagnostic check was also performed to make our model robust and reliable. Several maps were produced using ArcGIS Pro to visualize the distribution of coefficients of several factors.ResultsMore than half (58%) of the total CS deliveries were voluntary; however, choropleth maps have shown that southern peninsular and northeastern districts have a higher prevalence than the rest. The distribution of voluntary CS deliveries with a Moran’s Index of 0.22 was also spatially autocorrelated. Moreover, the values of Jarque–Bera statistics and Koenker (BP) Statistics show that the OLS model residuals were not normally distributed and reflected heteroskedasticity. The estimates from the MGWR show that four and above ANC visits and mass media exposure show a positive relationship with voluntary CS deliveries. The local R-squared map explains nearly half of the variation for voluntary CS deliveries for all northeastern districts. The model fit metric shows higher R square and lower corrected Akaike’s Information Criterion (AICc) values for the MGWR model, which validates that the MGWR model is more robust and parsimonious.ConclusionThe present study revealed that although voluntary CS deliveries are at an increasingly alarming pace, the geographic distribution of voluntary CS deliveries is, albeit unequally. Therefore, this investigation has made an effort to examine how spatial analysis can help identify the geographic patterns and factors that explain this variation in voluntary CS deliveries across space. These findings add value to our understanding of how geographic space and scale matter when health is examined. Public health practitioners can recognize such variations when devising targeted interventions for CS deliveries in India and other developing nations.

ENHANCING MOBILITY WITH IOT-BASED AUTONOMOUS WHEELCHAIR

The shortage of healthcare workers and continuous influx of patients have significantly increased workloads in hospitals. Concurrently, there is an increasing number of people with disabilities, resulting in a higher need for wheelchairs. Yet, many hospitals continue to rely on traditional, manually operated wheelchairs. As a result, disabled patients often face prolonged waits for assistance from hospital staff to reach their intended destinations. To address hospital issues, the idea of autonomous wheelchairs, known as autonomous robotic wheelchairs (ARWs), has been proposed and developed. In contrast to traditional wheelchairs, ARWs integrate advanced features, including obstacle detection and avoidance, local path mapping through line-following technology, and user-friendly interactions. Patients in developing countries, including the Kurdistan region of Iraq (KRI), rely on hospital workers for transportation, which poses difficulties, especially during emergencies, such as overcrowding, delays due to a lack of available porters, etc. Implementing ARWs would significantly reduce these issues. Thus, the main aim of this study is to develop ARWs using technologies like Arduino and sensors to enhance hospital efficiency and reduce reliance on porters, potentially transforming patient mobility and care in healthcare facilities. The testing results of the proposed system indicate that its implementation will greatly assist hospitals in addressing various issues, including those previously mentioned. In addition, it will enhance hospitals’ intelligence and autonomy.

EgoHDM: A Real-time Egocentric-Inertial Human Motion Capture, Localization, and Dense Mapping System

We present EgoHDM, an online egocentric-inertial human motion capture (mocap), localization, and dense mapping system. Our system uses 6 inertial measurement units (IMUs) and a commodity head-mounted RGB camera. EgoHDM is the first human mocap system that offers dense scene mapping in near real-time. Further, it is fast and robust to initialize and fully closes the loop between physically plausible map-aware global human motion estimation and mocap-aware 3D scene reconstruction. To achieve this, we design a tightly coupled mocap-aware dense bundle adjustment and physics-based body pose correction module leveraging a local body-centric elevation map. The latter introduces a novel terrain-aware contact PD controller, which enables characters to physically contact the given local elevation map thereby reducing human floating or penetration. We demonstrate the performance of our system on established synthetic and real-world benchmarks. The results show that our method reduces human localization, camera pose, and mapping accuracy error by 41%, 71%, 46%, respectively, compared to the state of the art. Our qualitative evaluations on newly captured data further demonstrate that EgoHDM can cover challenging scenarios in non-flat terrain including stepping over stairs and outdoor scenes in the wild. Our project page: https://handiyin.github.io/EgoHDM/

A 30 Arcsec Resolution VS30 Map for Mainland China Using Refined Topographic Slope-Based Cokriging (SCK) Model

ABSTRACT In 2022, we proposed a VS30 proxy model based on the Cokriging method, integrating local VS30 measurements with topographic slope data (SCK model, Zhou et al., 2022). Utilizing this model, a VS30 map for mainland China was developed based on 7797 local boreholes. However, the initial VS30 map was produced with a coarse grid resolution of 360 arcsec and did not comprehensively cover mainland China due to sparse distribution of boreholes in certain areas. In this study, we refined the SCK model by explicitly accounting for the influence of surrounding topographic slopes, incorporating more distant VS30 measurements, and improving the spatial distribution of VS30 measurements used in the model calculation. Tests based on virtual cases were conducted to evaluate the characteristics of the refined SCK model. In addition, 142 VS30 measurements from strong-motion stations in western China were added to enhance spatial coverage and model accuracy in these regions. Employing the refined SCK model and the expanded VS30 data set, a new VS30 map for mainland China was generated. This updated map features a higher grid resolution of 30 arcsec and covers the entirety of mainland China, showing reduced estimation errors and improved continuity. Comparisons with the U.S. Geological Survey Global VS30 Mosaic indicate that the results from the refined SCK model are more reasonable, as evidenced by local topography, VS profiles, lithologic profiles, and geology map, which better capture the impact of local geomorphological differences and climatic conditions on VS30 values.

Secondary Structure Detection and Structure Modeling for Cryo-EM.

Rapid advancements in cryogenic electron microscopy (cryo-EM) have revolutionized the field of structural biology by enabling the determination of complex macromolecular structures at unprecedented resolutions. When cryo-EM density maps have a resolution around 3Å, the atomic structure can be modeled manually. However, as the resolution decreases, analyzing these density maps becomes increasingly challenging. For modeling structures in lower resolution maps, deep learning can be used to identify structural features in the maps to assist in structure modeling.Here, we present a suite of deep learning-based tools developed by our lab that enable structural biologists to work with cryo-EM maps of a wide range of resolutions. For cryo-EM maps at near-atomic resolution (5Å or better), DeepMainmast automatically models all-atom structures by tracing the main chain from local map features of amino acids and atoms detected by deep learning; DAQ score quantifies map-model fit and indicates potential misassignments in protein models. In intermediate resolution maps (5-10Å), Emap2sec and Emap2sec+ can accurately detect protein secondary structures and nucleic acids. These tools and more are available at our web server: https://em.kiharalab.org/ .

A Monocular Ranging Method for Ship Targets Based on Unmanned Surface Vessels in a Shaking Environment

Aiming to address errors in the estimation of the position and attitude of an unmanned vessel, especially during vibration, where the rapid loss of feature point information hinders continuous attitude estimation and global trajectory mapping, this paper improves the monocular ORB-SLAM framework based on the characteristics of the marine environment. In general, we extract the location area of the artificial sea target in the video, build a virtual feature set for it, and filter the background features. When shaking occurs, GNSS information is combined and the target feature set is used to complete the map reconstruction task. Specifically, firstly, the sea target area of interest is detected by YOLOv5, and the feature extraction and matching method is optimized in the front-end tracking stage to adapt to the sea environment. In the key frame selection and local map optimization stage, the characteristics of the feature set are improved to further improve the positioning accuracy, to provide more accurate position and attitude information about the unmanned platform. We use GNSS information to provide the scale and world coordinates for the map. Finally, the target distance is measured by the beam ranging method. In this paper, marine unmanned platform data, GNSS, and AIS position data are autonomously collected, and experiments are carried out using the proposed marine ranging system. Experimental results show that the maximum measurement error of this method is 9.2%, and the average error is 4.7%.

Quantifying the Mesoscale Contribution to FACs During a Magnetospheric Substorm

AbstractMesoscales, which couple small to large scales, and vice‐versa, are critical to the magnetosphere‐ionosphere coupling. Optical and radar measurements indicate that dynamical mesoscale features are present in the ionosphere, however quantifying their contribution to the overall dynamics remains a challenge. We use a new ionospheric data assimilation technique, Lompe (Local mapping of the polar ionospheric electrodynamics), to specify ionospheric electrodynamics using a wide variety of input data and a priori assumptions about the physical nature of the ionospheric electric field. We isolate the terms of the ionospheric Ohm's law and find that mesoscale structures in the FACs are driven by Hall gradients, while the larger scale patterns are associated with the divergence of the electric field. We calculate the relative contribution of mesoscales to the overall FAC patterns during a magnetospheric substorm, and find that in the nightside, mesoscale FACs contribute up to 60% of the total.

RUSNet: Robust fish segmentation in underwater videos based on adaptive selection of optical flow

Fish segmentation in underwater videos can be used to accurately determine the silhouette size of fish objects, which provides key information for fish population monitoring and fishery resources survey. Some researchers have utilized underwater optical flow to improve the fish segmentation accuracy of underwater videos. However, the underwater optical flow is not evaluated and screen in existing works, and its predictions are easily disturbed by motion of non-fish. Therefore, in this paper, by analyzing underwater optical flow data, we propose a robust underwater segmentation network, RUSNet, with adaptive screening and fusion of input information. First, to enhance the robustness of the segmentation model to low-quality optical flow inputs, a global optical flow quality evaluation module is proposed for evaluating and aligning the underwater optical flow. Second, a decoder is designed by roughly localizing the fish object and then applying the proposed multidimension attention (MDA) module to iteratively recover the rough localization map from the spatial and edge dimensions of the fish. Finally, a multioutput selective fusion method is proposed in the testing stage, in which the mean absolute error (MAE) of the prediction using a single input is compared with that obtained using multisource input. Then, the information with the highest confidence is selected for predictive fusion, which facilitates the acquisition of the ultimate underwater fish segmentation results. To verify the effectiveness of the proposed model, we trained and evaluated it using a publicly available joint underwater video dataset and a separate DeepFish public dataset. Compared with the advanced underwater fish segmentation model, the proposed model has greater robustness to low-quality background optical flow in the DeepFish dataset, with the mean pixel accuracy (mPA) and mean intersection over union (mIoU) values reaching 98.77% and 97.65%, respectively. On the joint dataset, the mPA and mIoU of the proposed model are 92.61% and 90.12%, respectively, which are 0.72% and 1.21% higher than those of the advanced underwater video object segmentation model MSGNet. The results indicate that the proposed model can adaptively select the input and accurately segment fish in complex underwater scenes, which provides an effective solution for investigating fishery resources.

Method of Motion Path Planning Based on a Deep Neural Network with Vector Input

The article deals with the problem of path planning in a two-dimensional environment based on deep learning neural networks. Deep neural networks require large amounts of data and place high computational requirements on computing tools. The lack of sufficient data leads to a decrease in the accuracy of the neural network, and high computational requirements at the learning stage limit the use of this technology in engineering practice. In this paper, the forms of representation of the environment for the input of a neural network are studied. Vector form allows to reduce the amount of information supplied to the input of a neural network, but it leads to the need to use more complex neural networks. In this article, a combined form of representation is proposed, including a vector global and local map layout. The vector part of the map includes the position of the robot, the position of the target point and a description of obstacles. The local raster map describes the area closest to the robot. Using numerical research, the effectiveness of this form of data representation for a precise neural network is shown, compared with the raster representation of the map. In this article, two structures of neural networks are studied, one of which uses 8 possible directions of movement, and the other uses 3 possible directions of movement. It is shown that when using 3 possible directions, the cycling of trajectories planned by the neural network is eliminated, which leads to an increase in accuracy.

Reflectance mapping with microsphere-assisted white light interference nanoscopy

The characterisation of novel materials presents a challenge that requires new and original developments. To face some of these demands for making measurements at the nanoscale, a new microsphere-assisted white light interference nanoscope performing local reflectance mapping is presented. This technique presents the advantages of being non-destructive, full-field and label-free. A 145 μm diameter microsphere, glued to the end of an optical fiber, is inserted inside the white light interference microscope to improve the lateral resolution from 940 nm to 520 nm. The acquisition and the Fourier transform processing of a stack of interference images superimposed on the virtual image produced by the microsphere allows the extraction of the local reflectance over a wavelength range of 460 nm to 900 nm and a field of view of 8 μm in diameter. The enhancement in the lateral resolution of the reflectance is demonstrated through the spectral distinction of neighboring ripples on a laser-textured colored stainless-steel sample that cannot be resolved without the microsphere, on regions with a surface of 279 × 279 nm2 horizontally spaced 279 nm apart. Future improvements could potentially lead to a lateral resolution of reflectance measurement over a 100 nm diameter area in air, paving the way to sub-diffraction reflectance mapping.

BreakNet: discontinuity-resilient multi-scale transformer segmentation of retinal layers

Visible light optical coherence tomography (vis-OCT) is gaining traction for retinal imaging due to its high resolution and functional capabilities. However, the significant absorption of hemoglobin in the visible light range leads to pronounced shadow artifacts from retinal blood vessels, posing challenges for accurate layer segmentation. In this study, we present BreakNet, a multi-scale Transformer-based segmentation model designed to address boundary discontinuities caused by these shadow artifacts. BreakNet utilizes hierarchical Transformer and convolutional blocks to extract multi-scale global and local feature maps, capturing essential contextual, textural, and edge characteristics. The model incorporates decoder blocks that expand pathways to enhance the extraction of fine details and semantic information, ensuring precise segmentation. Evaluated on rodent retinal images acquired with prototype vis-OCT, BreakNet demonstrated superior performance over state-of-the-art segmentation models, such as TCCT-BP and U-Net, even when faced with limited-quality ground truth data. Our findings indicate that BreakNet has the potential to significantly improve retinal quantification and analysis.

Local wavenumber imaging in anisotropic structures

Local wavenumber imaging has been proven effective for damage detection in thin-walled structures. Such structures are, however, often inherently anisotropic, posing serious challenges to the known algorithms of local wavenumber mapping. The use of local wavenumber mapping algorithm designed for isotropic, rather than anisotropic, structures results in processing artifacts that may prohibit correct detection and characterization of defects. Therefore, in this work, we propose an improved implementation of the local wavenumber estimation algorithm for anisotropic media (ALWE). The working principle and efficiency of the proposed algorithm are illustrated on datasets coming from numerical simulations and experimental testing on anisotropic carbon fiber reinforced polymer (CFRP) laminates. We demonstrate that the proposed algorithm successfully compensates for the processing artifacts and enhances the overall quality of the wavenumber maps facilitating unequivocal damage detection.

Grain-wise piezoelectric response of polycrystalline aluminum nitride thin films: an empirical investigation using piezo force microscopy

Abstract Local mapping of piezoelectric response across grains with different crystal planes is crucial to enhance the level of understanding of the piezoelectric behavior of wurtzite hexagonal (WH) AlN thin films. To investigate this phenomenon, polycrystalline WH-AlN thin films were deposited by pulsed-DC magnetron sputtering using a mixture of argon (Ar) and nitrogen (N2) with varying N2 content as 5%, 10%, and 20% in the form of the AlN/Ti/Si (100) heterostructure. Piezo force microscopy (PFM) was used to establish a correlation among piezo-response (PR) phase contrast, polarity, and crystal planes constituting the film surface. A phenomenological model is also presented to correlate the distribution of interface sheet charges with polar (0002)-, semi-polar (10 1 ¯ 1)-, and non-polar (10 1 ¯ 0)-planes present on the film surface. This interdependence eventually regulates the phase between piezoelectric oscillations and modulation voltage, which is ultimately translated into a grain-wise contrast obtained from PR-phase images.

TC10 differently controls the dynamics of Exo70 in growth cones of cortical and hippocampal neurons

The exocyst is an octameric protein complex that acts as a tether for GOLGI-derived vesicles at the plasma membrane during exocytosis. It is involved in membrane expansion during axonal outgrowth. Exo70 is a major subunit of the exocyst complex and is controlled by TC10, a Rho family GTPase. How TC10 affects the dynamics of Exo70 at the plasma membrane is not well understood. There is also evidence that TC10 controls Exo70 dynamics differently in nonpolar cells and axons. To address this, we used super-resolution microscopy to study the spatially resolved effects of TC10 on Exo70 dynamics in HeLa cells and the growth cone of cortical and hippocampal neurons. We generated single particle localization and trajectory maps and extracted mean square displacements, diffusion coefficients, and alpha coefficients to characterize Exo70 diffusion. We found that the diffusivity of Exo70 was different in nonpolar cells and in the growth cone of neurons. TC10 stimulated the mobility of Exo70 in HeLa cells, but decreased the diffusion of Exo70 in the growth cone of cortical neurons. In contrast to cortical neurons, TC10 overexpression did not affect the mobility of Exo70 in the axonal growth cone of hippocampal neurons. These data suggest that mainly exocyst tethering in cortical neurons was under the control of TC10.

Robust Collaborative Visual-Inertial SLAM for Mobile Augmented Reality.

Achieving precise real-time localization and ensuring robustness are critical challenges in multi-user mobile AR applications. Leveraging collaborative information to augment tracking accuracy on lightweight devices and fortify overall system robustness emerges as a crucial necessity. In this paper, we propose a robust centralized collaborative rnulti-agent VI-SLAM system for mobile AR interaction and server-side efficient consistent mapping. The system deploys a lightweight VIO frontend on mobile devices for real-time tracking, and a backend running on a remote server to update multiple submaps. When overlapping areas between submaps across agents are detected, the system performs submap fusion to establish a globally consistent map. Additionally, we propose a map registration and fusion strategy based on covisibility areas for online registration and fusion in multi-agent scenarios. To improve the tracking accuracy of the frontend on agent, we introduce a strategy for updating the global map to the local map at a moderate frequency between the camera-rate pose estimation of the frontend VIO and the low-frequency global map optimization, using a tightly coupled strategy to achieve consistency of the multi-agent frontend poses estimation in the global map. The effectiveness of the proposed method is further confirmed by executing backend mapping on the server and deploying VIO frontends on multiple mobile devices for AR demostration. Additionally, we discuss the scalability of the proposed system by analyzing network traffic, synchronization frequency, and other factors at both the agent and server ends.

TLCSFI: A Pose-Guided Person Re-Identification Method with Two-Level Channel–Spatial Feature Integration

Person re-identification methods currently encounter challenges in feature learning, primarily due to difficulties in expressing the correlation between local features and integrating global and local features effectively. To address these issues, a pose-guided person re-identification method with Two-Level Channel–Spatial Feature Integration (TLCSFI) is proposed. In TLCSFI, a two-level integration mechanism is implemented. At the first level, TLCSFI integrates the spatial information from local features to generate fine-grained spatial features. At the second level, the fine-grained spatial feature and the coarse-grained channel feature are integrated together to complete channel–spatial feature integration. In the method, a Pose-based Spatial Feature Integration (PSFI) module is introduced to generate the pose union feature, which calculates intra-body affinity to guide the integration of spatial information among local pose feature maps. Then, a Channel and Spatial Union Feature Integration (CSUFI) module is proposed to efficiently integrate the channel information of the global feature and the spatial information of the pose union feature. Two individual networks are designed to extract channel and spatial information, respectively, in CSUFI, which are then weighted and integrated. Experiments are conducted on three publicly available datasets to evaluate TLCSFI, and the experimental results demonstrate its competitive performance.