Search Results (14)

Oriented object detection in remote sensing images is a particularly challenging task, especially when it involves detecting tiny, densely arranged, or occluded objects. Moreover, such remote sensing images are often susceptible to noise, which significantly increases the difficulty of the task. To address these challenges, we introduce the Wavelet-Domain Adaptive Receptive Field Network (WDARFNet), a novel architecture that combines Convolutional Neural Networks (CNNs) with Discrete Wavelet Transform (DWT) to enhance feature extraction and noise robustness. WDARFNet employs DWT to decompose feature maps into four distinct frequency components. Through ablation experiments, we demonstrate that selectively combining specific high-frequency and low-frequency features enhances the network’s representational capacity. Discarding diagonal high-frequency features, which contain significant noise, further enhances the model’s noise robustness. In addition, to capture long-range contextual information and adapt to varying object sizes and occlusions, WDARFNet incorporates a selective kernel mechanism. This strategy dynamically adjusts the receptive field based on the varying shapes of objects, ensuring optimal feature extraction for diverse objects. The streamlined and efficient WDARFNet achieves state-of-the-art performance on three challenging remote sensing object detection benchmarks: DOTA-v1.0, DIOR-R, and HRSC2016. Full article

Accurate trajectory prediction of surrounding vehicles is a fundamental challenge in autonomous driving, requiring sophisticated modeling of complex vehicle interactions, traffic dynamics, and contextual dependencies. This paper introduces Adaptive Speed-Aware Prediction (ADSAP), a novel trajectory prediction framework that advances the state of the art through innovative mechanisms for adaptive attention modulation and knowledge transfer. At its core, ADSAP employs an adaptive deformable speed-aware pooling mechanism that dynamically adjusts the model’s attention distribution and receptive field based on instantaneous vehicle states and interaction patterns. This adaptive architecture enables fine-grained modeling of diverse traffic scenarios, from sparse highway conditions to dense urban environments. The framework incorporates a sophisticated speed-aware multi-scale feature aggregation module that systematically combines spatial and temporal information across multiple scales, facilitating comprehensive scene understanding and robust trajectory prediction. To bridge the gap between model complexity and computational efficiency, we propose an adversarial knowledge distillation approach that effectively transfers learned representations and decision-making strategies from a high-capacity teacher model to a lightweight student model. This novel distillation mechanism preserves prediction accuracy while significantly reducing computational overhead, making the framework suitable for real-world deployment. Extensive empirical evaluation on the large-scale NGSIM and highD naturalistic driving datasets demonstrates ADSAP’s superior performance. The ADSAP framework achieves an 18.7% reduction in average displacement error and a 22.4% improvement in final displacement error compared to state-of-the-art methods while maintaining consistent performance across varying traffic densities (0.05–0.85 vehicles/meter) and speed ranges (0–35 m/s). Moreover, ADSAP exhibits robust generalization capabilities across different driving scenarios and weather conditions, with the lightweight student model achieving 95% of the teacher model’s accuracy while offering a 3.2× reduction in inference time. Comprehensive experimental results supported by detailed ablation studies and statistical analyses validate ADSAP’s effectiveness in addressing the trajectory prediction challenge. Our framework provides a novel perspective on integrating adaptive attention mechanisms with efficient knowledge transfer, contributing to the development of more reliable and intelligent autonomous driving systems. Significant improvements in prediction accuracy, computational efficiency, and generalization capability demonstrate ADSAP’s potential ability to advance autonomous driving technology. Full article

This paper addresses the challenge of small-object detection in traffic surveillance by proposing a hybrid network architecture that combines attention mechanisms with convolutional layers. The network introduces an innovative attention mechanism into the YOLOv8 backbone, which effectively enhances the detection accuracy and robustness of small objects through fine-grained and coarse-grained attention routing on feature maps. During the feature fusion stage, we employ adaptive dilated convolution, which dynamically adjusts the dilation rate spatially based on frequency components. This adaptive convolution kernel helps preserve the details of small objects while strengthening their feature representation. It also expands the receptive field, which is beneficial for capturing contextual information and the overall features of small objects. Our method demonstrates an improvement in Average Precision (AP) by 1% on the UA-DETRAC-test dataset and 3% on the VisDrone-test dataset when compared to state-of-the-art methods. The experiments indicate that the new architecture achieves significant performance improvements across various evaluation metrics. To fully leverage the potential of our approach, we conducted extended research on radar–camera systems. Full article

The YOLO series of object detection methods have achieved significant success in a wide range of computer vision tasks due to their efficiency and accuracy. However, detecting small objects in UAV images remains a formidable challenge due to factors such as a low resolution, complex background interference, and significant scale variations, which collectively degrade the quality of feature extraction and limit detection performance. To address these challenges, we propose the receptive field attention-guided YOLO (RFAG-YOLO) method, an advanced adaptation of YOLOv8 tailored for small-object detection in UAV imagery, with a focus on improving feature representation and detection robustness. To this end, we introduce a novel network building block, termed the receptive field network block (RFN block), which leverages dynamic kernel parameter adjustments to enhance the model’s ability to capture fine-grained local details. To effectively harness multi-scale features, we designed an enhanced FasterNet module based on RFN blocks as the core component of the backbone network in RFAG-YOLO, enabling robust feature extraction across varying resolutions. This approach achieves a balance of semantic information by employing staged downsampling and a hierarchical arrangement of RFN blocks, ensuring optimal feature representation across different resolutions. Additionally, we introduced a Scale-Aware Feature Amalgamation (SAF) component prior to the detection head of RFAG-YOLO. This component employs a scale attention mechanism to dynamically weight features from both higher and lower layers, facilitating richer information flow and significantly improving the model’s robustness to complex backgrounds and scale variations. Experimental results on the VisDrone2019 dataset demonstrated that RFAG-YOLO outperformed state-of-the-art models, including YOLOv7, YOLOv8, YOLOv10, and YOLOv11, in terms of detection accuracy and efficiency. In particular, RFAG-YOLO achieved an mAP50 of 38.9%, representing substantial improvements over multiple baseline models: a 12.43% increase over YOLOv7, a 5.99% improvement over YOLOv10, and significant gains of 16.12% compared to YOLOv8n and YOLOv11. Moreover, compared to the larger YOLOv8s model, RFAG-YOLO achieved 97.98% of its mAP50 performance while utilizing only 53.51% of the parameters, highlighting its exceptional efficiency in terms of the performance-to-parameter ratio and making it highly suitable for resource-constrained UAV applications. These results underscore the substantial potential of RFAG-YOLO for real-world UAV applications, particularly in scenarios demanding accurate detection of small objects under challenging conditions such as varying lighting, complex backgrounds, and diverse scales. Full article

Satellites frequently encounter atmospheric haze during imaging, leading to the loss of detailed information in remote sensing images and significantly compromising image quality. This detailed information is crucial for applications such as Earth observation and environmental monitoring. In response to the above issues, this paper proposes an end-to-end multi-scale adaptive feature extraction method for remote sensing image dehazing (MSD-Net). In our network model, we introduce a dilated convolution adaptive module to extract global and local detail features of remote sensing images. The design of this module can extract important image features at different scales. By expanding convolution, the receptive field is expanded to capture broader contextual information, thereby obtaining a more global feature representation. At the same time, a self-adaptive attention mechanism is also used, allowing the module to automatically adjust the size of its receptive field based on image content. In this way, important features suitable for different scales can be flexibly extracted to better adapt to the changes in details in remote sensing images. To fully utilize the features at different scales, we also adopted feature fusion technology. By fusing features from different scales and integrating information from different scales, more accurate and rich feature representations can be obtained. This process aids in retrieving lost detailed information from remote sensing images, thereby enhancing the overall image quality. A large number of experiments were conducted on the HRRSD and RICE datasets, and the results showed that our proposed method can better restore the original details and texture information of remote sensing images in the field of dehazing and is superior to current state-of-the-art methods. Full article

With the prosperous development and successful application of deep learning technologies in the field of remote sensing, numerous deep-learning-based methods have emerged for synthetic aperture radar (SAR) automatic target recognition (ATR) tasks over the past few years. Generally, most deep-learning-based methods can achieve outstanding recognition performance on the condition that an abundance of labeled samples are available to train the model. However, in real application scenarios, it is difficult and costly to acquire and to annotate abundant SAR images due to the imaging mechanism of SAR, which poses a big challenge to existing SAR ATR methods. Therefore, SAR target recognition in the situation of few-shot, where only a scarce few labeled samples are available, is a fundamental problem that needs to be solved. In this paper, a new method named enhanced prototypical network with customized region-aware convolution (CRCEPN) is put forward to specially tackle the few-shot SAR ATR tasks. To be specific, a feature-extraction network based on a customized and region-aware convolution is first developed. This network can adaptively adjust convolutional kernels and their receptive fields according to each SAR image’s own characteristics as well as the semantical similarity among spatial regions, thus augmenting its capability to extract more informative and discriminative features. To achieve accurate and robust target identity prediction under the few-shot condition, an enhanced prototypical network is proposed. This network can improve the representation ability of the class prototype by properly making use of training and test samples together, thus effectively raising the classification accuracy. Meanwhile, a new hybrid loss is designed to learn a feature space with both inter-class separability and intra-class tightness as much as possible, which can further upgrade the recognition performance of the proposed method. Experiments performed on the moving and stationary target acquisition and recognition (MSTAR) dataset, the OpenSARShip dataset, and the SAMPLE+ dataset demonstrate that the proposed method is competitive with some state-of-the-art methods for few-shot SAR ATR tasks. Full article

To address the challenges of excessive model parameters and low detection accuracy in printed circuit board (PCB) defect detection, this paper proposes a novel PCB defect detection model based on the improved RTDETR (Real-Time Detection, Embedding and Tracking) method, named MFAD-RTDETR. Specifically, the proposed model introduces the designed Detail Feature Retainer (DFR) into the original RTDETR backbone to capture and retain local details. Subsequently, based on the Mamba architecture, the Visual State Space (VSS) module is integrated to enhance global attention while reducing the original quadratic complexity to a linear level. Furthermore, by exploiting the deformable attention mechanism, which dynamically adjusts reference points, the model achieves precise localization of target defects and improves the accuracy of the transformer in complex visual tasks. Meanwhile, a receptive field synthesis mechanism is incorporated to enrich multi-scale semantic information and reduce parameter complexity. In addition, the scheme proposes a novel Multi-frequency Aggregation and Diffusion feature composite paradigm (MFAD-feature composite paradigm), which consists of the Aggregation Diffusion Fusion (ADF) module and the Refiner Feature Composition (RFC) module. It aims to strengthen features with fine-grained awareness while preserving a certain level of global attention. Finally, the Wise IoU (WIoU) dynamic nonmonotonic focusing mechanism is used to reduce competition among high-quality anchor boxes and mitigate the effects of the harmful gradients from low-quality examples, thereby concentrating on anchor boxes of average quality to promote the overall performance of the detector. Extensive experiments are conducted on the PCB defect dataset released by Peking University to validate the effectiveness of the proposed model. The experimental results show that our approach achieves the 97.0% and 51.0% performance in mean Average Precision (mAP)@0.5 and mAP@0.5:0.95, respectively, which significantly outperforms the original RTDETR. Moreover, the model reduces the number of parameters by approximately 18.2% compared to the original RTDETR. Full article

Similar to convolutional neural networks for image processing, existing analysis methods for 3D point clouds often require the designation of a local neighborhood to describe the local features of the point cloud. This local neighborhood is typically manually specified, which makes it impossible for the network to dynamically adjust the receptive field’s range. If the range is too large, it tends to overlook local details, and if it is too small, it cannot establish global dependencies. To address this issue, we introduce in this paper a new concept: receptive field space (RFS). With a minor computational cost, we extract features from multiple consecutive receptive field ranges to form this new receptive field space. On this basis, we further propose a receptive field space attention mechanism, enabling the network to adaptively select the most effective receptive field range from RFS, thus equipping the network with the ability to adjust granularity adaptively. Our approach achieved state-of-the-art performance in both point cloud classification, with an overall accuracy (OA) of 94.2%, and part segmentation, achieving an mIoU of 86.0%, demonstrating the effectiveness of our method. Full article

High-precision landcover classification is a fundamental prerequisite for resource and environmental monitoring and land-use status surveys. Imbued with intricate spatial information and texture features, very high spatial resolution remote sensing images accentuate the divergence between features within the same category, thereby amplifying the complexity of landcover classification. Consequently, semantic segmentation models leveraging deep backbone networks have emerged as stalwarts in landcover classification tasks owing to their adeptness in feature representation. However, the classification efficacy of a solitary backbone network model fluctuates across diverse scenarios and datasets, posing a persistent challenge in the construction or selection of an appropriate backbone network for distinct classification tasks. To elevate the classification performance and bolster the generalization of semantic segmentation models, we propose a novel semantic segmentation network architecture, named the aggregated backbone network (ABNet), for the meticulous landcover classification. ABNet aggregates three prevailing backbone networks (ResNet, HRNet, and VoVNet), distinguished by significant structural disparities, using a same-stage fusion approach. Subsequently, it amalgamates these networks with the Deeplabv3+ head after integrating the convolutional block attention mechanism (CBAM). Notably, this amalgamation harmonizes distinct scale features extracted by the three backbone networks, thus enriching the model’s spatial contextual comprehension and expanding its receptive field, thereby facilitating more effective semantic feature extraction across different stages. The convolutional block attention mechanism primarily orchestrates channel adjustments and curtails redundant information within the aggregated feature layers. Ablation experiments demonstrate an enhancement of no less than 3% in the mean intersection over union (mIoU) of ABNet on both the LoveDA and GID15 datasets when compared with a single backbone network model. Furthermore, in contrast to seven classical or state-of-the-art models (UNet, FPN, PSPNet, DANet, CBNet, CCNet, and UPerNet), ABNet evinces excellent segmentation performance across the aforementioned datasets, underscoring the efficiency and robust generalization capabilities of the proposed approach. Full article

X-ray package security check systems are widely used in public places, but they face difficulties in accurately detecting prohibited items due to the stacking and diversity of shapes of the objects inside the luggage, posing a threat to personal safety in public places. The existing methods for X-ray image object detection suffer from low accuracy and poor generalization, mainly due to the lack of large-scale and high-quality datasets. To address this gap, a novel large-scale X-ray image dataset for object detection, LSIray, is provided, consisting of high-quality X-ray images of luggage and objects of 21 types and sizes. LSIray covers some common categories that were neglected in previous research. The dataset provides more realistic and rich data resources for X-ray image object detection. To address the problem of poor security inspection, an improved model based on YOLOv8 is proposed, named SC- YOLOv8, consisting of two new modules: CSPnet Deformable Convolution Network Module (C2F_DCN) and Spatial Pyramid Multi-Head Attention Module (SPMA). C2F_DCN uses deformable convolution, which can adaptively adjust the position and shape of the receptive field to accommodate the diversity of targets. SPMA adopts the spatial pyramid head attention mechanism, which can utilize feature information from different scales and perspectives to enhance the representation ability of targets. The proposed method is evaluated through extensive experiments using the LSIray dataset and comparisons with the existing methods. The results show that the method surpasses the state-of-the-art methods on various indicators. Experimenting using the LSIray dataset and the OPIXray dataset, our SC-YOlOv8 model achieves 82.7% and 89.2% detection accuracies, compared to the YOLOv8 model, which is an improvement of 1.4% and 1.2%, respectively. The work not only provides valuable data resources, but also offers a novel and effective solution for the X-ray image security check problem. Full article

In this article, we propose an effective grasp detection network based on an improved deformable convolution and spatial feature center mechanism (DCSFC-Grasp) to precisely grasp unidentified objects. DCSFC-Grasp includes three key procedures as follows. First, improved deformable convolution is introduced to adaptively adjust receptive fields for multiscale feature information extraction. Then, an efficient spatial feature center (SFC) layer is explored to capture the global remote dependencies through a lightweight multilayer perceptron (MLP) architecture. Furthermore, a learnable feature center (LFC) mechanism is reported to gather local regional features and preserve the local corner region. Finally, a lightweight CARAFE operator is developed to upsample the features. Experimental results show that DCSFC-Grasp achieves a high accuracy (99.3% and 96.1% for the Cornell and Jacquard grasp datasets, respectively) and even outperforms the existing state-of-the-art grasp detection models. The results of real-world experiments on the six-DoF Realman RM65 robotic arm further demonstrate that our DCSFC-Grasp is effective and robust for the grasping of unknown targets. Full article

Endometriosis and autoimmune diseases share a hyper-inflammatory state that might negatively impact the embryo–endometrium crosstalk. Inflammatory and immune deregulatory mechanisms have been shown to impair both endometrial receptivity and embryo competence at the implantation site. The aim of this study was to investigate the potential additional impact of co-existing autoimmunity in women affected by endometriosis on the early stages of reproduction. This was a retrospective, multicenter case-control study enrolling N = 600 women with endometriosis who underwent in vitro fertilization–embryo transfer cycles between 2007 and 2021. Cases were women with endometriosis and concomitant autoimmunity matched based on age and body mass index to controls with endometriosis only in a 1:3 ratio. The primary outcome was the cumulative clinical pregnancy rate (cCPR). The study found significantly lower cleavage (p = 0.042) and implantation (p = 0.029) rates among cases. Autoimmunity (p = 0.018), age (p = 0.007), and expected poor response (p = 0.014) were significant negative predictors of cCPR, with an adjusted odds ratio of 0.54 (95% CI, 0.33–0.90) for autoimmunity. These results suggest that the presence of concomitant autoimmunity in endometriosis has a significant additive negative impact on embryo implantation. This effect might be due to several immunological and inflammatory mechanisms that interfere with both endometrial receptivity and embryo development and deserves further consideration. Full article

Energy harvesting wireless sensor network (EH-WSN) is considered to be one of the key enabling technologies for the internet of things (IoT) construction. Although the introduced EH technology can alleviate the energy limitation problem that occurs in the traditional wireless sensor network (WSN), most of the current studies on EH-WSN fail to adequately consider the relationship between energy state and data buffer constraint, and thereby they do not address well the issues of energy efficiency and long end-to-end delay. In view of the above problems, a brand new greedy strategy-based energy-efficient routing protocol is proposed in this paper. Firstly, in the system modeling process, we construct an energy evaluation model, which comprehensively considers the energy harvesting, energy consumption and energy classification factors, to identify the energy state of node. Then, we establish a channel feature-based communication range judgment model to determine the transmission area of nodes. Combining these two models, a reception state adjustment mechanism is designed. It takes the buffer occupancy and the MAC layer protocol into account to adjust the data reception state of nodes. On this basis, we propose a greedy strategy-based routing algorithm. In addition, we also analyze the correctness and computational complexity of the proposed algorithm. Finally, we conduct extensive simulation experiments to show that our algorithm achieves optimum performance in energy consumption, packet delivery ratio, average hop count and end-to-end delay and acceptable performance in energy variance. Full article

The detection of arbitrary-oriented and multi-scale objects in satellite optical imagery is an important task in remote sensing and computer vision. Despite significant research efforts, such detection remains largely unsolved due to the diversity of patterns in orientation, scale, aspect ratio, and visual appearance; the dense distribution of objects; and extreme imbalances in categories. In this paper, we propose an adaptive dynamic refined single-stage transformer detector to address the aforementioned challenges, aiming to achieve high recall and speed. Our detector realizes rotated object detection with RetinaNet as the baseline. Firstly, we propose a feature pyramid transformer (FPT) to enhance feature extraction of the rotated object detection framework through a feature interaction mechanism. This is beneficial for the detection of objects with diverse patterns in terms of scale, aspect ratio, visual appearance, and dense distributions. Secondly, we design two special post-processing steps for rotated objects with arbitrary orientations, large aspect ratios and dense distributions. The output features of FPT are fed into post-processing steps. In the first step, it performs the preliminary regression of locations and angle anchors for the refinement step. In the refinement step, it performs adaptive feature refinement first and then gives the final object detection result precisely. The main architecture of the refinement step is dynamic feature refinement (DFR), which is proposed to adaptively adjust the feature map and reconstruct a new feature map for arbitrary-oriented object detection to alleviate the mismatches between rotated bounding boxes and axis-aligned receptive fields. Thirdly, the focus loss is adopted to deal with the category imbalance problem. Experiments on two challenging satellite optical imagery public datasets, DOTA and HRSC2016, demonstrate that the proposed ADT-Det detector achieves a state-of-the-art detection accuracy (79.95% mAP for DOTA and 93.47% mAP for HRSC2016) while running very fast (14.6 fps with a 600 × 600 input image size). Full article