A New Hybrid ConvViT Model for Dangerous Farm Insect Detection by Anil Utku, Mahmut Kaya, Yavuz Canbay

“…This study proposes a novel hybrid convolution and vision transformer model (ConvViT) designed to detect harmful insect species that adversely affect agricultural production and play a critical role in global food security. …”
Get full text

A Malware Classification Method Based on Knowledge Distillation and Feature Fusion by Xin Guan, Guodong Zhang

“…This approach incorporates image texture features with enhanced Local Binary Pattern (LBP), providing insights into the local structure and layout of images and aiding the model in better understanding image details and internal structure, thus enhancing classification performance. …”
Get full text

TMAR: 3-D Transformer Network via Masked Autoencoder Regularization for Hyperspectral Sharpening by Zeinab Dehghan, Jingxiang Yang, Mehran Yazdi, Abdolraheem Khader, Liang Xiao

“…In this study, we focus on leveraging the power of CNN and transformer models and propose a multistage deep transformer-based super-resolution network that is regularized via an asymmetric autoencoder structure. In addition, we utilize a 3-D convolution layer in the light transformer structure because it allows for more flexible computation of correlations between HSI layers and better capturing of dependencies within spectral–spatial features. …”
Get full text

Generation driven understanding of localized 3D scenes with 3D diffusion model by Hao Sun, Junping Qin, Zheng Liu, Xinglong Jia, Kai Yan, Lei Wang, Zhiqiang Liu, Shaofei Gong

“…However, the existing diffusion models primarily focus on the global structure and are constrained by predefined dataset categories, which are unable to accurately resolve the detailed structure of complex 3D scenes. …”
Get full text

GIVTED-Net: GhostNet-Mobile Involution ViT Encoder-Decoder Network for Lightweight Medical Image Segmentation by Resha Dwika Hefni Al-Fahsi, Ahmad Naghim Fauzaini Prawirosoenoto, Hanung Adi Nugroho, Igi Ardiyanto

“…Nevertheless, conventional CNN layers, such as convolution and pooling, demonstrate a spatial inductive bias that constrains their ability to instantly capture global context information. …”
Get full text

Edge-aware joint neural denoising and normal estimation for mobile and handheld laser point clouds by T. Zhang, S. Filin

“…To extract contextual information it introduces densely packed graph convolution layers and a global attention mechanism. …”
Get full text

DEFIF-Net: A lightweight dual-encoding feature interaction fusion network for medical image segmentation. by Zhanlin Ji, Shengnan Hao, Quanming Zhao, Zidong Yu, Hongjiu Liu, Lei Li, Ivan Ganchev

“…Firstly, in the encoding stage of DEFIF-Net, a global dependency fusion branch is introduced as an additional encoder to capture distant feature dependencies, whereby the neighboring and distant feature dependencies are effectively integrated by the newly designed feature interaction fusion convolution. …”
Get full text

3D-SCUMamba: An Abdominal Tumor Segmentation Model by Juwita, Ghulam Mubashar Hassan, Amitava Datta

“…Existing deep learning models typically adopt encoder-decoder architectures integrating convolutional layers with global dependency modeling to capture broader contextual information around tumors. …”
Get full text

YOLO-DAFS: A Composite-Enhanced Underwater Object Detection Algorithm by Shengfu Luo, Chao Dong, Guixin Dong, Rongmin Chen, Bing Zheng, Ming Xiang, Peng Zhang, Zhanwei Li

“…The backbone incorporates a DualBottleneck module to enhance feature extraction, replacing the standard bottleneck structure in C3k, thus enhancing the feature extraction and the channel aggregation. …”
Get full text

Short-term rainfall prediction based on radar echo using an efficient spatio-temporal recurrent unit by Dali Wu, Shunli Zhang, Guohong Zhao, Yongchao Feng, Yuan Ma, Yue Zhang

“…The combined effect of the Self-Attention (SA) mechanism and convolution allows the model to focus on both global and local dependencies in spatial information, improving the clarity of the generated images. …”
Get full text

A lightweight steel surface defect detection network based on YOLOv9 by Tianyi Zheng, Ling Yu, Yongbao Shi, Fanglin Niu

“…Next, we replace the regular convolution blocks in the model network with spatial-to-depth convolutions, further reducing the model’s computational complexity while retaining global feature information. …”
Get full text

Reliable Multistate RRAM Devices for Reconfigurable CAM and IMC Applications by Shengpeng Xing, Zijian Wang, Zhen Wang, Pengtao Li, Xuemeng Fan, Ziyang Zhang, Guobin Zhang, Jianhao Kan, Qi Luo, Shuai Zhong, Yishu Zhang

Get full text

SSATNet: Spectral-spatial attention transformer for hyperspectral corn image classification by Bin Wang, Gongchao Chen, Juan Wen, Linfang Li, Songlin Jin, Yan Li, Ling Zhou, Weidong Zhang

“…Specifically, SSATNet utilizes 3D and 2D convolutions to effectively extract local spatial, spectral, and textural features from the data while incorporating spectral and spatial morphological structures to understand the internal structure of the data better. …”
Get full text

Fiber Visualization with LIC Maps Using Multidirectional Anisotropic Glyph Samples by Mark Höller, Kay-M. Otto, Uwe Klose, Samuel Groeschel, Hans-H. Ehricke

“…Line integral convolution (LIC) is used as a texture-based technique in computer graphics for flow field visualization. …”
Get full text

DSCONV-GAN: a UAV-BASED model for Verticillium Wilt disease detection in Chinese cabbage in complex growing environments by Jun Zhang, Dongfang Zhang, Jingyan Liu, Yuhong Zhou, Xiaoshuo Cui, Xiaofei Fan

“…Based on YOLOv8, with the addition of the dynamic snake convolution (DSConv) module and the improved loss function maximum possible distance intersection-over-union (MPDIoU), we acquired enhanced complex structures and global characteristics in Chinese cabbage images under different growth conditions. …”
Get full text

Path planning of intelligent tennis ball picking robot integrating twin network target tracking algorithm by Zegang Wang

“…Additionally, the Transformer structure improves tracking accuracy by capturing the global context relationship. …”
Get full text

Improvement in Pavement Defect Scenarios Using an Improved YOLOv10 with ECA Attention, RefConv and WIoU by Xiaolin Zhang, Lei Lu, Hanyun Luo, Lei Wang

“…The RefConv dual-branch structure achieves feature complementarity between local details and global context (mAP increased by 2.1%), the ECA mechanism models channel relationships using 1D convolution (small-object recall rate increased by 27%), and the WIoU loss optimizes difficult sample regression through a dynamic weighting mechanism (location accuracy improved by 37%). …”
Get full text

Foreign object detection on coal conveyor belt enhanced by attention mechanism by ZHANG Yang, CHENG Zhiyu, CHEN Yunjiang, ZHANG Jiannan, YUAN Wensheng, ZHANG Hui

“…A unique combination of convolution and pooling operations was used by the CPCA attention mechanism to perform global average pooling and maximum pooling on the input feature map, multi-dimensional feature information was deeply mined, and then attention weights for each channel and spatial position were accurately generated through nonlinear transformation, guiding the model to focus on the key feature areas of foreign objects and enhance feature extraction capabilities. …”
Get full text

RETINA: Reconstruction-based pre-trained enhanced TransUNet for electron microscopy segmentation on the CEM500K dataset. by Cheng Xing, Ronald Xie, Gary D Bader

“…We developed the RETINA method, which combines pre-training on the large, unlabeled CEM500K EM image dataset with a hybrid neural-network model architecture that integrates both local (convolutional layer) and global (transformer layer) image processing to learn from manual image annotations. …”
Get full text

Attention-enhanced StrongSORT for robust vehicle tracking in complex environments by Wei Xu, Xiaodong Du, Ruochen Li, Bingjie Li, Yuhu Jiao, Lei Xing

“…To address these challenges, we propose AE-StrongSORT (Attention-Enhanced StrongSORT), an attention-enhanced tracking framework featuring three systematic innovations: first, the GAM-YOLO (global attention mechanism-YOLO)hybrid architecture integrates multi-scale feature fusion with a global attention mechanism (GC2f structure). …”
Get full text