相关参考文献
注意:仅列出部分参考文献,下载原文获取全部文献信息。
Article
Computer Science, Software Engineering
Zhongkang Lin et al.
Summary: This paper proposes a semantic segmentation network with multi-path structure, attention reweighting, and multi-scale encoding structure. By combining multi-scale information, high-level semantic information, and global context information, the network improves the accuracy of semantic segmentation while ensuring high computational efficiency.
Article
Computer Science, Artificial Intelligence
Qing Liu et al.
Summary: In this paper, a multi-stage context refinement network (MCRNet) is proposed for semantic segmentation. By constructing the Lowest-resolution Chain Context Aggregation (LCCA) module and the High-resolution Context Attention Refinement (HCAR) module, MCRNet can encode rich semantic information while preserving spatial details, resulting in improved image segmentation performance.
Article
Computer Science, Artificial Intelligence
Yongsheng Dong et al.
Summary: This paper proposes a field-matching attention network (FMANet) for object detection, which normalizes the receptive fields between features at different stages to the same scale and captures spatial information and details using spatial and channel attention mechanisms. Experimental results show that FMANet achieves competitive performance in object detection.
Article
Environmental Sciences
Yang Yang et al.
Summary: In this study, we proposed an attention-based multiscale max-pooling dense network (DMAU-Net) based on U-Net for ground object classification. The network incorporates a max-pooling module in the encoder part to enhance the quality of the feature map and an Efficient Channel Attention (ECA) module in the decoder part to strengthen effective features and suppress irrelevant information. Experimental results show that DMAU-Net effectively improves the accuracy of feature classification of high-resolution remote-sensing images.
Article
Computer Science, Artificial Intelligence
Yongsheng Dong et al.
Summary: In this paper, we propose a CartoonLossGAN based on cartoon loss for generating cartoon-style images. By reusing the encoder part of the discriminator and introducing a new loss function, the network can learn the smooth surface and coloring process of cartoon images, resulting in high-quality cartoon-style images. Furthermore, an initialization strategy is proposed to simplify and stabilize the model training.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2022)
Article
Computer Science, Artificial Intelligence
Qiang Wang et al.
Summary: This paper proposes a new unsupervised domain adaptation framework, CASA, for addressing the medical domain mismatch problem. The framework preserves the synergistic fusion of adaptation knowledge from the perspectives of appearance and semantic and utilizes Characterization Transfer Module (CTM) and Representation Transfer Module (RTM) to transform the appearance and features of medical lesions across domains. Experimental results demonstrate the superior performance of CASA in medical image segmentation.
Article
Environmental Sciences
Hong Wang et al.
Summary: This study discusses the efficient method of semantic segmentation using remote sensing images for agricultural crop classification and the challenges it faces. A novel architecture named CCTNet is proposed to address these challenges, along with two fusion modules and three effective methods aimed at improving classification accuracy and image completeness. Experimental results demonstrate that CCTNet outperforms single CNN or Transformer methods in terms of mean Intersection over Union (mIoU) scores, making it a competitive option for crop segmentation through remote sensing images.
Article
Engineering, Civil
Hai Wang et al.
Summary: Considerable progress has been made in semantic segmentation of images in favorable environments in recent years, but the environmental perception of autonomous driving under adverse weather conditions remains challenging. This paper aims to explore image segmentation in low-light scenarios to expand the application range of autonomous vehicles. We propose a novel nighttime segmentation framework and demonstrate its effectiveness through experiments.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2022)
Article
Computer Science, Interdisciplinary Applications
Pengcheng Li et al.
Summary: An accurate tooth identification and delineation method is proposed for dental CBCT images. A semantic graph-based approach is used to model the spatial associations between teeth and achieve precise delineation. The method demonstrates superior performance compared to other state-of-the-art approaches.
IEEE TRANSACTIONS ON MEDICAL IMAGING
(2022)
Article
Computer Science, Artificial Intelligence
Zechao Li et al.
Summary: This study proposes a novel Context-based Tandem Network (CTNet) that explores spatial and channel contextual information for semantic segmentation. The CTNet demonstrates superior performance by adaptively integrating the results of two context modules, leading to improved learning representations.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2022)
Article
Environmental Sciences
Yuqi Dai et al.
Summary: This article proposes a lightweight ViT-based terrain segmentation method, SegMarsViT, for planetary rover systems. The method utilizes the mobile vision transformer block to extract spatial and contextual information, and incorporates cross-scale feature fusion modules and a compact feature aggregation module for multi-level feature representation. Extensive experiments on three public datasets demonstrate the effectiveness and efficiency of the proposed method.
Article
Computer Science, Interdisciplinary Applications
Jiahuan Song et al.
Summary: This paper proposes a Global and Local Feature Reconstruction Network (GLFRNet) which efficiently captures global context features and restores spatial information through global feature reconstruction module and local feature reconstruction module, improving the performance of medical image segmentation.
IEEE TRANSACTIONS ON MEDICAL IMAGING
(2022)
Article
Environmental Sciences
Bo Liu et al.
Summary: This article introduces a new framework called PGNet for semantic segmentation of VHR remote sensing images. It effectively improves the segmentation results through the positioning guidance module and the self-multiscale collection module. Experimental results show that PGNet achieves higher mIoU scores compared to FactSeg, with an improvement of 1.49% and 2.40% on the iSAID dataset and ISPRS Vaihingn dataset, respectively.
Article
Automation & Control Systems
Xuelong Li et al.
Summary: The precise location information of road and lane lines is crucial for autonomous and assisted driving, but detection inaccuracies are common. To address this, an attention-based spatial segmentation network has been proposed to improve network understanding of spatial information, effectively enhancing the performance of traffic scene understanding.
IEEE TRANSACTIONS ON CYBERNETICS
(2022)
Article
Environmental Sciences
Yan Chen et al.
Summary: This study proposes a lightweight global context semantic segmentation network to improve the effectiveness of semantic segmentation of remote sensing images. By utilizing global context data and reducing model parameters, the proposed network better extracts global context information. Additionally, the use of a parallel channel spatial attention module and a multi-scale fusion module further enhances the model's performance.
Article
Computer Science, Artificial Intelligence
Andre de Souza Brito et al.
Summary: This paper presents a novel multi-pooling architecture generated by combining the advantages of wavelet and max-pooling operations in convolutional neural networks (CNNs) for semantic segmentation tasks. Experimental results show that this multi-pooling architecture can enhance the performance of aerial image segmentation tasks, achieving results comparable to state-of-the-art approaches.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Computer Science, Artificial Intelligence
Quan Tang et al.
Summary: The paper proposes a novel Chained Context Aggregation Module (CAM) to enrich feature representation by capturing multi-scale contexts, and constructs the Chained Context Aggregation Network (CANet) which achieves state-of-the-art or competitive performance on six challenging datasets.
IMAGE AND VISION COMPUTING
(2021)
Article
Environmental Sciences
Zhiyong Xu et al.
Summary: The paper proposes a HRCNet model based on HRNet to address the loss of spatial information and lack of global context information in conventional CNN-based semantic segmentation methods. The model integrates LDA and FEFP structures to fuse contextual information of different scales and BA module combined with BAloss function to achieve boundary information. Experimental results show significant improvement in boundary and segmentation performance on Potsdam and Vaihingen datasets, with overall accuracy scores increased up to 92.0% and 92.3% respectively.
Article
Computer Science, Artificial Intelligence
Changqian Yu et al.
Summary: Separating low-level details and high-level semantics is key to achieving high accuracy and efficiency in real-time semantic segmentation. The proposed architecture, called Bilateral Segmentation Network (BiSeNet V2), effectively handles feature representations through detail and semantics branches, striking a balance between speed and accuracy to outperform existing methods.
INTERNATIONAL JOURNAL OF COMPUTER VISION
(2021)
Article
Computer Science, Artificial Intelligence
Zhen Zhou et al.
Summary: This paper introduces a self-attention feature fusion network for semantic segmentation, which improves performance by introducing vertical and horizontal compression attention modules and unequal channel pyramid pooling modules. The proposed model achieves high performance on the PASCAL VOC2012 and Cityscapes datasets.
Proceedings Paper
Computer Science, Artificial Intelligence
Taehun Kim et al.
Summary: The paper introduces a method called Spatial Context Memoization (SpaM) to add a bypassing branch in semantic segmentation networks, retaining input dimension and continuously communicating spatial context and semantic information with the backbone network. The study also proposes Meshgrid Atrous Convolution Consensus (MetroCon(2)) to address misalignment issues in multi-scale context schemes.
2021 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP)
(2021)
Article
Computer Science, Artificial Intelligence
Qichuan Geng et al.
Summary: The proposed GPSNet method achieves good performance in semantic segmentation tasks by dynamically selecting receptive fields and aggregating dense semantic context information.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2021)
Article
Robotics
Ping Hu et al.
Summary: This paper introduces a novel deep CNN architecture for semantic segmentation of high-resolution images and videos, achieving state-of-the-art performance with the use of fast spatial attention and additional spatial reduction. Experimental results demonstrate superior accuracy and speed compared to existing approaches.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2021)
Article
Computer Science, Artificial Intelligence
Jun Fu et al.
PATTERN RECOGNITION
(2020)
Article
Computer Science, Artificial Intelligence
Henghui Ding et al.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2020)
Proceedings Paper
Computer Science, Artificial Intelligence
Hang Zhang et al.
2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2019)
(2019)
Proceedings Paper
Computer Science, Artificial Intelligence
Yizhou Zhou et al.
2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2019)
(2019)
Review
Computer Science, Artificial Intelligence
Alberto Garcia-Garcia et al.
APPLIED SOFT COMPUTING
(2018)
Article
Computer Science, Artificial Intelligence
Vijay Badrinarayanan et al.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2017)
Article
Computer Science, Hardware & Architecture
Alex Krizhevsky et al.
COMMUNICATIONS OF THE ACM
(2017)