Article
Chemistry, Analytical
Kangwei Liu, Tianchi Qiu, Yinfeng Yu, Songlin Li, Xiuhong Li
Summary: This paper proposes a novel multi-level feature integration network (MFNet) for camouflaged object detection. The network incorporates an edge guidance module (EGM) to provide additional boundary semantic information and a multi-level feature integration module (MFIM) to fuse global and local features. A context aggregation refinement module (CARM) is also proposed to refine the prediction maps. Experimental results show the effectiveness of the MFNet model in camouflaged object detection.
Article
Remote Sensing
Zhenrong Zhuang, Wenzao Shi, Wenting Sun, Pengyu Wen, Lei Wang, Weiqi Yang, Tian Li
Summary: Change detection in remote sensing images has a significant impact on various applications. Recent advances have been made in change detection methods for different ground objects, but there are still limitations in feature recognition, resulting in unclear boundaries and a need for improved accuracy. To address these issues, the use of HRNet and new data augmentation methods are introduced to enhance accuracy. Additionally, the integration of CSWin and HRNet models improves performance, and a feature fusion network named A-FPN is designed to enhance perception of ground objects at different scales.
INTERNATIONAL JOURNAL OF REMOTE SENSING
(2023)
Article
Physics, Multidisciplinary
Hailu Yang, Qian Liu, Jin Zhang, Xiaoyu Ding, Chen Chen, Lili Wang
Summary: This paper proposes a multi-view integration method for community detection in semantic social network, which outperforms traditional methods and multi-view clustering algorithms in semantic information analysis.
Article
Environmental Sciences
Jinquan Lu, Xiangchao Meng, Qiang Liu, Zhiyong Lv, Gang Yang, Weiwei Sun, Wei Jin
Summary: This paper proposes a hierarchical feature association and global correction network (HFA-GCN) for change detection, aiming to address the issues of feature association and network accuracy in deep learning methods. Experiments show that the proposed method outperforms existing change detection models on publicly available databases.
Article
Environmental Sciences
Junkang Xue, Hao Xu, Hui Yang, Biao Wang, Penghai Wu, Jaewan Choi, Lixiao Cai, Yanlan Wu
Summary: This research proposes a multi-branched network structure to effectively fuse semantic information of building changes at different levels, achieving high accuracies. The model features auxiliary branches and a feature enhancement layer, enhancing the identification and merging of building changes.
Article
Computer Science, Information Systems
Samia Zouaoui, Khaled Rezeg
Summary: This paper proposes a novel approach called Multi-agents Indexing System to address plagiarism in Arabic documents. The system consists of three phases: natural language processing, indexing, and evaluation. The results show that the proposed system improves the performance of plagiarism detection in Arabic documents with semantic indexing and multi-agents system.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2022)
Article
Environmental Sciences
Jian Xu, Yuchun Huang, Dakan Ying
Summary: The study uses YOLOv5 as a single classification detector for traffic sign localization, and proposes a hierarchical classification model (HCM) for specific classification, which reduces class imbalance significantly. To address the limitations of a single image, a training-free multi-frame information integration module (MIM) is constructed to extract the detection sequence of traffic signs based on the embedding generated by HCM. Experimental results demonstrate that the improved HCM-YOLOv5 achieves a 79.0 mAP on two publicly available datasets, TT100K and ONCE, exceeding the state-of-the-art methods, with an inference speed of 22.7 FPS. Moreover, MIM further enhances model performance by integrating multi-frame information with only a slight increase in computational resource consumption.
Article
Biochemical Research Methods
Zhong Li, Kaiyancheng Jiang, Shengwei Qin, Yijun Zhong, Arne Elofsson
Summary: The study demonstrated the importance of miRNA-disease associations in understanding pathogenicity and developing treatments. The newly proposed GCSENet model outperformed existing methods in accurately predicting miRNA-disease associations. This method showed promising results in cross-validation and case studies for various diseases.
PLOS COMPUTATIONAL BIOLOGY
(2021)
Article
Environmental Sciences
Yanpeng Zhou, Jinjie Wang, Jianli Ding, Bohua Liu, Nan Weng, Hongzhi Xiao
Summary: Detecting changes in urban areas is challenging due to complex features, fast-changing rates, and human-induced interference. This paper proposes SIGNet, a Siamese graph convolutional network, to address these challenges and improve the accuracy of urban multi-class change detection tasks. SIGNet combines joint pyramidal upsampling with graph convolution-based graph reasoning and graph cross-attention methods to capture contextual relationships and semantic correlations between different regions and categories. Experimental results demonstrate that SIGNet achieves state-of-the-art accuracy on various MCD datasets. Furthermore, a new well-labeled dataset, CNAM-CD, containing 2508 pairs of high-resolution images is introduced to the MCD domain.
Article
Computer Science, Information Systems
Shunxin Guo, Hong Zhao, Wenyuan Yang
Summary: Hierarchical feature selection method with a multi-granularity clustering structure proposed in this paper effectively alleviates the semantic gap problem and outperforms several state-of-the-art approaches in extensive experiments.
INFORMATION SCIENCES
(2021)
Article
Computer Science, Artificial Intelligence
Vishnu VandanaKolisetty, Dharmendra Singh Rajput
Summary: This study proposes an integration and classification method for big data, which trains a model to understand the relationships and dependencies between data and use this knowledge to accurately map and classify incoming data. Experimental results demonstrate the efficiency and improved prediction accuracy of this method.
COMPLEX & INTELLIGENT SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Tianling Jiang, Zefan Zhang, Xin Li, Yi Ji, Chunping Liu
Summary: In this work, a novel method called Multi-View Semantic Understanding for Visual Dialog (MVSU) is proposed to resolve the visual coreference resolution problem. The model consists of two main textual processing modules, SRR and CRoT. Experimental results demonstrate that MVSU enhances the ability to understand the semantical information in the VisDial v1.0 dataset.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Multidisciplinary Sciences
Dongen Guo, Zechen Wu, Jiangfan Feng, Tao Zou
Summary: In this paper, a novel multi-scale semantic enhancement feature pyramid network (MSE-FPN) is proposed to address the semantic gap and feature aliasing issues in feature fusion for object detection. The MSE-FPN consists of three effective modules: semantic enhancement, semantic injection, and gated channel guidance, which improve the performance of FPN-based detectors.
SCIENTIFIC REPORTS
(2023)
Article
Computer Science, Information Systems
Jiexian Zeng, Huan Ouyang, Min Liu, Lu Leng, Xiang Fu
Summary: This paper proposes a multi-scale instance segmentation method that improves network performance and segmentation accuracy by enhancing the response of prototype masks. The experimental results demonstrate that the method achieves higher segmentation accuracy while maintaining a relatively high speed.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2022)
Article
Computer Science, Artificial Intelligence
Yiyang Zhou, Qinghai Zheng, Shunshun Bai, Jihua Zhu
Summary: This work focuses on the challenging task of unsupervised multi-view representation learning, aiming to learn a unified feature representation from multiple views. A novel approach called Semantically Consistent Multi-view Representation Learning (SCMRL) is proposed, which excavates underlying multi-view semantic consensus information and utilizes it to guide the unified feature representation learning process. SCMRL consists of within view reconstruction module and unified feature representation learning module, which are elegantly integrated using a contrastive learning strategy to align semantic labels and constrain the learning process.
KNOWLEDGE-BASED SYSTEMS
(2023)