检索结果-内蒙古大学图书馆

Geometry-sensitive semantic modeling in visual and visual-language domains for image captioning

ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE 2025年 147卷

作者： Zhu, Wencai Jiang, Zetao He, Yuting Guilin Univ Elect Technol Guangxi Key Lab Image & Graph Intelligent Proc Guilin 541004 Peoples R China

Transformer-based models with grid features as visual representations perform well in image captioning. However, the division and flattening operations increase the difficulty of capturing objects and their relationships via pure semantic modeling. Furthermore, the natural language generated by the current Transformer model still suffers from semantic overconcentration. In this paper, we aim to improve the attention modules in two ways to solve the above issues. We first propose a Geometry-Sensitive Self-Attention (GSSA) module, subdivide geometric signals in the visual domain into relative position and distance, and assist the semantic modeling process according to their unique characteristics. It compensates for the lack of objects and their relationships in the grid features. Then, we propose a Geometry-Sensitive Cross-Attention (GSCA) module, which perceives the source neighboring relationships between images and text in the visual-language domain from a geometric perspective and uses these relationships to adjust the semantic correspondences between the two dynamically. It spreads overly focused attention to surrounding grids to improve understanding of full image content during captioning. To prove our designs, we apply GSSA and GSCA to a standard Transformer to construct a novel Geometry-Sensitive Transformer Network (GSTNet), which conducts geometry-sensitive semantic modeling in visual and visual-language domains. Extensive experiments are conducted to verify the effectiveness of our proposal. The results show that our GSTNet achieves superior performance compared to many state-of-theart image captioning models on the Microsoft Common Objects in Context (MSCOCO) dataset. Besides, the generalization of GSTNet is also verified on the Flickr30k dataset.

关键词： image captioning Grid features Transformer mechanism Geometry-sensitive semantic modeling

来源：评论

学校读者我要写书评

暂无评论

DMFormer: Dense Memory Linformer for image Captioning

引用

ELECTRONICS 2025年第9期14卷 1716-1716页

作者： He, Yuting Jiang, Zetao Guilin Univ Elect Technol Guangxi Key Lab Image & Graph Intelligent Proc Guilin 541004 Peoples R China

image captioning is a cross-task of computer vision and natural language processing, aiming to describe image content in natural language. Existing methods still have deficiencies in modeling the spatial location and semantic correlation between image regions. Furthermore, these methods often exhibit insufficient interaction between image features and text features. To address these issues, we propose a Linformer-based image captioning method, the Dense Memory Linformer for image Captioning (DMFormer), which has lower time and space complexity than the traditional Transformer architecture. The DMFormer contains two core modules: the Relation Memory Augmented Encoder (RMAE) and the Dense Memory Augmented Decoder (DMAD). In the RMAE, we propose Relation Memory Augmented Attention (RMAA), which combines explicit spatial perception and implicit spatial perception. It explicitly uses geometric information to model the geometric correlation between image regions and implicitly constructs memory unit matrices to learn the contextual information of image region features. In the DMAD, we introduce Dense Memory Augmented Cross Attention (DMACA). This module fully utilizes the low-level and high-level features generated by the RMAE through dense connections, and constructs memory units to store prior knowledge of image and text. It learns the cross-modal associations between visual and linguistic features through an adaptive gating mechanism. Experimental results on the MS-COCO dataset show that the descriptions generated by the DMFormer are richer and more accurate, with significant improvements in various evaluation metrics compared to mainstream methods.

关键词： image captioning Linformer relation memory augmented attention dense memory augmented cross attention

来源：评论

学校读者我要写书评

暂无评论

Linear self-attention with multi-relational graph for knowledge graph completion

引用

APPLIED INTELLIGENCE 2025年第8期55卷 1-15页

作者： Liu, Weida Qiang, Baohua Chen, Ruidong Xie, Yuan Chen, Lirui Chen, Zhiqin Guilin Univ Elect Technol Guangxi Key Lab Image & Graph Intelligent Proc Guilin 541004 Peoples R China

Knowledge graph completion (KGC) aims to infer missing facts based on the existing knowledge. graph Convolutional Networks (GCNs) have gained significant traction due to their proficiency in effectively modeling graph structures, especially within the realm of Knowledge graph Completion (KGC). In GCN-based KGC methodologies, GCNs are initially employed to generate comprehensive representations of entities, followed by the application of Knowledge graph Embedding (KGE) models to elucidate the interactions among entities and relations. However, most GCN-based KGC models ignore the long-range pairwise relationships in the graph. To address these limitations and enhance KGC, we propose a model called Linear Self-Attention with Multi-Relational graph Network (LTRGN). Specifically, this model merges GCN and linear self-attention to serve as the encoder. This model introduces a linear self-attention that can capture long-range node dependencies without introducing excessive computational overhead. Furthermore, we implement an attention mechanism designed to better assess the significance of various neighboring nodes relative to the source node. We demonstrate the effectiveness of the proposed LTRGN on the standard FB15k-237, WN18RR, Kinship, and UMLS datasets. On the dense graphs Kinship and UMLS, the MRR of our model improves by 1.3% and 4.1%, respectively, while Hits@1 increases by 1.7% and 6.4% compared to the best-performing model. The results show the efficacy of the model for the KGC task. The code is released at https://***/lixianqingliuyan/LTRGN.

关键词： Knowledge graph completion Knowledge graph embedding graph convolutional networks Linear self-attention

来源：评论

学校读者我要写书评

暂无评论

YOLO-TC: An Optimized Detection Model for Monitoring Safety-Critical Small Objects in Tower Crane Operations

引用

ALGORITHMS 2025年第1期18卷 27-27页

作者： Ding, Dong Deng, Zhengrong Yang, Rui Guilin Univ Elect Technol Guangxi Key Lab Images & Graph Intelligent Proc Guilin 541004 Peoples R China

Ensuring operational safety within high-risk environments, such as construction sites, is paramount, especially for tower crane operations where distractions can lead to severe accidents. Despite existing behavioral monitoring approaches, the task of identifying small yet hazardous objects like mobile phones and cigarettes in real time remains a significant challenge in ensuring operator compliance and site safety. Traditional object detection models often fall short in crane operator cabins due to complex lighting conditions, cluttered backgrounds, and the small physical scale of target objects. To address these challenges, we introduce YOLO-TC, a refined object detection model tailored specifically for tower crane monitoring applications. Built upon the robust YOLOv7 architecture, our model integrates a novel channel-spatial attention mechanism, ECA-CBAM, into the backbone network, enhancing feature extraction without an increase in parameter count. Additionally, we propose the HA-PANet architecture to achieve progressive feature fusion, addressing scale disparities and prioritizing small object detection while reducing noise from unrelated objects. To improve bounding box regression, the MPDIoU Loss function is employed, resulting in superior accuracy for small, critical objects in dense environments. The experimental results on both the PASCAL VOC benchmark and a custom dataset demonstrate that YOLO-TC outperforms baseline models, showcasing its robustness in identifying high-risk objects under challenging conditions. This model holds significant promise for enhancing automated safety monitoring, potentially reducing occupational hazards by providing a proactive, resilient solution for real-time risk detection in tower crane operations.

关键词： computer vision small object detection YOLOv7 attention mechanism feature fusion

来源：评论

学校读者我要写书评

暂无评论

Privacy-preserving face attribute classification via differential privacy

引用

NEUROCOMPUTING 2025年 626卷

作者： Zhang, Xiaoting Wang, Tao Ji, Junhao Zhang, Yushu Lan, Rushi Nanjing Univ Aeronaut & Astronaut Coll Comp Sci & Technol Nanjing 211106 Peoples R China Guilin Univ Elect Technol Guangxi Key Lab Image & Graph Intelligent Proc Guilin 541004 Peoples R China

The development of face attribute recognition technology has enhanced the intelligence capabilities in the retail industry. Merchants use the surveillance system to capture customers' face images, and analyze their basic characteristics to provide accurate product recommendations and optimize product configurations. However, these captured face images may contain sensitive visual information, especially identity-related data, which could lead to potential security and privacy risks. Current methods for face privacy protection cannot fully support privacy preserving face attributes classification. To this end, this paper proposes a privacy protection scheme that employs differential privacy in the frequency domain to mitigate risks in face attribute classification systems. Our main goal is to take the frequency domain features perturbed with differential privacy as the input of the face attribute classification model to resist privacy attacks. Specifically, the proposed scheme first transforms the original face image into the frequency domain using the discrete cosine transform (DCT) and removes the DC components that contain the visual information. Then the privacy budget allocation in the differential privacy framework is optimized based on the loss of the face attribute classification network. Finally, the corresponding differential privacy noise is added to the frequency representation. The utilization of differential privacy theoretically provides privacy guarantees. Sufficient experimental results show that the proposed scheme can well balance the privacy-utility.

关键词： Face attribute classification Privacy protection Differential privacy Learnable privacy budget

来源：评论

学校读者我要写书评

暂无评论

Word Vector Representation of Latin Cuengh Based on Root Feature Enhancement

引用

APPLIED SCIENCES-BASEL 2025年第1期15卷 211-211页

作者： Lyu, Weibin Chen, Jinlong Qin, Xingguo Li, Jun Guilin Univ Elect Technol Sch Comp Sci & Informat Secur Guilin 541004 Peoples R China Guilin Univ Elect Technol Guangxi Key Lab Image & Graph Intelligent Proc Guilin 541004 Peoples R China

The Latin Cuengh is a kind of language used in China's minority areas. Due to its complex pronunciation and semantic system, it is difficult to spread widely. To deal with and protect this language further, this paper considers using the current word vector representation technology to study it. Word vector representation is the basic method and an important foundation of current research on natural language processing. It relies on a large number of data resources and is obtained through the paradigm of pre-training and feature learning. Due to the extreme lack of Latin Cuengh corpus resources, it is very difficult to obtain word vectors by relying on a large amount of data training. In this study, we propose a word vector representation method that combines the root features of Latin Cuengh words. Specifically, while training and learning the Latin Cuengh language corpus, this method uses the special word roots in the Latin Cuengh language to modify the training process, which can enhance the expression ability of the root features. The method uses the mask method based on BERT to mask the word roots after word segmentation and predict the masked word roots in the output layer of the model to obtain a better vector representation of Latin Cuengh words. The experimental results show that the word vector representation method proposed in this paper is effective and has the ability to express Latin Cuengh semantics. The accuracy rate of words semantic is nearly 2% points higher than that of BERT representation, and the judgment of the semantic similarity of words is more accurate.

关键词： Latin Cuengh word vector representation root feature mask method

来源：评论

学校读者我要写书评

暂无评论

BENet: boundary-enhanced network for real-time semantic segmentation

引用

VISUAL COMPUTER 2025年第1期41卷 229-241页

作者： Lei, Xiaochun Chen, Zeyu Yu, Zhaoxin Jiang, Zetao Guilin Univ Elect Technol Sch Comp Sci & Informat Secur Guilin 541010 Guangxi Peoples R China Guilin Univ Elect Technol Guangxi Key Lab Image & G Intelligent Proc Guilin 541004 Guangxi Peoples R China

In the realm of real-time semantic segmentation, deep neural networks have demonstrated promising potential. However, current methods face challenges when it comes to accurately segmenting object boundaries and small objects. This limitation is partly attributed to the prevalence of convolutional neural networks, which often involve multiple sequential down-sampling operations, resulting in the loss of fine-grained details. To overcome this drawback, we introduce BENet, a real-time semantic segmentation network with a focus on enhancing object boundaries. The proposed BENet integrates two key components: the boundary extraction module (BEM) and the boundary adaption layer (BAL). The proposed BEM efficiently extracts boundary information, while the BAL guides the network using this information to preserve intricate details during the feature extraction process. Furthermore, to address the challenges associated with poor segmentation of elongated objects, we introduce the strip mixed aggregation pyramid pooling module (SMAPPM). This module employs strip pooling kernels to effectively expand the contextual representation and receptive field of the network, thereby enhancing overall segmentation performance. Our experiments conducted on a single RTX 3090 GPU show that our method achieves an mIoU of 79.4% at a speed of 45.5 FPS on the Cityscapes test set without imageNet pre-training.

关键词： Semantic segmentation Deep neural networks Real-time inference Boundary-enhanced

来源：评论

学校读者我要写书评

暂无评论

Robust corner detection in continuous space

引用

VISUAL COMPUTER 2025年第1期41卷 759-772页

作者： Wei, Xiyu Dong, Yanmei Liu, Qin Wang, Lei Lou, Liantang Guangxi Univ Sci & Technol Tus Coll Digit Key Lab Intelligent Informat Proc & Graph Proc Liuzhou 545006 Peoples R China South Cent Minzu Univ Sch Math & Stat Wuhan 430074 Peoples R China

Corner detection is important in image analysis and understanding, but most existing corner detectors are sensitive to image quality, lens radial distortion, and illumination. In this paper, we propose a corner detector for robust corner detection in continuous space. We use the open string theory to construct the continuous representation of an image. Defining a corner as the intersection of two or more curve edges or straight line edges, we design a corner response function for corner determination. In detail, for each integer point, we construct multiple grayscale-parallelograms by any two directed line segments of that point, and the corner response function is based on these grayscale-parallelograms. Finally, a point with a high response value is detected as a corner. Experimental results on conventional images, wide-angle images, and fisheye images show that the proposed method obtains state-of-the-art performance on conventional images and achieves superior performance on wide-angle images and fisheye images, even under weak lighting and low-quality conditions.

关键词： Corner detection Continuous space Line segment Radial distortion Continuous representation

来源：评论

学校读者我要写书评

暂无评论

TourismNER: A Tourism Named Entity Recognition method based on entity boundary joint prediction

引用

intelligent SYSTEMS WITH APPLICATIONS 2025年 25卷

作者： Gao, Kai Zhou, Jiahao Chi, Yunxian Wen, Yimin Hebei Univ Sci & Technol Sch Informat Sci & Engn Shijiazhuang 050018 Peoples R China Guilin Tourism Univ Guangxi Key Lab Culture & Tourism Smart Technol Guilin 541006 Peoples R China Guilin Univ Elect Technol Guangxi Key Lab Image & Graph Intelligent Proc Guilin 541004 Peoples R China

Tourism named entity recognition is indispensable in tourism information extraction, and plays a crucial role in constructing tourism knowledge map and enhancing tourism knowledge quiz system. The difficulty of tourism named entity recognition lies in its complex nested structure, and the lengthy entity naming length. To address these existing problems, we propose a tourism named entity recognition model that jointly predicts entity boundaries, adopting a training strategy of data preprocessing to enhance the model's ability for tourism named entity boundary recognition, while our model introduces a pre-trained Bert model as well as BiLSTM coding to enhance the representation of the model's contexts, and uses a combined predictor of Biaffine and MLP to enhance the model's recognition performance for boundaries, as well as introducing label smoothing cross entropy to smooth the target labels during the training process. Experiments are conducted on three datasets with different granularities. From the analysis of the experimental results, it can be seen that the named entity recognition method achieves higher accuracy and F1 value compared with the optimal baseline model, and also proves the effectiveness and generality of the modeling method proposed in this paper.

关键词： Natural Language processing Tourism Named Entity Recognition Entity boundary recognition Joint prediction

来源：评论

学校读者我要写书评

暂无评论

Deepfake Video Detection Using Facial Feature Points and Ch-Transformer

引用

ACM TRANSACTIONS ON MULTIMEDIA COMPUTING COMMUNICATIONS AND APPLICATIONS 2025年第2期21卷 1-22页

作者： Yang, Rui Lan, Rushi Deng, Zhenrong Luo, Xiaonan Sun, Xiyan Guilin Univ Elect Technol Guangxi Key Lab Image & G Intelligent Proc Guilin Peoples R China Guilin Univ Elect Technol Nanning Res Inst Nanning Peoples R China

With the development of Metaverse technology, the avatar in Metaverse has faced serious security and privacy concerns. Analyzing facial features to distinguish between genuine and manipulated facial videos holds significant research importance for ensuring the authenticity of characters in the virtual world and for mitigating discrimination as well as preventing malicious use of facial data. To address this issue, the Facial Feature Points and Class-head-Transformer (FFP-ChT) deepfake video detection model is designed based on the clues of different FFPs distribution in real and fake videos and different displacement distances of real and fake FFPs between frames. The face video input is first detected by the BlazeFace model, and the face detection results are fed into the FaceMesh model to extract 468 FFPs. Then, the Lucas-Kanade (LK) optical flow method is used to track the points of the face, the face calibration algorithm is introduced to re-calibrate the FFPs, and the jitter displacement is calculated by tracking the FFPs between frames. Finally, the Ch is designed in the transformer, and the FFPs and FFP displacement are jointly classified through the ChT model. In this way, the designed ChT classifier is able to accurately and effectively identify deepfake videos. Experiments on open datasets clearly demonstrate the effectiveness and generalization capabilities of our approach. CCS Concepts: center dot Security and privacy - Privacy protections;Social network security and privacy;center dot Theory of computation - Structured prediction

关键词： Transformer deepfake detection facial feature FFP-ChT facial feature points

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：