Expert Systems with Applications最新文献_第5页

Research on the generation and evaluation of bridge defect datasets for underwater environments utilizing CycleGAN networks 利用 CycleGAN 网络生成和评估水下环境桥梁缺陷数据集的研究

IF 7.5 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Expert Systems with Applications

Pub Date : 2024-11-06 DOI: 10.1016/j.eswa.2024.125576

Fei Zhang , Yeyang Gu , Ling Yin , Jialei Song , Chaochao Qiu , Zhengwei Ye , Xiangyin Chen , Jing Wu

The surface cracks on the underwater structures critically damages the overall reliability of the structures and reduces their strength. It is significant to monitor these cracks in timely manner. Recently, deep learning algorithms have been used for large scale data study and predictions. However, deep supervised learning algorithms need to get training on large scale data set which is time consuming and difficult to apply on the underwater structures. Therefore, it is highly needed to address these issues. Current research proposes an improved cycle-constraint generative adversarial algorithm for the timely detection of surface cracks in underwater structures. It utilizes an enhanced cycle-consistent generative adversarial network (CycleGAN). The proposed algorithm uses image processing techniques including DeblurGAN and Dark channel prior methods to get quality of dataset from underwater structures. The proposed Algorithm introduces a novel cross-domain VGG-cosine similarity assessment to precisely evaluate the performance of proposed algorithm to retain crack information etc. Moreover, performance of proposed algorithm is evaluated through both qualitative and quantitative methods. The quantitative results are directly obtained from the visual results are presented which are generated by the proposed Algorithm. Whereas, the performance of proposed algorithm based on quantitative results is obtained from metrics including PSNR, SSIM, and FID. Experimental results indicates that the proposed algorithm outperforms the original CycleGAN. End results indicate that the proposed algorithm decreased the value of FID by 20 % and increased the values of PSNR and SSIM by 2.37 % and 3.33 % respectively. Quantitative and qualitative results of the proposed algorithm give significant advantages during creating of surface crack images.

水下结构的表面裂缝会严重损害结构的整体可靠性并降低其强度。及时监测这些裂缝意义重大。最近，深度学习算法被用于大规模数据研究和预测。然而，深度监督学习算法需要在大规模数据集上进行训练，既耗时又难以应用于水下结构。因此，亟需解决这些问题。目前的研究提出了一种改进的周期约束生成对抗算法，用于及时检测水下结构的表面裂缝。该算法采用了增强型周期约束生成对抗网络（CycleGAN）。该算法采用了包括 DeblurGAN 和暗通道先验方法在内的图像处理技术，以获得高质量的水下结构数据集。拟议算法引入了一种新颖的跨域 VGG-余弦相似性评估，以精确评估拟议算法在保留裂缝信息等方面的性能。此外，还通过定性和定量方法评估了所提算法的性能。定量结果直接从所提出的算法生成的视觉结果中获得。而基于定量结果的拟议算法性能则是通过 PSNR、SSIM 和 FID 等指标获得的。实验结果表明，所提出的算法优于原始的 CycleGAN 算法。最终结果表明，所提算法的 FID 值降低了 20%，PSNR 和 SSIM 值分别提高了 2.37% 和 3.33%。所提算法的定量和定性结果在创建表面裂纹图像时具有显著优势。

{"title":"Research on the generation and evaluation of bridge defect datasets for underwater environments utilizing CycleGAN networks","authors":"Fei Zhang , Yeyang Gu , Ling Yin , Jialei Song , Chaochao Qiu , Zhengwei Ye , Xiangyin Chen , Jing Wu","doi":"10.1016/j.eswa.2024.125576","DOIUrl":"10.1016/j.eswa.2024.125576","url":null,"abstract":"<div><div>The surface cracks on the underwater structures critically damages the overall reliability of the structures and reduces their strength. It is significant to monitor these cracks in timely manner. Recently, deep learning algorithms have been used for large scale data study and predictions. However, deep supervised learning algorithms need to get training on large scale data set which is time consuming and difficult to apply on the underwater structures. Therefore, it is highly needed to address these issues. Current research proposes an improved cycle-constraint generative adversarial algorithm for the timely detection of surface cracks in underwater structures. It utilizes an enhanced cycle-consistent generative adversarial network (CycleGAN). The proposed algorithm uses image processing techniques including DeblurGAN and Dark channel prior methods to get quality of dataset from underwater structures. The proposed Algorithm introduces a novel cross-domain VGG-cosine similarity assessment to precisely evaluate the performance of proposed algorithm to retain crack information etc. Moreover, performance of proposed algorithm is evaluated through both qualitative and quantitative methods. The quantitative results are directly obtained from the visual results are presented which are generated by the proposed Algorithm. Whereas, the performance of proposed algorithm based on quantitative results is obtained from metrics including PSNR, SSIM, and FID. Experimental results indicates that the proposed algorithm outperforms the original CycleGAN. End results indicate that the proposed algorithm decreased the value of FID by 20 % and increased the values of PSNR and SSIM by 2.37 % and 3.33 % respectively. Quantitative and qualitative results of the proposed algorithm give significant advantages during creating of surface crack images.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"262 ","pages":"Article 125576"},"PeriodicalIF":7.5,"publicationDate":"2024-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142593683","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A framework to Prediction occupant injuries in rotated seating arrangements 旋转座椅安排中的乘员伤害预测框架

IF 7.5 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Expert Systems with Applications

Pub Date : 2024-11-06 DOI: 10.1016/j.eswa.2024.125698

Alexander Diederich, Christophe Bastien, Michael Blundell

This paper describes an innovative computational framework that can address the passive safety of occupants and the prediction of injuries for a wide range of rotated seat arrangements in future autonomous vehicles. An Active Human Model (AHM) wearing a 3-point seat belt with kinematics previously validated using test data was positioned in the rotated seats and simulations were performed for a pre-crash braking phase followed by a frontal collision. The pre-crash braking pulse was parameterised for intensity and shape, both of which affect the collision speed. Computer routines were created to adjust the restraint system response to the impact speed, as well as the crash pulse pattern, based on a Honda Accord MY2011 Finite Element (FE) model. A Design of Experiments (DoE) study was created to explore 400 possible scenarios, and 12 standard injury responses were extracted and converted into AIS2 + and AIS3 + risk values. A Reduced Order Model (ROM), based on the Proper Orthogonal Decomposition (POD) technique, was trained using the 400 scenarios and used to find the seating positions that resulted in AIS2 + and AIS3 + values higher than the base values obtained with the Honda Accord MY2011 FE model. The paper concludes that this new framework is capable of carry out fast computations of dangerous seating positions and the occupant’s kinematics in seconds. The novel framework provides vehicle designers and vehicle safety teams with the capability to identify potentially dangerous positions for the rotated seating arrangements that are envisaged to feature in the cabins of future autonomous vehicles.

本文介绍了一个创新的计算框架，该框架可以解决乘员的被动安全问题，并预测未来自动驾驶汽车中各种旋转座椅布置的伤害情况。将佩戴三点式安全带的主动人体模型（AHM）放置在旋转座椅上，并对碰撞前制动阶段和随后的正面碰撞进行了模拟。对碰撞前制动脉冲的强度和形状进行了参数化，这两个因素都会影响碰撞速度。根据本田雅阁 2011 款有限元 (FE) 模型创建了计算机例程，以调整约束系统对碰撞速度和碰撞脉冲模式的响应。实验设计（DoE）研究探索了 400 种可能的情况，提取了 12 种标准伤害反应，并将其转换为 AIS2 + 和 AIS3 + 风险值。基于适当正交分解（POD）技术的降序模型（ROM）利用这 400 种情景进行了训练，并用于找出导致 AIS2 + 和 AIS3 + 值高于本田雅阁 MY2011 FE 模型基础值的座椅位置。论文的结论是，这一新框架能够在数秒内快速计算危险的座椅位置和乘员的运动学特性。新框架为车辆设计师和车辆安全团队提供了识别旋转座椅布置潜在危险位置的能力，这种旋转座椅布置预计将出现在未来的自动驾驶车辆的驾驶室中。

{"title":"A framework to Prediction occupant injuries in rotated seating arrangements","authors":"Alexander Diederich, Christophe Bastien, Michael Blundell","doi":"10.1016/j.eswa.2024.125698","DOIUrl":"10.1016/j.eswa.2024.125698","url":null,"abstract":"<div><div>This paper describes an innovative computational framework that can address the passive safety of occupants and the prediction of injuries for a wide range of rotated seat arrangements in future autonomous vehicles. An Active Human Model (AHM) wearing a 3-point seat belt with kinematics previously validated using test data was positioned in the rotated seats and simulations were performed for a pre-crash braking phase followed by a frontal collision. The pre-crash braking pulse was parameterised for intensity and shape, both of which affect the collision speed. Computer routines were created to adjust the restraint system response to the impact speed, as well as the crash pulse pattern, based on a Honda Accord MY2011 Finite Element (FE) model. A Design of Experiments (DoE) study was created to explore 400 possible scenarios, and 12 standard injury responses were extracted and converted into AIS2 + and AIS3 + risk values. A Reduced Order Model (ROM), based on the Proper Orthogonal Decomposition (POD) technique, was trained using the 400 scenarios and used to find the seating positions that resulted in AIS2 + and AIS3 + values higher than the base values obtained with the Honda Accord MY2011 FE model. The paper concludes that this new framework is capable of carry out fast computations of dangerous seating positions and the occupant’s kinematics in seconds. The novel framework provides vehicle designers and vehicle safety teams with the capability to identify potentially dangerous positions for the rotated seating arrangements that are envisaged to feature in the cabins of future autonomous vehicles.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"263 ","pages":"Article 125698"},"PeriodicalIF":7.5,"publicationDate":"2024-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662116","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

An intelligent font generation system based on stroke inference, mitigating production labor and enhancing design experience 基于笔画推理的智能字体生成系统，减轻生产劳动强度，提升设计体验

IF 7.5 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Expert Systems with Applications

Pub Date : 2024-11-06 DOI: 10.1016/j.eswa.2024.125657

Bolin Wang , Kejun Zhang , Zewen Chen , Lyukesheng Shen , Xinyi Shen , Yu Liu , Jiang Bian , Hanshu Shen

Traditionally, font design has relied on manual craftsmanship by designers, a time-consuming and labor-intensive process that can take over a year to complete a new font family. Despite advancements in computer vision and graphics enabling the automation of font generation, creating high-quality fonts meeting commercial standards remains a technical challenge. Current automatic font generation technologies have not fully met production demands, mainly due to their lack of focus on generating glyphs that can be decomposed into strokes and their ineffective post-processing interaction, particularly for Chinese fonts. This study presents an innovative system for intelligently generating Chinese character fonts. The system utilizes a stroke database created by professional designers and combines font images generated through style transfer learning to perform stroke inference for font generation. The system’s core lies in its unique stroke inference mechanism, accurately identifying and matching strokes within font images to efficiently align with standard stroke data in the database. This approach not only improves the precision of font generation but also streamlines subsequent processing steps. Compared to traditional font design systems, our system shows significant advantages in generating fonts suitable for commercial use. It not only aids designers in enhancing work efficiency but also has the potential to greatly increase the production efficiency of font libraries. Moreover, the system’s design is scalable, offering extensive application prospects for future expansion to other East Asian scripts like Japanese and Korean.

传统上，字体设计依赖于设计师的手工制作，这是一个耗时耗力的过程，完成一个新的字体系列可能需要一年多的时间。尽管计算机视觉和图形学技术的进步使字体生成自动化成为可能，但创建符合商业标准的高质量字体仍然是一项技术挑战。目前的字体自动生成技术还不能完全满足生产需求，主要原因是这些技术不注重生成可分解为笔画的字形，而且后处理交互效果不佳，尤其是对中文字体而言。本研究提出了一种智能生成汉字字体的创新系统。该系统利用专业设计师创建的笔画数据库，结合通过风格转移学习生成的字体图像，进行笔画推理，从而生成字体。该系统的核心在于其独特的笔画推理机制，能准确识别和匹配字体图像中的笔画，从而有效地与数据库中的标准笔画数据保持一致。这种方法不仅提高了字体生成的精度，还简化了后续处理步骤。与传统的字体设计系统相比，我们的系统在生成适合商业用途的字体方面具有显著优势。它不仅能帮助设计人员提高工作效率，还有可能大大提高字体库的生产效率。此外，该系统的设计还具有可扩展性，为将来扩展到日文和韩文等其他东亚文字提供了广泛的应用前景。

{"title":"An intelligent font generation system based on stroke inference, mitigating production labor and enhancing design experience","authors":"Bolin Wang , Kejun Zhang , Zewen Chen , Lyukesheng Shen , Xinyi Shen , Yu Liu , Jiang Bian , Hanshu Shen","doi":"10.1016/j.eswa.2024.125657","DOIUrl":"10.1016/j.eswa.2024.125657","url":null,"abstract":"<div><div>Traditionally, font design has relied on manual craftsmanship by designers, a time-consuming and labor-intensive process that can take over a year to complete a new font family. Despite advancements in computer vision and graphics enabling the automation of font generation, creating high-quality fonts meeting commercial standards remains a technical challenge. Current automatic font generation technologies have not fully met production demands, mainly due to their lack of focus on generating glyphs that can be decomposed into strokes and their ineffective post-processing interaction, particularly for Chinese fonts. This study presents an innovative system for intelligently generating Chinese character fonts. The system utilizes a stroke database created by professional designers and combines font images generated through style transfer learning to perform stroke inference for font generation. The system’s core lies in its unique stroke inference mechanism, accurately identifying and matching strokes within font images to efficiently align with standard stroke data in the database. This approach not only improves the precision of font generation but also streamlines subsequent processing steps. Compared to traditional font design systems, our system shows significant advantages in generating fonts suitable for commercial use. It not only aids designers in enhancing work efficiency but also has the potential to greatly increase the production efficiency of font libraries. Moreover, the system’s design is scalable, offering extensive application prospects for future expansion to other East Asian scripts like Japanese and Korean.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"263 ","pages":"Article 125657"},"PeriodicalIF":7.5,"publicationDate":"2024-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662272","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A new parallel cooperative landscape smoothing algorithm and its applications on TSP and UBQP 一种新的并行协同景观平滑算法及其在 TSP 和 UBQP 上的应用

IF 7.5 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Expert Systems with Applications

Pub Date : 2024-11-06 DOI: 10.1016/j.eswa.2024.125611

Wei Wang , Jialong Shi , Jianyong Sun , Arnaud Liefooghe , Qingfu Zhang

Combinatorial optimization problem (COP) is difficult to solve because of the massive number of local optimal solutions in his solution space. Various methods have been put forward to smooth the solution space of COPs, including homotopic convex (HC) transformation for the traveling salesman problem (TSP). This paper extends the HC transformation approach to unconstrained binary quadratic programming (UBQP) by proposing a method to construct a unimodal toy UBQP of any size. We theoretically prove the unimodality of the constructed toy UBQP. After that, we apply this unimodal toy UBQP to smooth the original UBQP by using the HC transformation framework and empirically verify the smoothing effects. Subsequently, we introduce an iterative algorithmic framework incorporating HC transformation, referred as landscape smoothing iterated local search (LSILS). Our experimental analyses, conducted on various UBQP instances show the effectiveness of LSILS. Furthermore, this paper proposes a parallel cooperative variant of LSILS, denoted as PC-LSILS and apply it to both the UBQP and the TSP. Our experimental findings highlight that PC-LSILS improves the smoothing performance of the HC transformation, and further improves the overall performance of the algorithm.

组合优化问题（COP）因其解空间中存在大量局部最优解而难以解决。人们提出了各种方法来平滑 COP 的解空间，包括旅行推销员问题（TSP）的同位凸（HC）变换。本文将 HC 变换方法扩展到无约束二元二次编程（UBQP），提出了一种构建任意大小的单模态玩具 UBQP 的方法。我们从理论上证明了所构建的玩具 UBQP 的单模态性。之后，我们利用 HC 变换框架，将这种单模态玩具 UBQP 应用于平滑原始 UBQP，并通过经验验证了平滑效果。随后，我们介绍了一种包含 HC 变换的迭代算法框架，即景观平滑迭代局部搜索（LSILS）。我们在各种 UBQP 实例上进行的实验分析表明了 LSILS 的有效性。此外，本文还提出了 LSILS 的并行合作变体，称为 PC-LSILS，并将其应用于 UBQP 和 TSP。实验结果表明，PC-LSILS 提高了 HC 变换的平滑性能，并进一步提高了算法的整体性能。

{"title":"A new parallel cooperative landscape smoothing algorithm and its applications on TSP and UBQP","authors":"Wei Wang , Jialong Shi , Jianyong Sun , Arnaud Liefooghe , Qingfu Zhang","doi":"10.1016/j.eswa.2024.125611","DOIUrl":"10.1016/j.eswa.2024.125611","url":null,"abstract":"<div><div>Combinatorial optimization problem (COP) is difficult to solve because of the massive number of local optimal solutions in his solution space. Various methods have been put forward to smooth the solution space of COPs, including homotopic convex (HC) transformation for the traveling salesman problem (TSP). This paper extends the HC transformation approach to unconstrained binary quadratic programming (UBQP) by proposing a method to construct a unimodal toy UBQP of any size. We theoretically prove the unimodality of the constructed toy UBQP. After that, we apply this unimodal toy UBQP to smooth the original UBQP by using the HC transformation framework and empirically verify the smoothing effects. Subsequently, we introduce an iterative algorithmic framework incorporating HC transformation, referred as landscape smoothing iterated local search (LSILS). Our experimental analyses, conducted on various UBQP instances show the effectiveness of LSILS. Furthermore, this paper proposes a parallel cooperative variant of LSILS, denoted as PC-LSILS and apply it to both the UBQP and the TSP. Our experimental findings highlight that PC-LSILS improves the smoothing performance of the HC transformation, and further improves the overall performance of the algorithm.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"263 ","pages":"Article 125611"},"PeriodicalIF":7.5,"publicationDate":"2024-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662115","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

MuSe-CarASTE: A comprehensive dataset for aspect sentiment triplet extraction in automotive review videos MuSe-CarASTE：用于汽车评论视频中方面情感三元组提取的综合数据集

IF 7.5 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Expert Systems with Applications

Pub Date : 2024-11-06 DOI: 10.1016/j.eswa.2024.125695

Atiya Usmani , Saeed Hamood Alsamhi , Muhammad Jaleed Khan , John Breslin , Edward Curry

In the Aspect-Based Sentiment Analysis (ABSA) domain, the Aspect Sentiment Triplet Extraction (ASTE) task has emerged as a pivotal endeavor, offering insights into nuanced aspects, opinions, and sentiment relationships. This paper introduces “MuSe-CarASTE”, an extensive and meticulously curated dataset purpose-built to propel ASTE advancements within the automotive domain. The core emphasis of MuSe-CarASTE is on aspect, opinion, and sentiment triplets, facilitating a comprehensive analysis of product reviews. Comprising transcripts from MuSe-Car’s automotive video reviews, MuSe-CarASTE presents a sub-stantial collection of nearly 28,295 sentences organized into 5,500 segments. Each segment is meticulously annotated with multiple aspects, opinions, and sentiment labels, offering unprecedented granularity for ASTE tasks. The percentage agreement between annotated triples by different annotators over the randomly sampled subset of the dataset is 79.74 %, at similarity threshold τ = 0.60. We also experimented with four baseline models on our datset and report results. The distinctiveness of the dataset emerges from its extension into the automotive domain, shedding light on sentiment dynamics specific to vehicles. With the fusion of extensive content and real-world applicability, MuSe-CarASTE presents a fertile ground for Natural Language Processing (NLP) innovation. Researchers, practitioners, and data scientists can harness MuSe-CarASTE to build and evaluate NLP models tailored for challenges in ASTE. These challenges encompass intricate aspect-opinion relationships, multi-word aspect and opinion extraction, and the subtleties of vague language. Moreover, including aspects not verbatim in sentences introduces a practical dimension to our dataset, enabling real-world applications like review pattern analysis, summarization, and recommender system enhancement. As a pioneering benchmark for NLP model evaluation in ABSA, MuSe-CarASTE integrates content richness, real-world context, and sentiment complexity. The integration empowers the development of accurate, adaptable, and insightful sentiment analysis models within the automotive review landscape.

在基于方面的情感分析（ABSA）领域，方面情感三重抽取（ASTE）任务已成为一项至关重要的工作，它提供了对细微方面、观点和情感关系的洞察力。本文介绍了 "MuSe-CarASTE"，这是一个广泛而精心策划的数据集，旨在推动汽车领域 ASTE 的发展。MuSe-CarASTE 的核心重点是方面、观点和情感三元组，有助于对产品评论进行全面分析。MuSe-CarASTE 由 MuSe-Car 的汽车视频评论记录组成，提供了一个由近 28,295 个句子组成的子集，分为 5,500 个片段。每个片段都精心标注了多个方面、观点和情感标签，为 ASTE 任务提供了前所未有的粒度。在相似性阈值 τ = 0.60 的条件下，不同注释者在随机抽样的数据集子集上注释的三元组之间的一致率为 79.74%。我们还在数据集上试验了四种基线模型，并报告了结果。该数据集的独特之处在于它扩展到了汽车领域，揭示了汽车特有的情感动态。MuSe-CarASTE 融合了广泛的内容和现实世界的适用性，为自然语言处理 (NLP) 的创新提供了肥沃的土壤。研究人员、从业人员和数据科学家可以利用 MuSe-CarASTE 来构建和评估针对 ASTE 中的挑战而定制的 NLP 模型。这些挑战包括错综复杂的方面-观点关系、多词方面和观点提取以及模糊语言的微妙之处。此外，将句子中的非逐字方面纳入数据集还为我们的数据集引入了一个实用维度，使评论模式分析、总结和推荐系统增强等现实世界的应用成为可能。作为 ABSA 中 NLP 模型评估的先驱基准，MuSe-CarASTE 整合了内容丰富性、真实语境和情感复杂性。这种整合有助于在汽车评论领域开发准确、适应性强且具有洞察力的情感分析模型。

{"title":"MuSe-CarASTE: A comprehensive dataset for aspect sentiment triplet extraction in automotive review videos","authors":"Atiya Usmani , Saeed Hamood Alsamhi , Muhammad Jaleed Khan , John Breslin , Edward Curry","doi":"10.1016/j.eswa.2024.125695","DOIUrl":"10.1016/j.eswa.2024.125695","url":null,"abstract":"<div><div>In the Aspect-Based Sentiment Analysis (ABSA) domain, the Aspect Sentiment Triplet Extraction (ASTE) task has emerged as a pivotal endeavor, offering insights into nuanced aspects, opinions, and sentiment relationships. This paper introduces “MuSe-CarASTE”, an extensive and meticulously curated dataset purpose-built to propel ASTE advancements within the automotive domain. The core emphasis of MuSe-CarASTE is on aspect, opinion, and sentiment triplets, facilitating a comprehensive analysis of product reviews. Comprising transcripts from MuSe-Car’s automotive video reviews, MuSe-CarASTE presents a sub-stantial collection of nearly 28,295 sentences organized into 5,500 segments. Each segment is meticulously annotated with multiple aspects, opinions, and sentiment labels, offering unprecedented granularity for ASTE tasks. The percentage agreement between annotated triples by different annotators over the randomly sampled subset of the dataset is 79.74 %, at similarity threshold <em>τ</em> = 0.60. We also experimented with four baseline models on our datset and report results. The distinctiveness of the dataset emerges from its extension into the automotive domain, shedding light on sentiment dynamics specific to vehicles. With the fusion of extensive content and real-world applicability, MuSe-CarASTE presents a fertile ground for Natural Language Processing (NLP) innovation. Researchers, practitioners, and data scientists can harness MuSe-CarASTE to build and evaluate NLP models tailored for challenges in ASTE. These challenges encompass intricate aspect-opinion relationships, multi-word aspect and opinion extraction, and the subtleties of vague language. Moreover, including aspects not verbatim in sentences introduces a practical dimension to our dataset, enabling real-world applications like review pattern analysis, summarization, and recommender system enhancement. As a pioneering benchmark for NLP model evaluation in ABSA, MuSe-CarASTE integrates content richness, real-world context, and sentiment complexity. The integration empowers the development of accurate, adaptable, and insightful sentiment analysis models within the automotive review landscape.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"262 ","pages":"Article 125695"},"PeriodicalIF":7.5,"publicationDate":"2024-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662776","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

CTIFTrack: Continuous Temporal Information Fusion for object track CTIFTrack：用于物体跟踪的连续时态信息融合技术

IF 7.5 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Expert Systems with Applications

Pub Date : 2024-11-06 DOI: 10.1016/j.eswa.2024.125654

Zhiguo Zhang , Zhiqing Guo , Liejun Wang, Yongming Li

In visual tracking tasks, researchers usually focus on increasing the complexity of the model or only discretely focusing on the changes in the object itself to achieve accurate recognition and tracking of the moving object. However, they often overlook the significant contribution of video-level linear temporal information fusion and continuous spatiotemporal mapping to tracking tasks. This oversight may lead to poor tracking performance or insufficient real-time ability of the model in complex scenes. Therefore, this paper proposes a real-time tracker, namely Continuous Temporal Information Fusion Tracker (CTIFTrack). The key of CTIFTrack lies in its well-designed Temporal Information Fusion (TIF) module, which cleverly performs a linear fusion of the temporal information between the

(t - 1) -th

and the

t -th

frames and completes the spatiotemporal mapping. This enables the tracker to better understand the overall spatiotemporal information and contextual spatiotemporal correlations within the video, thereby having a positive impact on the tracking task. In addition, this paper also proposes the Object Template Feature Refinement (OTFR) module, which effectively captures the global information and local details of the object, and further improves the tracker’s understanding of the object features. Extensive experiments are conducted on seven benchmarks, such as LaSOT, GOT-10K, UAV123, NFS, TrackingNet, VOT2018 and OTB-100. The experimental results validate the significant contribution of the TIF module and OTFR module to the tracking task, as well as the effectiveness of CTIFTrack. It is worth noting that while maintaining excellent tracking performance, CTIFTrack also shows outstanding real-time tracking speed. On the Nvidia Tesla T4-16GB GPU, the

F P S

of CTIFTrack reaches 71.98. The code and demo materials will be available at https://github.com/vpsg-research/CTIFTrack.

在视觉跟踪任务中，研究人员通常专注于提高模型的复杂度，或仅离散地关注物体本身的变化，以实现对运动物体的精确识别和跟踪。然而，他们往往忽视了视频级线性时空信息融合和连续时空映射对跟踪任务的重要贡献。这种疏忽可能会导致复杂场景下的跟踪性能不佳或模型的实时性不足。因此，本文提出了一种实时跟踪器，即连续时空信息融合跟踪器（CTIFTrack）。CTIFTrack 的关键在于其精心设计的时空信息融合（Temporal Information Fusion，TIF）模块，该模块巧妙地将第 (t-1)-th 帧与第 t 帧之间的时空信息进行线性融合，并完成时空映射。这样，跟踪器就能更好地理解视频中的整体时空信息和上下文时空相关性，从而对跟踪任务产生积极影响。此外，本文还提出了物体模板特征提纯（OTFR）模块，该模块能有效捕捉物体的全局信息和局部细节，进一步提高跟踪器对物体特征的理解。在 LaSOT、GOT-10K、UAV123、NFS、TrackingNet、VOT2018 和 OTB-100 等七个基准上进行了广泛的实验。实验结果验证了 TIF 模块和 OTFR 模块对跟踪任务的重要贡献，以及 CTIFTrack 的有效性。值得注意的是，在保持出色跟踪性能的同时，CTIFTrack 还表现出了出色的实时跟踪速度。在 Nvidia Tesla T4-16GB GPU 上，CTIFTrack 的 FPS 达到 71.98。代码和演示材料将在 https://github.com/vpsg-research/CTIFTrack 网站上提供。

{"title":"CTIFTrack: Continuous Temporal Information Fusion for object track","authors":"Zhiguo Zhang , Zhiqing Guo , Liejun Wang, Yongming Li","doi":"10.1016/j.eswa.2024.125654","DOIUrl":"10.1016/j.eswa.2024.125654","url":null,"abstract":"<div><div>In visual tracking tasks, researchers usually focus on increasing the complexity of the model or only discretely focusing on the changes in the object itself to achieve accurate recognition and tracking of the moving object. However, they often overlook the significant contribution of video-level linear temporal information fusion and continuous spatiotemporal mapping to tracking tasks. This oversight may lead to poor tracking performance or insufficient real-time ability of the model in complex scenes. Therefore, this paper proposes a real-time tracker, namely Continuous Temporal Information Fusion Tracker (CTIFTrack). The key of CTIFTrack lies in its well-designed Temporal Information Fusion (TIF) module, which cleverly performs a linear fusion of the temporal information between the <span><math><mrow><mrow><mo>(</mo><mi>t</mi><mtext>-</mtext><mn>1</mn><mo>)</mo></mrow><mtext>-th</mtext></mrow></math></span> and the <span><math><mrow><mi>t</mi><mtext>-th</mtext></mrow></math></span> frames and completes the spatiotemporal mapping. This enables the tracker to better understand the overall spatiotemporal information and contextual spatiotemporal correlations within the video, thereby having a positive impact on the tracking task. In addition, this paper also proposes the Object Template Feature Refinement (OTFR) module, which effectively captures the global information and local details of the object, and further improves the tracker’s understanding of the object features. Extensive experiments are conducted on seven benchmarks, such as LaSOT, GOT-10K, UAV123, NFS, TrackingNet, VOT2018 and OTB-100. The experimental results validate the significant contribution of the TIF module and OTFR module to the tracking task, as well as the effectiveness of CTIFTrack. It is worth noting that while maintaining excellent tracking performance, CTIFTrack also shows outstanding real-time tracking speed. On the Nvidia Tesla T4-16GB GPU, the <span><math><mrow><mi>F</mi><mi>P</mi><mi>S</mi></mrow></math></span> of CTIFTrack reaches 71.98. The code and demo materials will be available at <span><span>https://github.com/vpsg-research/CTIFTrack</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"262 ","pages":"Article 125654"},"PeriodicalIF":7.5,"publicationDate":"2024-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662783","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Automatic detection of cyberbullying behaviour on social media using Stacked Bi-Gru attention with BERT model 使用堆叠双顾注意与 BERT 模型自动检测社交媒体上的网络欺凌行为

IF 7.5 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Expert Systems with Applications

Pub Date : 2024-11-05 DOI: 10.1016/j.eswa.2024.125641

Mohan K. Mali, Ranjeet R. Pawar, Sandeep A. Shinde, Satish D. Kale, Sameer V. Mulik, Asmita A. Jagtap, Pratibha A. Tambewagh, Punam U. Rajput

Cyberbullying behaviour has drawn more attention as social media usage has grown. Teen suicide has been related to cyberbullying, among other serious and harmful effects on a person’s life. Using the appropriate natural language processing and machine learning techniques, it is possible to proactively identify bullying content to reduce and eventually eradicate cyberbullying. Accordingly, the article proposed an automated deep-learning model for detecting aggressive activity in cyberbullying. Initially, the data was extracted from the social media platform using Formspring, Instagram and MySpace datasets for perceiving cyberbullying behaviour, then the collected data are input for preprocessing. To remove the raw data, several preprocessing processes have been introduced. They consist of removing stop words, white spaces for punctuation, and changing the comments to lowercase. Lexical Density (LD) has been one of the metrics used to gauge language complexity generally. As a result, the study made use of the Feature Density (FD) to calculate how complicated certain natural language datasets are using the linguistically backed preprocessing model. After preprocessing, the data are input to the feature selection process which selects the pertinent features or attributes to include in predictive modelling and which to leave out. Since, the article proposed a Binary Chimp Optimization (BCO)-based Feature Selection (BCO-FSS) technique, which selects the subset of features for classification performance improvement. The selected features are exploited for cyberbullying behaviour detection. To identify the exploit of social media for cyberbullying text content, the article suggested Stacked Bidirectional Gated Recurrent Unit (SBiGRU) Attention for learning spatial location information and sequential semantic representations using a Bi-GRU. Additionally, the BERT model is employed as a base classifier to recognize and categorise aggressive behaviour in the textual content. The Matlab software is employed for simulation. For accuracy, precision, recall, and F1-Score, this experiment yielded a practically perfect outcome with values of 99.12%, 94.73%, 97.45%, and 93.91% respectively.

随着社交媒体使用的增加，网络欺凌行为引起了更多的关注。青少年自杀与网络欺凌有关，网络欺凌还对人的一生造成了其他严重有害影响。利用适当的自然语言处理和机器学习技术，可以主动识别欺凌内容，从而减少并最终消除网络欺凌。因此，文章提出了一种自动深度学习模型，用于检测网络欺凌中的攻击性活动。首先，使用 Formspring、Instagram 和 MySpace 数据集从社交媒体平台中提取数据，用于感知网络欺凌行为，然后将收集到的数据输入预处理。为了去除原始数据，引入了几个预处理过程。这些预处理包括删除停顿词、标点符号留白和将注释改为小写。词汇密度（LD）一直是用来衡量语言复杂性的指标之一。因此，本研究使用了特征密度（FD），利用语言支持的预处理模型来计算某些自然语言数据集的复杂程度。经过预处理后，数据被输入到特征选择流程，该流程会选择相关的特征或属性，并将其纳入预测建模中，而将其排除在外。因此，文章提出了基于二进制黑猩猩优化（BCO）的特征选择（BCO-FSS）技术，该技术可选择特征子集以提高分类性能。所选特征可用于网络欺凌行为检测。为了识别社交媒体中的网络欺凌文本内容，文章建议使用堆叠双向门控递归单元（SBiGRU）注意学习空间位置信息和顺序语义表征。此外，还采用 BERT 模型作为基础分类器，对文本内容中的攻击行为进行识别和分类。模拟使用了 Matlab 软件。在准确率、精确度、召回率和 F1-Score 方面，本实验取得了几乎完美的结果，数值分别为 99.12%、94.73%、97.45% 和 93.91%。

{"title":"Automatic detection of cyberbullying behaviour on social media using Stacked Bi-Gru attention with BERT model","authors":"Mohan K. Mali, Ranjeet R. Pawar, Sandeep A. Shinde, Satish D. Kale, Sameer V. Mulik, Asmita A. Jagtap, Pratibha A. Tambewagh, Punam U. Rajput","doi":"10.1016/j.eswa.2024.125641","DOIUrl":"10.1016/j.eswa.2024.125641","url":null,"abstract":"<div><div>Cyberbullying behaviour has drawn more attention as social media usage has grown. Teen suicide has been related to cyberbullying, among other serious and harmful effects on a person’s life. Using the appropriate natural language processing and machine learning techniques, it is possible to proactively identify bullying content to reduce and eventually eradicate cyberbullying. Accordingly, the article proposed an automated deep-learning model for detecting aggressive activity in cyberbullying. Initially, the data was extracted from the social media platform using Formspring, Instagram and MySpace datasets for perceiving cyberbullying behaviour, then the collected data are input for preprocessing. To remove the raw data, several preprocessing processes have been introduced. They consist of removing stop words, white spaces for punctuation, and changing the comments to lowercase. Lexical Density (LD) has been one of the metrics used to gauge language complexity generally. As a result, the study made use of the Feature Density (FD) to calculate how complicated certain natural language datasets are using the linguistically backed preprocessing model. After preprocessing, the data are input to the feature selection process which selects the pertinent features or attributes to include in predictive modelling and which to leave out. Since, the article proposed a Binary Chimp Optimization (BCO)-based Feature Selection (BCO-FSS) technique, which selects the subset of features for classification performance improvement. The selected features are exploited for cyberbullying behaviour detection. To identify the exploit of social media for cyberbullying text content, the article suggested Stacked Bidirectional Gated Recurrent Unit (SBiGRU) Attention for learning spatial location information and sequential semantic representations using a Bi-GRU. Additionally, the BERT model is employed as a base classifier to recognize and categorise aggressive behaviour in the textual content. The Matlab software is employed for simulation. For accuracy, precision, recall, and F1-Score, this experiment yielded a practically perfect outcome with values of 99.12%, 94.73%, 97.45%, and 93.91% respectively.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"262 ","pages":"Article 125641"},"PeriodicalIF":7.5,"publicationDate":"2024-11-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662845","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Using cross-domain knowledge augmentation to explore comorbidity in electronic health records data 利用跨域知识扩增探索电子健康记录数据中的合并症

IF 7.5 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Expert Systems with Applications

Pub Date : 2024-11-05 DOI: 10.1016/j.eswa.2024.125644

Kaiyuan Zhang , Buyue Qian , Xiyuan Zhang , Qinghua Zheng

Research concentrating on specific diseases or employing single datasets, such as medical histories and thematic data, has garnered considerable attention. However, there has been limited investigation into comorbidities. Although some ad-hoc methods have been utilized in the medical field, there is a scarcity of systematic approaches to address this challenge. The task of expressing patient features using heterogeneous and cross-domain data presents considerable difficulties. Directly mapping this data into a matrix frequently results in issues such as high dimensionality, sparsity, redundancy, and noise. Additionally, given the critical role of supervisory information in medicine, acquiring accurate information is paramount. To address these issues, we propose an enhanced clustering method that capitalizes on cross-domain knowledge augmentation. This method can iteratively learn clustering outcomes and cross-domain knowledge. The cross-domain knowledge matrix produced by our approach can be interpreted as a measure of similarity between instances across domains. We validate our proposed model in real-world electronic health records (EHR) data, and achieve significant performance improvement compared to the baseline method, successfully completing the task of exploring comorbidity. Due to the privacy of EHR data, we also conduct extensive experiments on the publicly available datasets DBLP and UCI. The experimental results show that our algorithm is superior to the baseline algorithm and has strong generality.

专注于特定疾病或采用单一数据集（如病史和专题数据）的研究已引起了广泛关注。然而，对合并症的研究却十分有限。虽然在医疗领域已经使用了一些临时方法，但还缺乏系统的方法来应对这一挑战。使用异构和跨域数据来表达患者特征的任务相当困难。将这些数据直接映射到矩阵中经常会导致高维、稀疏、冗余和噪声等问题。此外，鉴于监督信息在医学中的关键作用，获取准确的信息至关重要。为了解决这些问题，我们提出了一种利用跨领域知识增强的增强聚类方法。这种方法可以反复学习聚类结果和跨领域知识。我们的方法产生的跨领域知识矩阵可以解释为跨领域实例之间相似性的度量。我们在真实世界的电子健康记录（EHR）数据中验证了我们提出的模型，与基线方法相比，性能有了显著提高，成功完成了探索合并症的任务。由于电子病历数据的隐私性，我们还在公开数据集 DBLP 和 UCI 上进行了大量实验。实验结果表明，我们的算法优于基线算法，并且具有很强的通用性。

{"title":"Using cross-domain knowledge augmentation to explore comorbidity in electronic health records data","authors":"Kaiyuan Zhang , Buyue Qian , Xiyuan Zhang , Qinghua Zheng","doi":"10.1016/j.eswa.2024.125644","DOIUrl":"10.1016/j.eswa.2024.125644","url":null,"abstract":"<div><div>Research concentrating on specific diseases or employing single datasets, such as medical histories and thematic data, has garnered considerable attention. However, there has been limited investigation into comorbidities. Although some ad-hoc methods have been utilized in the medical field, there is a scarcity of systematic approaches to address this challenge. The task of expressing patient features using heterogeneous and cross-domain data presents considerable difficulties. Directly mapping this data into a matrix frequently results in issues such as high dimensionality, sparsity, redundancy, and noise. Additionally, given the critical role of supervisory information in medicine, acquiring accurate information is paramount. To address these issues, we propose an enhanced clustering method that capitalizes on cross-domain knowledge augmentation. This method can iteratively learn clustering outcomes and cross-domain knowledge. The cross-domain knowledge matrix produced by our approach can be interpreted as a measure of similarity between instances across domains. We validate our proposed model in real-world electronic health records (EHR) data, and achieve significant performance improvement compared to the baseline method, successfully completing the task of exploring comorbidity. Due to the privacy of EHR data, we also conduct extensive experiments on the publicly available datasets DBLP and UCI. The experimental results show that our algorithm is superior to the baseline algorithm and has strong generality.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"263 ","pages":"Article 125644"},"PeriodicalIF":7.5,"publicationDate":"2024-11-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662079","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Enhancing rumor detection with data augmentation and generative pre-trained transformer 利用数据扩增和生成式预训练变换器加强谣言检测

IF 7.5 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Expert Systems with Applications

Pub Date : 2024-11-05 DOI: 10.1016/j.eswa.2024.125649

Mojgan Askarizade

The advent of social networks has facilitated the rapid dissemination of false information, including rumors, leading to significant societal and individual damages. Extensive research has been dedicated to rumor detection, ranging from machine learning techniques to neural networks. However, the existing methods could not learn the deep concepts of the rumor text to detect the rumor. In addition, imbalanced datasets in the rumor domain reduce the effectiveness of these algorithms. This study addresses this challenge by leveraging the Generative Pre-trained Transformer 2 (GPT-2) model to generate rumor-like texts, thus creating a balanced dataset. Subsequently, a novel approach for classifying rumor texts is proposed by modifying the GPT-2 model. We compare our results with state-of-art machine learning and deep learning methods as well as pre-trained models on the PHEME, Twitter15, and Twitter16 datasets. Our findings demonstrate that the proposed model, implementing advanced artificial intelligence techniques, has improved accuracy and F-measure in the application of detecting rumors compared to previous methods.

社交网络的出现促进了包括谣言在内的虚假信息的快速传播，给社会和个人造成了巨大损失。从机器学习技术到神经网络，人们对谣言检测进行了广泛的研究。然而，现有方法无法学习谣言文本的深层概念来检测谣言。此外，谣言领域的不平衡数据集也降低了这些算法的有效性。本研究利用生成预训练转换器 2（GPT-2）模型生成类似谣言的文本，从而创建了一个平衡的数据集，从而解决了这一难题。随后，通过修改 GPT-2 模型，提出了一种对谣言文本进行分类的新方法。我们在 PHEME、Twitter15 和 Twitter16 数据集上将我们的结果与最先进的机器学习和深度学习方法以及预训练模型进行了比较。我们的研究结果表明，与以前的方法相比，采用先进人工智能技术的拟议模型在检测谣言的应用中提高了准确性和 F-measure。

引用次数: 0

Enhancing inertial sensor-based sports activity recognition through reduction of the signals and deep learning 通过减少信号和深度学习提高基于惯性传感器的体育活动识别能力

IF 7.5 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Expert Systems with Applications

Pub Date : 2024-11-05 DOI: 10.1016/j.eswa.2024.125693

Pajak Grzegorz , Patalas-Maliszewska Justyna , Krutz Pascal , Rehm Matthias , Pajak Iwona , Schlegel Holger , Dix Martin

The increasing demand for sport training and health monitoring aligns with contemporary lifestyle trends. Developing a system to support sport training and verify exercise correctness can significantly enhance the acquisition of detailed, subject-specific data. This study aims to evaluate the accuracy of sport exercise recognition while minimizing the number of sensors required. The dataset includes 8,968 samples of exercises such as bar dips, squats, dips, lunges, pull-ups, sit-ups, and push-ups. Data were collected from 60 signals using mobile sensors positioned at the chest, right hand, and right foot of 21 subjects. The objective is to identify the most efficient method for recognizing these activities. Our methodology involved experiments with 21 participants in a custom-built setup and the application of deep learning techniques. Initially, a novel algorithm identified the optimal set of signals. We then tested two scenarios: training a Convolutional Neural Network (CNN) on raw signals and on pre-processed signals to reduce noise. The findings indicate that the magnetic field (MF) signal is crucial for recognizing exercises in both filtered and unfiltered data sets. The CNN’s accuracy was 2–3% higher with unfiltered data and remained robust at 93.7% for training and 90.0% for testing, despite a reduction in model complexity. This method’s practical implications are significant, enhancing sports training systems by reducing the number of sensors needed, thereby improving user comfort. Additionally, it contributes valuable insights to the field of Human Activity Recognition.

运动训练和健康监测的需求与日俱增，这与当代生活方式的发展趋势不谋而合。开发一个支持体育训练和验证运动正确性的系统，可以大大提高获取详细的、针对特定对象的数据的能力。本研究旨在评估体育锻炼识别的准确性，同时尽量减少所需的传感器数量。数据集包括 8968 个运动样本，如单杠下蹲、深蹲、俯卧撑、仰卧起坐、引体向上、仰卧起坐和俯卧撑。数据是使用安装在 21 名受试者胸部、右手和右脚的移动传感器从 60 个信号中收集的。目的是找出识别这些活动的最有效方法。我们的方法包括在定制的装置中对 21 名参与者进行实验，并应用深度学习技术。最初，一种新颖的算法确定了最佳信号集。然后，我们测试了两种情况：在原始信号上训练卷积神经网络（CNN），以及在预处理信号上训练卷积神经网络以减少噪音。研究结果表明，磁场（MF）信号对于识别过滤和未过滤数据集中的练习至关重要。在未过滤数据中，CNN 的准确率高出 2-3%，尽管模型复杂度有所降低，但训练和测试的准确率仍分别为 93.7% 和 90.0%。这种方法具有重要的实际意义，它通过减少所需的传感器数量来增强运动训练系统，从而提高用户的舒适度。此外，它还为人类活动识别领域提供了宝贵的见解。

{"title":"Enhancing inertial sensor-based sports activity recognition through reduction of the signals and deep learning","authors":"Pajak Grzegorz , Patalas-Maliszewska Justyna , Krutz Pascal , Rehm Matthias , Pajak Iwona , Schlegel Holger , Dix Martin","doi":"10.1016/j.eswa.2024.125693","DOIUrl":"10.1016/j.eswa.2024.125693","url":null,"abstract":"<div><div>The increasing demand for sport training and health monitoring aligns with contemporary lifestyle trends. Developing a system to support sport training and verify exercise correctness can significantly enhance the acquisition of detailed, subject-specific data. This study aims to evaluate the accuracy of sport exercise recognition while minimizing the number of sensors required. The dataset includes 8,968 samples of exercises such as bar dips, squats, dips, lunges, pull-ups, sit-ups, and push-ups. Data were collected from 60 signals using mobile sensors positioned at the chest, right hand, and right foot of 21 subjects. The objective is to identify the most efficient method for recognizing these activities. Our methodology involved experiments with 21 participants in a custom-built setup and the application of deep learning techniques. Initially, a novel algorithm identified the optimal set of signals. We then tested two scenarios: training a Convolutional Neural Network (CNN) on raw signals and on pre-processed signals to reduce noise. The findings indicate that the magnetic field (MF) signal is crucial for recognizing exercises in both filtered and unfiltered data sets. The CNN’s accuracy was 2–3% higher with unfiltered data and remained robust at 93.7% for training and 90.0% for testing, despite a reduction in model complexity. This method’s practical implications are significant, enhancing sports training systems by reducing the number of sensors needed, thereby improving user comfort. Additionally, it contributes valuable insights to the field of Human Activity Recognition.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"263 ","pages":"Article 125693"},"PeriodicalIF":7.5,"publicationDate":"2024-11-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662183","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0