Barbell trajectory provides much kinematic information which can indicate the lifter's performance. However, kinematic parameters are not only gathering difficult but also hard to understand. This paper proposes a barbell trajectory evaluation inference that indicates the lifter's snatch performance from the barbell trajectory. We gathered four competitions and obtained the barbell trajectories from each lifter's attempt. Furthermore, five weightlifting experts recruit to indicate the performance categories, which are goodlift-good posture, goodlift-normal posture, nolift-good posture, and nolift-normal posture, as our data label. VGG16 convolution neural network utilize in our trajectory evaluation inference. The accuracy of the proposed inference is approximate 71.11%. From these results, our proposed barbell trajectory inference can provide a high accuracy performance evaluator for athlete self-training and competition performance analysis.
{"title":"Machine Learning for the Posture Evaluation of Women Snatch Barbell Trajectory","authors":"Jen-Shi Chen, Ching-Ting Hsu, Wei-Hua Ho, Chiao-Yin Hsu","doi":"10.1145/3571560.3571567","DOIUrl":"https://doi.org/10.1145/3571560.3571567","url":null,"abstract":"Barbell trajectory provides much kinematic information which can indicate the lifter's performance. However, kinematic parameters are not only gathering difficult but also hard to understand. This paper proposes a barbell trajectory evaluation inference that indicates the lifter's snatch performance from the barbell trajectory. We gathered four competitions and obtained the barbell trajectories from each lifter's attempt. Furthermore, five weightlifting experts recruit to indicate the performance categories, which are goodlift-good posture, goodlift-normal posture, nolift-good posture, and nolift-normal posture, as our data label. VGG16 convolution neural network utilize in our trajectory evaluation inference. The accuracy of the proposed inference is approximate 71.11%. From these results, our proposed barbell trajectory inference can provide a high accuracy performance evaluator for athlete self-training and competition performance analysis.","PeriodicalId":143909,"journal":{"name":"Proceedings of the 6th International Conference on Advances in Artificial Intelligence","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124604015","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Gergely Pap, Krisztian Adam, Zoltan Gyorgypal, Laszlo Toth, Z. Hegedus
Classifying DNA sequences based on a nucleotide representation has enjoyed considerable success with the advancement of Deep Learning, as the proper usage and combination of different layers and architecture choices led to an increase in performance. The most common approaches rely on convolutional, recurrent and attention layer types. Moreover, the inclusion of further information in addition to the nucleotide sequence provides increases in performance, even though the methods of combining the input feature representations with distinct model structures could pose a challenge. To examine this topic, we applied depthwise separable convolutional layers to a physicochemical DNA sequence representation and train models to detect the binding sites of DNA binding proteins. While convolutional kernels learn the local feature patterns of motifs, the behaviour of the depthwise separable convolution better exploits the feature, shape and physicochemical information that could be stored in the input representation. Our models with depthwise separable convolution achieve increases in accuracy compared to the convolutional and nucleotide-based approaches on several datasets.
{"title":"Depthwise Convolutions using Physicochemical Features of DNA for Transcription Factor Binding Site Classification: Physicochemical Features for DNA-Protein Classification with Depthwise Convolutions","authors":"Gergely Pap, Krisztian Adam, Zoltan Gyorgypal, Laszlo Toth, Z. Hegedus","doi":"10.1145/3571560.3571563","DOIUrl":"https://doi.org/10.1145/3571560.3571563","url":null,"abstract":"Classifying DNA sequences based on a nucleotide representation has enjoyed considerable success with the advancement of Deep Learning, as the proper usage and combination of different layers and architecture choices led to an increase in performance. The most common approaches rely on convolutional, recurrent and attention layer types. Moreover, the inclusion of further information in addition to the nucleotide sequence provides increases in performance, even though the methods of combining the input feature representations with distinct model structures could pose a challenge. To examine this topic, we applied depthwise separable convolutional layers to a physicochemical DNA sequence representation and train models to detect the binding sites of DNA binding proteins. While convolutional kernels learn the local feature patterns of motifs, the behaviour of the depthwise separable convolution better exploits the feature, shape and physicochemical information that could be stored in the input representation. Our models with depthwise separable convolution achieve increases in accuracy compared to the convolutional and nucleotide-based approaches on several datasets.","PeriodicalId":143909,"journal":{"name":"Proceedings of the 6th International Conference on Advances in Artificial Intelligence","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131221033","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Atharv Nagrikar* Project Engineer, High-Performance Computing Technologies Group, Centre for Development of Advanced Computing (C-DAC) Pune, India Rahul Singh Dangi Senior Technical Officer, High-Performance Computing Technologies Group, Centre for Development of Advanced Computing (C-DAC) Pune, India Samrit Kumar Maity Joint Director, High-Performance Computing Technologies Group, Centre for Development of Advanced Computing (C-DAC) Pune, India
{"title":"Incremental Learning of Classification models in Deep Learning","authors":"Atharv Nagarikar, Rahul Singh Dangi, Samrit Kumar Maity, Ashish Kuvelkar, Sanjay Wandhekar","doi":"10.1145/3571560.3571568","DOIUrl":"https://doi.org/10.1145/3571560.3571568","url":null,"abstract":"Atharv Nagrikar* Project Engineer, High-Performance Computing Technologies Group, Centre for Development of Advanced Computing (C-DAC) Pune, India Rahul Singh Dangi Senior Technical Officer, High-Performance Computing Technologies Group, Centre for Development of Advanced Computing (C-DAC) Pune, India Samrit Kumar Maity Joint Director, High-Performance Computing Technologies Group, Centre for Development of Advanced Computing (C-DAC) Pune, India","PeriodicalId":143909,"journal":{"name":"Proceedings of the 6th International Conference on Advances in Artificial Intelligence","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130733120","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Machine Learning, a subtype of AI, enables computers to mimic human behavior without explicit programming. Machine learning models aren't used very often in diagnostic imaging because there isn't enough knowledge and resources to do so. Hence, this study aims to apply automated machine learning to the diagnosis of medical images to make machine learning more accessible to non-experts. In this study, a dataset containing 2313 images each of covid-19, pneumonia and normal chest x-rays were selected and divided into testing, training, and validation datasets. The AutoGluon library was used to train and produce a model that would classify an input image and infer the probable diagnosis from the diseases it was trained upon. This study can prove that applying hyperparameter optimization and neural architecture search is able to produce high accuracy models for medical image diagnosis.
{"title":"AutoMID : A Novel Framework For Automated Computer Aided Diagnosis Of Medical Images","authors":"Ayeshmantha Wijegunathileke, A. Aponso","doi":"10.1145/3571560.3571571","DOIUrl":"https://doi.org/10.1145/3571560.3571571","url":null,"abstract":"Machine Learning, a subtype of AI, enables computers to mimic human behavior without explicit programming. Machine learning models aren't used very often in diagnostic imaging because there isn't enough knowledge and resources to do so. Hence, this study aims to apply automated machine learning to the diagnosis of medical images to make machine learning more accessible to non-experts. In this study, a dataset containing 2313 images each of covid-19, pneumonia and normal chest x-rays were selected and divided into testing, training, and validation datasets. The AutoGluon library was used to train and produce a model that would classify an input image and infer the probable diagnosis from the diseases it was trained upon. This study can prove that applying hyperparameter optimization and neural architecture search is able to produce high accuracy models for medical image diagnosis.","PeriodicalId":143909,"journal":{"name":"Proceedings of the 6th International Conference on Advances in Artificial Intelligence","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127127548","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Breast cancer is one of the diseases that led to a huge number of deaths in the recent decades. One of the major issues that affect the recovery procedure is the early detection of the disease. Thus, in this paper, several machine learning algorithms that support the early detection process, along with the impact on combining these algorithms with hyperparameter tuning optimization techniques will be presented. Moreover, we conducted a comparison among proposed techniques to figure out which classifier model can achieve better detection accuracy of the disease.
{"title":"A Comprehensive Study on Machine Learning in Breast Cancer Detection and Classification","authors":"A. B. Nassif, Aya Al-Chikh Omar","doi":"10.1145/3571560.3571572","DOIUrl":"https://doi.org/10.1145/3571560.3571572","url":null,"abstract":"Breast cancer is one of the diseases that led to a huge number of deaths in the recent decades. One of the major issues that affect the recovery procedure is the early detection of the disease. Thus, in this paper, several machine learning algorithms that support the early detection process, along with the impact on combining these algorithms with hyperparameter tuning optimization techniques will be presented. Moreover, we conducted a comparison among proposed techniques to figure out which classifier model can achieve better detection accuracy of the disease.","PeriodicalId":143909,"journal":{"name":"Proceedings of the 6th International Conference on Advances in Artificial Intelligence","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127365849","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The airport industry is a highly competitive market that has expanded quickly during the last two decades. Airport management usually measures the level of passenger satisfaction by applying the traditional methods, such as user surveys and expert opinions, which require time and effort to analyse. Recently, there has been considerable attention on employing machine learning techniques and sentiment analysis for measuring the level of passenger satisfaction. Sentiment analysis can be implemented using a range of different methods. However, it is still uncertain which techniques are better suited for recognising the sentiment for a particular subject domain or dataset. In this paper, we analyse the sentiment of air travellers using five different algorithms, namely Logistic Regression, XGBoost, Support Vector Machine, Random Forest and Naïve Bayes. We obtain our data set through the SKYTRAX website which is a collection of reviews of around 600 airports. We apply some pre-processing steps, such as converting the textual reviews into numerical form, by using the term frequency-inverse document frequency. We also remove stopwords from the text using the NLTK list of stopwords. We evaluate our results using the accuracy, precision, recall and F1_score performance metrics. Our analysis shows that XGBoost provides the most accurate results when compared with other algorithms.
{"title":"Measuring Airport Service Quality Using Machine Learning Algorithms","authors":"Mohammed Salih Homaid, I. Moulitsas","doi":"10.1145/3571560.3571562","DOIUrl":"https://doi.org/10.1145/3571560.3571562","url":null,"abstract":"The airport industry is a highly competitive market that has expanded quickly during the last two decades. Airport management usually measures the level of passenger satisfaction by applying the traditional methods, such as user surveys and expert opinions, which require time and effort to analyse. Recently, there has been considerable attention on employing machine learning techniques and sentiment analysis for measuring the level of passenger satisfaction. Sentiment analysis can be implemented using a range of different methods. However, it is still uncertain which techniques are better suited for recognising the sentiment for a particular subject domain or dataset. In this paper, we analyse the sentiment of air travellers using five different algorithms, namely Logistic Regression, XGBoost, Support Vector Machine, Random Forest and Naïve Bayes. We obtain our data set through the SKYTRAX website which is a collection of reviews of around 600 airports. We apply some pre-processing steps, such as converting the textual reviews into numerical form, by using the term frequency-inverse document frequency. We also remove stopwords from the text using the NLTK list of stopwords. We evaluate our results using the accuracy, precision, recall and F1_score performance metrics. Our analysis shows that XGBoost provides the most accurate results when compared with other algorithms.","PeriodicalId":143909,"journal":{"name":"Proceedings of the 6th International Conference on Advances in Artificial Intelligence","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115891823","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
A deep understanding of the brain can lead to significant breakthroughs in Artificial Intelligence. Many researchers concentrate their efforts on simulating the human mind to comprehend its complexities better. With the intention of better understanding the episodic memory aspect of the human mind, we propose a deep learning model to implement the detection and retrieval properties of human episodic memory, a part of long-term memory. A model based on LSTM and CNN is proposed, which follows the architectural methodology of Rosenblatt’s experiential memory model. A comparison of detection efficiency and accuracy and the proposed model’s retrieval property with a recently suggested method demonstrate its effectiveness and superiority.
{"title":"An Effective Implementation of Detection and Retrieval Property of Episodic Memory","authors":"Aniket Sharma, Pramod Kumar Singh, J. Prakash","doi":"10.1145/3571560.3571582","DOIUrl":"https://doi.org/10.1145/3571560.3571582","url":null,"abstract":"A deep understanding of the brain can lead to significant breakthroughs in Artificial Intelligence. Many researchers concentrate their efforts on simulating the human mind to comprehend its complexities better. With the intention of better understanding the episodic memory aspect of the human mind, we propose a deep learning model to implement the detection and retrieval properties of human episodic memory, a part of long-term memory. A model based on LSTM and CNN is proposed, which follows the architectural methodology of Rosenblatt’s experiential memory model. A comparison of detection efficiency and accuracy and the proposed model’s retrieval property with a recently suggested method demonstrate its effectiveness and superiority.","PeriodicalId":143909,"journal":{"name":"Proceedings of the 6th International Conference on Advances in Artificial Intelligence","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114746211","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Geonsik Youn, Bohan Yoon, Seungbin Ji, Dahee Ko, J. Rhee
Named Entity Recognition (NER) is one of the first stages in deep natural language understanding. The state-of-the-art deep NER models are dependent on high-quality and massive datasets. Also, the NER tasks require token-level labels. For this reason, there is a problem that annotating many sentences for the NER tasks is time-consuming and expensive. To solve this problem, many prior studies have been conducted to use the auto annotated weakly labeled data. However, the weakly labeled data contains a lot of noises that are obstructive to the training of NER models. We propose to use MixUp and cross-consistency training (CCT) together as a strategy to use weakly labeled data for NER tasks. In this study, the proposed method stems from the idea that MixUp, which was recently considered the data augmentation strategy, hinders the NER model training. Inspired by this point, we propose to use MixUp as a perturbation of cross-consistency training for NER. Experiments conducted on several NER benchmarks demonstrate the proposed method achieves improved performance compared to employing only a few human-annotated data.
{"title":"MixUp based Cross-Consistency Training for Named Entity Recognition","authors":"Geonsik Youn, Bohan Yoon, Seungbin Ji, Dahee Ko, J. Rhee","doi":"10.1145/3571560.3571576","DOIUrl":"https://doi.org/10.1145/3571560.3571576","url":null,"abstract":"Named Entity Recognition (NER) is one of the first stages in deep natural language understanding. The state-of-the-art deep NER models are dependent on high-quality and massive datasets. Also, the NER tasks require token-level labels. For this reason, there is a problem that annotating many sentences for the NER tasks is time-consuming and expensive. To solve this problem, many prior studies have been conducted to use the auto annotated weakly labeled data. However, the weakly labeled data contains a lot of noises that are obstructive to the training of NER models. We propose to use MixUp and cross-consistency training (CCT) together as a strategy to use weakly labeled data for NER tasks. In this study, the proposed method stems from the idea that MixUp, which was recently considered the data augmentation strategy, hinders the NER model training. Inspired by this point, we propose to use MixUp as a perturbation of cross-consistency training for NER. Experiments conducted on several NER benchmarks demonstrate the proposed method achieves improved performance compared to employing only a few human-annotated data.","PeriodicalId":143909,"journal":{"name":"Proceedings of the 6th International Conference on Advances in Artificial Intelligence","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130433218","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In the context of a fast-aging population, the health problem of the elderly has become a hot spot in the world [1]. In terms of long-term care, wearable health monitoring technology is an excellent way to address these issues [2]. As a kind of wearable system, the smart vest is used as the object of this study. In order to encourage the elderly to use and wear smart vests, this study investigated the factors affecting the acceptance of smart vests; moreover, it developed the technology acceptance model of smart vests for the elderly population. The model was certified with a sample of 152 older adults aged 60 and above. The results show that aesthetics as an external factor has a significant positive impact on perceived usefulness and ease of use. Perceived usefulness positively impacts the attitude of the elderly to use smart vests. This research provides valuable insights for future researchers and practitioners to improve the acceptance of smart vests among older adults.
{"title":"Design for the elderly: the acceptance of smart vests in the senior population","authors":"Zhuozhen Xie","doi":"10.1145/3571560.3571580","DOIUrl":"https://doi.org/10.1145/3571560.3571580","url":null,"abstract":"In the context of a fast-aging population, the health problem of the elderly has become a hot spot in the world [1]. In terms of long-term care, wearable health monitoring technology is an excellent way to address these issues [2]. As a kind of wearable system, the smart vest is used as the object of this study. In order to encourage the elderly to use and wear smart vests, this study investigated the factors affecting the acceptance of smart vests; moreover, it developed the technology acceptance model of smart vests for the elderly population. The model was certified with a sample of 152 older adults aged 60 and above. The results show that aesthetics as an external factor has a significant positive impact on perceived usefulness and ease of use. Perceived usefulness positively impacts the attitude of the elderly to use smart vests. This research provides valuable insights for future researchers and practitioners to improve the acceptance of smart vests among older adults.","PeriodicalId":143909,"journal":{"name":"Proceedings of the 6th International Conference on Advances in Artificial Intelligence","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125608476","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In various areas of study, standard established procedures are critical for the successful accomplishment of a kinaesthetic task. Such standard procedures are important in various industries like engineering and health. This study makes a case for the development of intelligent activity monitoring systems for learning purposes through a proof of concept in first-aid training. Minor accidents such as simple cuts, bruises and minor burns are frequently treated without the need of emergency medical services. However, an incorrect first-aid procedure may lead to medical complications. This study aims to aid a learner to train how to perform a first-aid procedure for treating a wound through real-time monitoring, instructions and feedback. We propose a three-phase system where fast object detection, activity recognition in a temporal dimension and sequencing are used to semantically understand leaner actions. The You Only Look Once (YOLOv5) was used in phase 1 to detect multiple objects like wounds and bandages and Mediapipe to detect hand landmarks. Each class was assigned a different threshold for more accurate detections. The object detection model achieved a mean Average Precision (mAP) of 72.74% on the validation set and was subsequently used in a temporal manner to recognize an action. This temporal method to recognize the action of applying pressure over a wound, achieved an F1-Score of 91.67%. The method using an ontology-based technique to recognize the action of applying a bandage, achieved an F1-Score of 90.91%. The optimum distance from camera was found to be the actor placed at a position where the arm of the wounded actor occupies a significant portion of the viewport, whilst the optimum camera angle was found to be 110°. The created sequencing algorithm was tested using three different scenarios with the aid of a number of participants. The overall accuracy was 83.33%, wherein the result highlights that the algorithm is able to identify the sequence being conducted even with minimal movement involved during bandage application. The proposed system has high prospects of addressing challenges in a real-world environment.
在不同的研究领域，标准的既定程序是成功完成动觉任务的关键。这样的标准程序在工程和卫生等各个行业都很重要。本研究通过在急救培训中的概念验证，为学习目的的智能活动监测系统的发展提出了一个案例。诸如简单的割伤、瘀伤和轻微烧伤等轻微事故往往不需要紧急医疗服务就能得到治疗。然而，不正确的急救程序可能导致医学并发症。本研究旨在通过实时监测、指导和反馈来帮助学习者训练如何执行急救程序来处理伤口。我们提出了一个三相系统，其中使用快速对象检测，时间维度的活动识别和排序来从语义上理解更精简的动作。第一阶段使用You Only Look Once (YOLOv5)来检测多个物体，如伤口和绷带，使用Mediapipe来检测手部地标。为了更准确的检测，每个类别都被分配了不同的阈值。目标检测模型在验证集上的平均精度(mAP)达到72.74%，随后以时间方式用于识别动作。该方法用于识别在创面上施加压力的动作，f1评分为91.67%。该方法使用基于本体的技术来识别绷带的动作，获得了90.91%的f1评分。我们发现，演员与摄像机的最佳距离是放置在受伤演员的手臂占据视口很大一部分的位置，而最佳摄像机角度是110°。在许多参与者的帮助下，用三种不同的场景测试了所创建的排序算法。总体准确率为83.33%，其中结果突出表明，即使在绷带应用过程中涉及的最小运动，该算法也能够识别正在进行的序列。所提出的系统在解决现实环境中的挑战方面具有很高的前景。
{"title":"A semantic real-time activity recognition system for sequential procedures in vocational learning","authors":"J. Magro, Daren Scerri","doi":"10.1145/3571560.3571579","DOIUrl":"https://doi.org/10.1145/3571560.3571579","url":null,"abstract":"In various areas of study, standard established procedures are critical for the successful accomplishment of a kinaesthetic task. Such standard procedures are important in various industries like engineering and health. This study makes a case for the development of intelligent activity monitoring systems for learning purposes through a proof of concept in first-aid training. Minor accidents such as simple cuts, bruises and minor burns are frequently treated without the need of emergency medical services. However, an incorrect first-aid procedure may lead to medical complications. This study aims to aid a learner to train how to perform a first-aid procedure for treating a wound through real-time monitoring, instructions and feedback. We propose a three-phase system where fast object detection, activity recognition in a temporal dimension and sequencing are used to semantically understand leaner actions. The You Only Look Once (YOLOv5) was used in phase 1 to detect multiple objects like wounds and bandages and Mediapipe to detect hand landmarks. Each class was assigned a different threshold for more accurate detections. The object detection model achieved a mean Average Precision (mAP) of 72.74% on the validation set and was subsequently used in a temporal manner to recognize an action. This temporal method to recognize the action of applying pressure over a wound, achieved an F1-Score of 91.67%. The method using an ontology-based technique to recognize the action of applying a bandage, achieved an F1-Score of 90.91%. The optimum distance from camera was found to be the actor placed at a position where the arm of the wounded actor occupies a significant portion of the viewport, whilst the optimum camera angle was found to be 110°. The created sequencing algorithm was tested using three different scenarios with the aid of a number of participants. The overall accuracy was 83.33%, wherein the result highlights that the algorithm is able to identify the sequence being conducted even with minimal movement involved during bandage application. The proposed system has high prospects of addressing challenges in a real-world environment.","PeriodicalId":143909,"journal":{"name":"Proceedings of the 6th International Conference on Advances in Artificial Intelligence","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114133171","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：