Pub Date : 2023-08-31DOI: 10.9717/kmms.2023.26.8.1042
SukHyun Yu
In this study, we used Layer-wise Relevance Propagation (LRP) to analyze the level of contribution of input factors to the predictive results of the PM2.5 predictive model. First, we trained the DNN prediction model using data from 2015 to 2020, and then evaluated it using data from 2021. Next, we performed LRP on the evaluation data to analyze the importance of input factors in the prediction results. As a result, factors with consistently high importance regardless of concentration were O_TA, O_TD, O_RH, O_U, O_V, and O_PA, whereas PMSUB10/SUB and O_RN_ACC were observed to have lower importance. Furthermore, to analyze the characteristics of high-concentration data that are generally difficult to predict compared to low-concentration data, we divided the data by concentration and analyzed the importance of input factors. As a result, the importance of O_PMSUB2.5/SUB was high in the high concentration pattern and the importance of O_radiation was low, while the opposite trend was observed in the low concentration pattern. In particular, for high-concentration patterns that started suddenly and lasted more than three days, we analyzed the importance of input factors by time and factor. These high-concentration patterns with these characteristics showed significantly increased importance in the O_PMSUB2.5/SUB factor in the T12 interval closest to the prediction time, and it was observed that the importance of the F_PMSUB2.5/SUB factor increased slightly. Applying the factor importance results analyzed in this study to the PMSUB2.5/SUB prediction model is expected to improve prediction accuracy for high concentration patterns that are difficult to predict compared to general patterns.
{"title":"Analysis of Factor Importance of PM<SUB>2.5</SUB> High Concentration Case Using DNN and Layer-wise Relevance Propagation","authors":"SukHyun Yu","doi":"10.9717/kmms.2023.26.8.1042","DOIUrl":"https://doi.org/10.9717/kmms.2023.26.8.1042","url":null,"abstract":"In this study, we used Layer-wise Relevance Propagation (LRP) to analyze the level of contribution of input factors to the predictive results of the PM2.5 predictive model. First, we trained the DNN prediction model using data from 2015 to 2020, and then evaluated it using data from 2021. Next, we performed LRP on the evaluation data to analyze the importance of input factors in the prediction results. As a result, factors with consistently high importance regardless of concentration were O_TA, O_TD, O_RH, O_U, O_V, and O_PA, whereas PMSUB10/SUB and O_RN_ACC were observed to have lower importance. Furthermore, to analyze the characteristics of high-concentration data that are generally difficult to predict compared to low-concentration data, we divided the data by concentration and analyzed the importance of input factors. As a result, the importance of O_PMSUB2.5/SUB was high in the high concentration pattern and the importance of O_radiation was low, while the opposite trend was observed in the low concentration pattern. In particular, for high-concentration patterns that started suddenly and lasted more than three days, we analyzed the importance of input factors by time and factor. These high-concentration patterns with these characteristics showed significantly increased importance in the O_PMSUB2.5/SUB factor in the T12 interval closest to the prediction time, and it was observed that the importance of the F_PMSUB2.5/SUB factor increased slightly. Applying the factor importance results analyzed in this study to the PMSUB2.5/SUB prediction model is expected to improve prediction accuracy for high concentration patterns that are difficult to predict compared to general patterns.","PeriodicalId":16316,"journal":{"name":"Journal of Korea Multimedia Society","volume":"57 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135991533","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}
Pub Date : 2023-08-31DOI: 10.9717/kmms.2023.26.8.898
Myeong-Hun Bae, Jun Park, Se-Hoon Jung, Chun-Bo Sim
In fish farms, the overuse of feed can lead to residue and fish excrement polluting the water quality environment, thereby increasing the probability of pathogen proliferation and disease incidence in fish. In order to minimize the occurrence of diseases, it is crucial to administer an appropriate amount of feed and manage breeding diligently while mitigating any stress factors affecting the fish. This study involves the development of a fish species and disease detection system, where models are trained to identify different types of fish and their diseases. The system is designed to be used by fish farmers, offering a user-friendly interface through the Web. In the model training, the YOLOv7 model demonstrated high performance, achieving over 0.9 accuracy in detecting fish species. Meanwhile, for fish disease detection, the YOLOv5l model exhibited overall superior performance. However, there was a limitation in the dataset for fish disease detection, with only a small number of samples available. To overcome this, the fish species and disease detection system, developed in conjunction with the YOLOv5l model, was incorporated into the web page. This system aims to help identify the species, disease status, and specific affected regions in the fish population.
{"title":"Fish Species and Disease Detection System Using Deep Learning-Based Object Detection Model","authors":"Myeong-Hun Bae, Jun Park, Se-Hoon Jung, Chun-Bo Sim","doi":"10.9717/kmms.2023.26.8.898","DOIUrl":"https://doi.org/10.9717/kmms.2023.26.8.898","url":null,"abstract":"In fish farms, the overuse of feed can lead to residue and fish excrement polluting the water quality environment, thereby increasing the probability of pathogen proliferation and disease incidence in fish. In order to minimize the occurrence of diseases, it is crucial to administer an appropriate amount of feed and manage breeding diligently while mitigating any stress factors affecting the fish. This study involves the development of a fish species and disease detection system, where models are trained to identify different types of fish and their diseases. The system is designed to be used by fish farmers, offering a user-friendly interface through the Web. In the model training, the YOLOv7 model demonstrated high performance, achieving over 0.9 accuracy in detecting fish species. Meanwhile, for fish disease detection, the YOLOv5l model exhibited overall superior performance. However, there was a limitation in the dataset for fish disease detection, with only a small number of samples available. To overcome this, the fish species and disease detection system, developed in conjunction with the YOLOv5l model, was incorporated into the web page. This system aims to help identify the species, disease status, and specific affected regions in the fish population.","PeriodicalId":16316,"journal":{"name":"Journal of Korea Multimedia Society","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135991670","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}
Pub Date : 2023-08-31DOI: 10.9717/kmms.2023.26.8.874
Junyeong Kum, Taeyeon Kim, Myungho Lee
The demand for conversational systems using digital humans has increased across various fields, including education and psychological counseling. Conversational digital humans can be categorized into avatar and agent. This paper introduces the Hybrid Avatar Agent System (HAAS), designed to implement a conversational digital human capable of addressing diverse questions. To facilitate seamless transitions between avatars and agents, we developed a Korean Dialogue Breakdown Detection (DBD) model. While a digital human typically functions as an agent, the DBD model identifies conversational pauses, prompting human operator intervention for user assistance. We trained DBD model using the RoBERTa_base model, achieving an accuracy of 60.77%. In a user study, a comparison was drawn between HAAS and the Agent Only System (AOS). Users noted that HAAS provided more appropriate answers than AOS.
包括教育和心理咨询在内的各个领域对使用数字人的对话系统的需求都在增加。会说话的数字人可以分为化身和代理。本文介绍了混合化身代理系统(HAAS)，旨在实现一个能够解决各种问题的对话数字人。为了促进化身和代理之间的无缝转换，我们开发了一个韩文对话中断检测(DBD)模型。虽然数字人通常充当代理，但DBD模型识别会话暂停，提示人工操作员进行干预以提供用户帮助。我们使用RoBERTa_base模型训练DBD模型，准确率达到60.77%。在一项用户研究中，对HAAS和Agent Only System (AOS)进行了比较。用户注意到HAAS比AOS提供了更合适的答案。
{"title":"Evaluating the Usability of an LLM-Aided Hybrid Avatar Agent System","authors":"Junyeong Kum, Taeyeon Kim, Myungho Lee","doi":"10.9717/kmms.2023.26.8.874","DOIUrl":"https://doi.org/10.9717/kmms.2023.26.8.874","url":null,"abstract":"The demand for conversational systems using digital humans has increased across various fields, including education and psychological counseling. Conversational digital humans can be categorized into avatar and agent. This paper introduces the Hybrid Avatar Agent System (HAAS), designed to implement a conversational digital human capable of addressing diverse questions. To facilitate seamless transitions between avatars and agents, we developed a Korean Dialogue Breakdown Detection (DBD) model. While a digital human typically functions as an agent, the DBD model identifies conversational pauses, prompting human operator intervention for user assistance. We trained DBD model using the RoBERTa_base model, achieving an accuracy of 60.77%. In a user study, a comparison was drawn between HAAS and the Agent Only System (AOS). Users noted that HAAS provided more appropriate answers than AOS.","PeriodicalId":16316,"journal":{"name":"Journal of Korea Multimedia Society","volume":"68 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135991694","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}
Pub Date : 2023-07-31DOI: 10.9717/kmms.2023.26.7.833
Ji-Young Choi, Mahnwoo Kwon, Chee-Yong Kim, Mikyung Hwang
{"title":"Analysis of the Usability of Domestic Mobile Applications for Sexual Crime Prevention","authors":"Ji-Young Choi, Mahnwoo Kwon, Chee-Yong Kim, Mikyung Hwang","doi":"10.9717/kmms.2023.26.7.833","DOIUrl":"https://doi.org/10.9717/kmms.2023.26.7.833","url":null,"abstract":"","PeriodicalId":16316,"journal":{"name":"Journal of Korea Multimedia Society","volume":"213 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75008644","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}
Pub Date : 2023-07-31DOI: 10.9717/kmms.2023.26.7.813
Yae-Sop Lee
{"title":"Depth-guided Hole-filling Algorithm for View Synthesis","authors":"Yae-Sop Lee","doi":"10.9717/kmms.2023.26.7.813","DOIUrl":"https://doi.org/10.9717/kmms.2023.26.7.813","url":null,"abstract":"","PeriodicalId":16316,"journal":{"name":"Journal of Korea Multimedia Society","volume":"36 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"76678339","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}
Pub Date : 2023-07-31DOI: 10.9717/kmms.2023.26.7.805
JiHee Park, Y. Ha, J. Shim
{"title":"Face Recognition with Knowledge Distillation on Arcface","authors":"JiHee Park, Y. Ha, J. Shim","doi":"10.9717/kmms.2023.26.7.805","DOIUrl":"https://doi.org/10.9717/kmms.2023.26.7.805","url":null,"abstract":"","PeriodicalId":16316,"journal":{"name":"Journal of Korea Multimedia Society","volume":"140 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"84223767","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}
Pub Date : 2023-07-31DOI: 10.9717/kmms.2023.26.7.850
Haram Choi, S. Nam
{"title":"Development of Mixed Reality Regional-Culture Realistic Content based on Affordance for User Participation","authors":"Haram Choi, S. Nam","doi":"10.9717/kmms.2023.26.7.850","DOIUrl":"https://doi.org/10.9717/kmms.2023.26.7.850","url":null,"abstract":"","PeriodicalId":16316,"journal":{"name":"Journal of Korea Multimedia Society","volume":"1 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74428938","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}
Pub Date : 2023-07-31DOI: 10.9717/kmms.2023.26.7.821
Jong-hee Kim, H. Lee
{"title":"Comparison of Usability Evaluation of Mobile GUI Components by Operating System","authors":"Jong-hee Kim, H. Lee","doi":"10.9717/kmms.2023.26.7.821","DOIUrl":"https://doi.org/10.9717/kmms.2023.26.7.821","url":null,"abstract":"","PeriodicalId":16316,"journal":{"name":"Journal of Korea Multimedia Society","volume":"222 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75900708","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}
Pub Date : 2023-07-31DOI: 10.9717/kmms.2023.26.7.842
Young-Suk Lee
{"title":"A Study on the Actor Recommendation Algorithm for Animal Characters","authors":"Young-Suk Lee","doi":"10.9717/kmms.2023.26.7.842","DOIUrl":"https://doi.org/10.9717/kmms.2023.26.7.842","url":null,"abstract":"","PeriodicalId":16316,"journal":{"name":"Journal of Korea Multimedia Society","volume":"186 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"77037954","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}
Pub Date : 2023-06-30DOI: 10.9717/kmms.2023.26.6.753
H. Jeon, Dong-min Cho
{"title":"The Retargeting of Character Animation Using Deep Learning Cloud Service Technology","authors":"H. Jeon, Dong-min Cho","doi":"10.9717/kmms.2023.26.6.753","DOIUrl":"https://doi.org/10.9717/kmms.2023.26.6.753","url":null,"abstract":"","PeriodicalId":16316,"journal":{"name":"Journal of Korea Multimedia Society","volume":"10 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"73780388","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：