Pub Date : 2014-08-26DOI: 10.1109/SIBGRAPI.2014.26
M. Kurzynski, Maciej Krysmann
The concept of classifier competence in the feature space is fundamental to dynamic classifier selection in multiple classifier systems (MCS). Competence function (measure) of base classifier can be determined using validation set in the two step procedure. The first step consists in creating competence set, i.e. the set of classifier competences for all validation objects. To this end a hypothetical classifier called randomized reference classifier (RRC) is constructed. Since RRC - on average - acts like the evaluated classifier, the competence of the classifier at validation point is calculated as the probability of correct classification at this point of the respective RRC. In the second step, the competences calculated for a validation set are generalised to an entire feature space by constructing a competence function based on a supervised learning procedure. In this study, the second step of the above procedure is addressed by developing the fuzzy inference methods of learning competence functions. Two fuzzy inference systems are developed and applied to the supervised learning competence function of base classifiers in a MCS system with dynamic classifier selection (DCS) and dynamic ensemble selection (DES) scheme: Mamdani fuzzy inference system and Sugeno fuzzy inference system. Both fuzzy inference systems were experimentally tested and compared against 4 literature methods of learning classifier competence (potential function, regression model, multilayer perceptron, k-nearest neighbor scheme) using 9 databases taken from the UCI Machine Learning Repository. The experimental results clearly show the effectiveness of the proposed supervised learning competence function using fuzzy inference systems regardless of the ensemble type used (homogeneous or heterogeneous).
{"title":"Fuzzy Inference Methods Applied to the Learning Competence Measure in Dynamic Classifier Selection","authors":"M. Kurzynski, Maciej Krysmann","doi":"10.1109/SIBGRAPI.2014.26","DOIUrl":"https://doi.org/10.1109/SIBGRAPI.2014.26","url":null,"abstract":"The concept of classifier competence in the feature space is fundamental to dynamic classifier selection in multiple classifier systems (MCS). Competence function (measure) of base classifier can be determined using validation set in the two step procedure. The first step consists in creating competence set, i.e. the set of classifier competences for all validation objects. To this end a hypothetical classifier called randomized reference classifier (RRC) is constructed. Since RRC - on average - acts like the evaluated classifier, the competence of the classifier at validation point is calculated as the probability of correct classification at this point of the respective RRC. In the second step, the competences calculated for a validation set are generalised to an entire feature space by constructing a competence function based on a supervised learning procedure. In this study, the second step of the above procedure is addressed by developing the fuzzy inference methods of learning competence functions. Two fuzzy inference systems are developed and applied to the supervised learning competence function of base classifiers in a MCS system with dynamic classifier selection (DCS) and dynamic ensemble selection (DES) scheme: Mamdani fuzzy inference system and Sugeno fuzzy inference system. Both fuzzy inference systems were experimentally tested and compared against 4 literature methods of learning classifier competence (potential function, regression model, multilayer perceptron, k-nearest neighbor scheme) using 9 databases taken from the UCI Machine Learning Repository. The experimental results clearly show the effectiveness of the proposed supervised learning competence function using fuzzy inference systems regardless of the ensemble type used (homogeneous or heterogeneous).","PeriodicalId":146229,"journal":{"name":"2014 27th SIBGRAPI Conference on Graphics, Patterns and Images","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2014-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133333314","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 : 2014-08-01DOI: 10.1109/SIBGRAPI.2014.18
Fabiane Queiroz, Ing Ren Tsang
Endoscopy is a minimally invasive medical diagnostic procedure that is used to provide a realistic view of the surfaces of organs inside human body. Images taken during such procedures largely show tissues of human organs. Due to the presence of mucosa of the gastrointestinal tract or other characteristics of the human body, these surfaces usually have a glossy appearance showing specular reflections. For many image analysis algorithms, these distinct and bright visual mark can be a significant source of error. On other hand, these features can also be useful for image restoration and for the construction of 3D model of the organs. In this article, we propose a segmentation method of the specular regions based on sparse and low-rank decomposition using a robust PCA via accelerated proximal gradient algorithm. In contrast to the existing approaches, the proposed segmentation works without using colour image thresholds. Moreover, the proposed method presents more precise segmentation results represented by grayscale masks instead of binary masks.
{"title":"Automatic Segmentation of Specular Reflections for Endoscopic Images Based on Sparse and Low-Rank Decomposition","authors":"Fabiane Queiroz, Ing Ren Tsang","doi":"10.1109/SIBGRAPI.2014.18","DOIUrl":"https://doi.org/10.1109/SIBGRAPI.2014.18","url":null,"abstract":"Endoscopy is a minimally invasive medical diagnostic procedure that is used to provide a realistic view of the surfaces of organs inside human body. Images taken during such procedures largely show tissues of human organs. Due to the presence of mucosa of the gastrointestinal tract or other characteristics of the human body, these surfaces usually have a glossy appearance showing specular reflections. For many image analysis algorithms, these distinct and bright visual mark can be a significant source of error. On other hand, these features can also be useful for image restoration and for the construction of 3D model of the organs. In this article, we propose a segmentation method of the specular regions based on sparse and low-rank decomposition using a robust PCA via accelerated proximal gradient algorithm. In contrast to the existing approaches, the proposed segmentation works without using colour image thresholds. Moreover, the proposed method presents more precise segmentation results represented by grayscale masks instead of binary masks.","PeriodicalId":146229,"journal":{"name":"2014 27th SIBGRAPI Conference on Graphics, Patterns and Images","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2014-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127659029","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 : 2014-06-07DOI: 10.1109/SIBGRAPI.2014.15
P. Freitas, Mylène C. Q. Farias, Aleteia P. F. Araujo
In this paper, we propose a framework for acquiring super-resolution videos from low-resolution originals. Given that super-resolution conversion algorithms require a large amount of data processing, the proposed framework uses a set of strategies to improve performance and computational efficiency. The strategies consists of a combination of data simplification and parallel processing techniques. The simplification strategies are used to decrease the amount of data to process and, consequently, the required processing time. The parallel processing techniques are designed so that major modifications of the super-resolution algorithms are not required. The framework is fast and makes the video resolution increase timely.
{"title":"A Parallel Framework for Video Super-Resolution","authors":"P. Freitas, Mylène C. Q. Farias, Aleteia P. F. Araujo","doi":"10.1109/SIBGRAPI.2014.15","DOIUrl":"https://doi.org/10.1109/SIBGRAPI.2014.15","url":null,"abstract":"In this paper, we propose a framework for acquiring super-resolution videos from low-resolution originals. Given that super-resolution conversion algorithms require a large amount of data processing, the proposed framework uses a set of strategies to improve performance and computational efficiency. The strategies consists of a combination of data simplification and parallel processing techniques. The simplification strategies are used to decrease the amount of data to process and, consequently, the required processing time. The parallel processing techniques are designed so that major modifications of the super-resolution algorithms are not required. The framework is fast and makes the video resolution increase timely.","PeriodicalId":146229,"journal":{"name":"2014 27th SIBGRAPI Conference on Graphics, Patterns and Images","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2014-06-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125641685","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：