With the proliferation of image data, the need to search and retrieve images efficiently and accurately from a large image database or a collection of image databases has drastically increased. To address such a demand, a unified framework called Markov Model Mediators (MMMs) is proposed in this paper to facilitate conceptual database clustering and to improve the query processing performance by analyzing the summarized knowledge. The unique characteristics of MMMs are that it provides the capabilities of exploring the affinity relations among the images at the database level and among the databases at the cluster level respectively, using an effective data mining process. At the database level, each database is modeled by an intra-database MMM which enables accurate image retrieval within the database. Then the conceptual database clustering is performed and cluster-level knowledge summarization is conducted to reduce the cost of retrieving images across the databases. This framework has been tested using a set of image databases, which contain various numbers of images with different dimensions and concept categories. The experimental results demonstrate that our framework achieves better retrieval accuracy via inter-cluster retrieval than that of intra-cluster retrieval with minimal extra effort.
随着图像数据的激增，从大型图像数据库或图像数据库集合中高效、准确地搜索和检索图像的需求急剧增加。针对这一需求，本文提出了一种统一的马尔可夫模型中介(Markov Model mediator, MMMs)框架，以促进概念数据库聚类，并通过分析汇总的知识来提高查询处理性能。mm的独特之处在于，它提供了使用有效的数据挖掘过程分别在数据库级别和集群级别探索图像之间和数据库之间的亲和关系的能力。在数据库级别上，每个数据库都由数据库内的MMM建模，从而在数据库内实现准确的图像检索。然后进行概念数据库聚类，并进行聚类级知识汇总，以降低跨数据库检索图像的成本。这个框架已经使用一组图像数据库进行了测试，这些数据库包含不同尺寸和概念类别的不同数量的图像。实验结果表明，该框架通过簇间检索比簇内检索获得了更好的检索精度。
{"title":"A unified framework for image database clustering and content-based retrieval","authors":"M. Shyu, Shu‐Ching Chen, Min Chen, Chengcui Zhang","doi":"10.1145/1032604.1032609","DOIUrl":"https://doi.org/10.1145/1032604.1032609","url":null,"abstract":"With the proliferation of image data, the need to search and retrieve images efficiently and accurately from a large image database or a collection of image databases has drastically increased. To address such a demand, a unified framework called <i>Markov Model Mediators</i> (MMMs) is proposed in this paper to facilitate conceptual database clustering and to improve the query processing performance by analyzing the summarized knowledge. The unique characteristics of MMMs are that it provides the capabilities of exploring the affinity relations among the images at the database level and among the databases at the cluster level respectively, using an effective data mining process. At the database level, each database is modeled by an intra-database MMM which enables accurate image retrieval within the database. Then the conceptual database clustering is performed and cluster-level knowledge summarization is conducted to reduce the cost of retrieving images across the databases. This framework has been tested using a set of image databases, which contain various numbers of images with different dimensions and concept categories. The experimental results demonstrate that our framework achieves better retrieval accuracy via inter-cluster retrieval than that of intra-cluster retrieval with minimal extra effort.","PeriodicalId":415406,"journal":{"name":"ACM International Workshop on Multimedia Databases","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2004-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125843194","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}
This paper presents the use of relevance feedback to the problem of content-based sub-image retrieval (CBsIR). Relevance feedback is used to improve the accuracy of successive retrievals via a tile re-weighting scheme that assigns penalties to each tile of database images and updates the tile penalties for all relevant images retrieved at each iteration using both the relevant (positive) and irrelevant (negative) images identified by the user. Performance evaluation on a dataset of over 10,000 images shows the effectiveness and efficiency of the proposed framework. Using 64 quantized colors in the RGB color space, the system can achieve a stable average recall value of 70% within the top 20 retrieved (and presented) images after only 5 iterations, with each such iteration taking about 2 seconds.
{"title":"Content-based sub-image retrieval using relevance feedback","authors":"Jie Luo, M. Nascimento","doi":"10.1145/1032604.1032607","DOIUrl":"https://doi.org/10.1145/1032604.1032607","url":null,"abstract":"This paper presents the use of relevance feedback to the problem of content-based sub-image retrieval (CBsIR). Relevance feedback is used to improve the accuracy of successive retrievals via a tile re-weighting scheme that assigns penalties to each tile of database images and updates the tile penalties for all relevant images retrieved at each iteration using both the relevant (positive) and irrelevant (negative) images identified by the user. Performance evaluation on a dataset of over 10,000 images shows the effectiveness and efficiency of the proposed framework. Using 64 quantized colors in the RGB color space, the system can achieve a stable average recall value of 70% within the top 20 retrieved (and presented) images after only 5 iterations, with each such iteration taking about 2 seconds.","PeriodicalId":415406,"journal":{"name":"ACM International Workshop on Multimedia Databases","volume":"62 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2004-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125439604","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}
Multivariate time series (MTS) datasets are common in various multimedia, medical and financial applications. We propose a similarity measure for MTS datasets, ErosExtended Frobenius norm), which is based on Principal Component Analysis (PCA). Eros applies PCA to MTS datasets represented as matrices to generate principal components and associated eigenvalues. These principal components and eigenvalues are then used to compare the similarity between MTS matrices. Though Eros in itself does not satisfy the triangle inequality, without which existing multidimensional indexing structures may not be utilized, the lower and upper bounds to satisfy the triangle inequality are obtained. In order to show the validity of Eros for similarity search on MTS datasets, we performed several experiments on three datasets (2 real-world and 1 synthetic). The results show the superiority of our approaches as compared to the traditional similarity measures for MTS datasets, such as Euclidean Distance (ED), Dynamic Time Warping (DTW), Weighted Sum SVD (WSSVD) and PCA similarity factor (SPCA) in precision/recall.
{"title":"A PCA-based similarity measure for multivariate time series","authors":"Kiyoung Yang, C. Shahabi","doi":"10.1145/1032604.1032616","DOIUrl":"https://doi.org/10.1145/1032604.1032616","url":null,"abstract":"Multivariate time series (MTS) datasets are common in various multimedia, medical and financial applications. We propose a similarity measure for MTS datasets, <i>Eros</i> <i>E</i>xtended F<i>ro</i>beniu<i>s</i> norm), which is based on Principal Component Analysis (PCA). <i>Eros</i> applies PCA to MTS datasets represented as matrices to generate principal components and associated eigenvalues. These principal components and eigenvalues are then used to compare the similarity between MTS matrices. Though <i>Eros</i> in itself does not satisfy the triangle inequality, without which existing multidimensional indexing structures may not be utilized, the lower and upper bounds to satisfy the triangle inequality are obtained. In order to show the validity of <i>Eros</i> for similarity search on MTS datasets, we performed several experiments on three datasets (2 real-world and 1 synthetic). The results show the superiority of our approaches as compared to the traditional similarity measures for MTS datasets, such as Euclidean Distance (ED), Dynamic Time Warping (DTW), Weighted Sum SVD (WSSVD) and PCA similarity factor (S<sc>PCA</sc>) in precision/recall.","PeriodicalId":415406,"journal":{"name":"ACM International Workshop on Multimedia Databases","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2004-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130798355","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}
MPEG-7 documents, which are primarily for multimedia information exchange, are also data-centric XML documents. Due to its advantages, the relational DBMS is the best choice for storing such XML documents. Storing XML data in relational DBMS can be classified into two classes of storage model: structure-mapping and model-mapping. However, the structure-mapping model cannot support complex Xpath-based query efficiently and model mapping approach lacks the flexible capability in representing all kinds of datatypes. In this paper, we present a new storage approach, called SM3. As an XML document, MPEG-7 document can be viewed as XML tree. Such a tree graph, where the internal nodes are element type with element contents, represents the structure of document and can be viewed as nodes which are meaningful only for document traversal. The leaf node, which is a single-valued attribute or element type with text content, has little usage for XML tree routing as it is the end-point of Xpath. So it can be viewed as the special node which only holds value. In this paper, SM3 was designed to use model-mapping approach to store all internal nodes and structure-mapping model to store all leaf nodes. SM3 integrate the advantages of those two models and avoid the main drawbacks from each method. Performance studies are conducted by comparing SM3 with XParent (a pure model-mapping method) and SM3 with XML-DBMS (a pure structure-mapping method). The experimental results are presented in the paper and initial results are encouraging.
{"title":"Looking at mapping, indexing & querying of MPEG-7 descriptors in RDBMS with SM3","authors":"Yang Chu, L. Chia, S. Bhowmick","doi":"10.1145/1032604.1032615","DOIUrl":"https://doi.org/10.1145/1032604.1032615","url":null,"abstract":"MPEG-7 documents, which are primarily for multimedia information exchange, are also data-centric XML documents. Due to its advantages, the relational DBMS is the best choice for storing such XML documents. Storing XML data in relational DBMS can be classified into two classes of storage model: structure-mapping and model-mapping. However, the structure-mapping model cannot support complex Xpath-based query efficiently and model mapping approach lacks the flexible capability in representing all kinds of datatypes. In this paper, we present a new storage approach, called SM3. As an XML document, MPEG-7 document can be viewed as XML tree. Such a tree graph, where the internal nodes are element type with element contents, represents the structure of document and can be viewed as nodes which are meaningful only for document traversal. The leaf node, which is a single-valued attribute or element type with text content, has little usage for XML tree routing as it is the end-point of Xpath. So it can be viewed as the special node which only holds value. In this paper, SM3 was designed to use model-mapping approach to store all internal nodes and structure-mapping model to store all leaf nodes. SM3 integrate the advantages of those two models and avoid the main drawbacks from each method. Performance studies are conducted by comparing SM3 with XParent (a pure model-mapping method) and SM3 with XML-DBMS (a pure structure-mapping method). The experimental results are presented in the paper and initial results are encouraging.","PeriodicalId":415406,"journal":{"name":"ACM International Workshop on Multimedia Databases","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2004-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127877114","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 development of technology generates huge amounts of non-textual information, such as images. An efficient image annotation and retrieval system is highly desired. Clustering algorithms make it possible to represent visual features of images with finite symbols. Based on this, many statistical models, which analyze correspondence between visual features and words and discover hidden semantics, have been published. These models improve the annotation and retrieval of large image databases. However, image data usually have a large number of dimensions. Traditional clustering algorithms assign equal weights to these dimensions, and become confounded in the process of dealing with these dimensions. In this paper, we propose a top-down, subspace clustering algorithm as a solution to this problem. For a given cluster, we determine relevant features based on histogram analysis and assign greater weight to relevant features as compared to less relevant features. We have implemented four different models to link visual tokens with keywords based on the clustering results of our clustering algorithm and K-means algorithm, and evaluated performance using precision, recall and correspondence accuracy using benchmark dataset. The results show that our algorithm is better than traditional ones for automatic image annotation and retrieval.
{"title":"Automatic image annotation and retrieval using subspace clustering algorithm","authors":"Lei Wang, Li Liu, L. Khan","doi":"10.1145/1032604.1032621","DOIUrl":"https://doi.org/10.1145/1032604.1032621","url":null,"abstract":"The development of technology generates huge amounts of non-textual information, such as images. An efficient image annotation and retrieval system is highly desired. Clustering algorithms make it possible to represent visual features of images with finite symbols. Based on this, many statistical models, which analyze correspondence between visual features and words and discover hidden semantics, have been published. These models improve the annotation and retrieval of large image databases. However, image data usually have a large number of dimensions. Traditional clustering algorithms assign equal weights to these dimensions, and become confounded in the process of dealing with these dimensions. In this paper, we propose a top-down, subspace clustering algorithm as a solution to this problem. For a given cluster, we determine relevant features based on histogram analysis and assign greater weight to relevant features as compared to less relevant features. We have implemented four different models to link visual tokens with keywords based on the clustering results of our clustering algorithm and K-means algorithm, and evaluated performance using precision, recall and correspondence accuracy using benchmark dataset. The results show that our algorithm is better than traditional ones for automatic image annotation and retrieval.","PeriodicalId":415406,"journal":{"name":"ACM International Workshop on Multimedia Databases","volume":"126 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2004-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115962781","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 importance of automatic discrimination between speech signals and music signals has evolved as a research topic over recent years. The need to classify audio into categories such as speech or music is an important aspect of many multimedia document retrieval systems. Several approaches have been previously used to discriminate between speech and music data. In this paper, we propose the use of the mean and variance of the discrete wavelet transform in addition to other features that have been used previously for audio classification. We have used Multi-Layer Perceptron (MLP) Neural Networks as a classifier. Our initial tests have shown encouraging results that indicate the viability of our approach.
{"title":"Automatic classification of speech and music using neural networks","authors":"M. K. S. Khan, W. Al-Khatib, M. Moinuddin","doi":"10.1145/1032604.1032620","DOIUrl":"https://doi.org/10.1145/1032604.1032620","url":null,"abstract":"The importance of automatic discrimination between speech signals and music signals has evolved as a research topic over recent years. The need to classify audio into categories such as speech or music is an important aspect of many multimedia document retrieval systems. Several approaches have been previously used to discriminate between speech and music data. In this paper, we propose the use of the mean and variance of the discrete wavelet transform in addition to other features that have been used previously for audio classification. We have used Multi-Layer Perceptron (MLP) Neural Networks as a classifier. Our initial tests have shown encouraging results that indicate the viability of our approach.","PeriodicalId":415406,"journal":{"name":"ACM International Workshop on Multimedia Databases","volume":"58 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2004-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128502205","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}
Samira Hammiche, S. Benbernou, Mohand-Said Hacid, A. Vakali
This paper deals with the problem of finding multimedia data that fulfill the requirements of user queries. We assume both the user query and the multimedia data are expressed by MPEG-7 standard. The MPEG-7 formalism lacks the semantics and reasoning support in many ways. For example, the search of the implicit data can not be achieved, due to its description based on XML schema. We propose a framework for querying multimedia data based on a tree embedding approximation algorithm, combining the MPEG-7 standard and an ontology.
{"title":"Semantic retrieval of multimedia data","authors":"Samira Hammiche, S. Benbernou, Mohand-Said Hacid, A. Vakali","doi":"10.1145/1032604.1032612","DOIUrl":"https://doi.org/10.1145/1032604.1032612","url":null,"abstract":"This paper deals with the problem of finding multimedia data that fulfill the requirements of user queries. We assume both the user query and the multimedia data are expressed by MPEG-7 standard. The MPEG-7 formalism lacks the semantics and reasoning support in many ways. For example, the search of the implicit data can not be achieved, due to its description based on XML schema. We propose a framework for querying multimedia data based on a tree embedding approximation algorithm, combining the MPEG-7 standard and an ontology.","PeriodicalId":415406,"journal":{"name":"ACM International Workshop on Multimedia Databases","volume":"64 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2004-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133664483","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}
Development of Web-based multimedia applications is expected to hold central importance for engineering and technological progress during the rest of this decade. It is already opening up new research frontiers in a number of areas such as multimedia data modeling and indexing, data mining, multimedia document management, semantic Web, pervasive computing, distributed sensor networks, computer security, real-time operating systems, human-computer interaction, and storage technology etc. As a result of concerted effort in these areas, many Web-accessible multimedia applications involving different media types, e.g., video, audio, text, images, animation and graphics, are rapidly emerging Examples of such applications abound in the domains of health care, education, entertainment, manufacturing, e-commerce, digital libraries as well as military and critical national infrastructures. The premise is that the integration of Web and multimedia technologies can provide cost effective solutions for management and dissemination of information, which is a primary tool for increasing economic efficiency. Development of Web-based multimedia applications needs a broad range of technological solutions that deal with organizing, storing, and delivering multimedia information in an integrated, secure and timely manner with guaranteed quality of service (QoS). Multimedia database management, when viewed in conjunction with integration of contents from independent Web-based data sources, present formidable research and development challenges. Key challenges include: • content analysis and indexing of distributed multimedia data and documents • semantic modeling and knowledge-based representation of multimedia data • transformation and organization of multimedia data semantics as a part of Semantic Web • security, privacy and QoS related issues concerning Web-based multimedia database applications • emerging Web standards and their role in managing distributed multimedia databases In this talk, we elaborate on these challenges and describe several solutions and tools that have been developed for Web-based multimedia database systems. Acknowledgement: This research was in part supported by the National Science Foundation under the awards IIS0209111 and EIC-9972883. Copyright is held by the author/owner(s). MMDB'04, November 13, 2004, Washington, DC, USA. ACM 1-58113-975-6/04/0011.
{"title":"Web-based multimedia databases: prospects and challenges","authors":"A. Ghafoor","doi":"10.1145/1032604.1032605","DOIUrl":"https://doi.org/10.1145/1032604.1032605","url":null,"abstract":"Development of Web-based multimedia applications is expected to hold central importance for engineering and technological progress during the rest of this decade. It is already opening up new research frontiers in a number of areas such as multimedia data modeling and indexing, data mining, multimedia document management, semantic Web, pervasive computing, distributed sensor networks, computer security, real-time operating systems, human-computer interaction, and storage technology etc. As a result of concerted effort in these areas, many Web-accessible multimedia applications involving different media types, e.g., video, audio, text, images, animation and graphics, are rapidly emerging Examples of such applications abound in the domains of health care, education, entertainment, manufacturing, e-commerce, digital libraries as well as military and critical national infrastructures. The premise is that the integration of Web and multimedia technologies can provide cost effective solutions for management and dissemination of information, which is a primary tool for increasing economic efficiency. Development of Web-based multimedia applications needs a broad range of technological solutions that deal with organizing, storing, and delivering multimedia information in an integrated, secure and timely manner with guaranteed quality of service (QoS). Multimedia database management, when viewed in conjunction with integration of contents from independent Web-based data sources, present formidable research and development challenges. Key challenges include: • content analysis and indexing of distributed multimedia data and documents • semantic modeling and knowledge-based representation of multimedia data • transformation and organization of multimedia data semantics as a part of Semantic Web • security, privacy and QoS related issues concerning Web-based multimedia database applications • emerging Web standards and their role in managing distributed multimedia databases In this talk, we elaborate on these challenges and describe several solutions and tools that have been developed for Web-based multimedia database systems. Acknowledgement: This research was in part supported by the National Science Foundation under the awards IIS0209111 and EIC-9972883. Copyright is held by the author/owner(s). MMDB'04, November 13, 2004, Washington, DC, USA. ACM 1-58113-975-6/04/0011.","PeriodicalId":415406,"journal":{"name":"ACM International Workshop on Multimedia Databases","volume":"60 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2004-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133221162","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}
This paper describes a fully automatic content-based approach for browsing and retrieval of MPEG-2 compressed video. The first step of the approach is the detection of shot boundaries based on motion vectors available from the compressed video stream. The next step involves the construction of a scene tree from the shots obtained earlier. The scene tree is shown to capture some semantic information as well as to provide a construct for hierarchical browsing of compressed videos. Finally, we build a new model for video similarity based on global as well as local motion associated with each node in the scene tree. To this end, we propose new approaches to camera motion and object motion estimation. The experimental results demonstrate that the integration of the above techniques results in an efficient framework for browsing and searching large video databases.
{"title":"A motion based scene tree for browsing and retrieval of compressed videos","authors":"Haoran Yi, D. Rajan, L. Chia","doi":"10.1145/1032604.1032608","DOIUrl":"https://doi.org/10.1145/1032604.1032608","url":null,"abstract":"This paper describes a fully automatic content-based approach for browsing and retrieval of MPEG-2 compressed video. The first step of the approach is the detection of shot boundaries based on motion vectors available from the compressed video stream. The next step involves the construction of a scene tree from the shots obtained earlier. The scene tree is shown to capture some semantic information as well as to provide a construct for hierarchical browsing of compressed videos. Finally, we build a new model for video similarity based on global as well as local motion associated with each node in the scene tree. To this end, we propose new approaches to camera motion and object motion estimation. The experimental results demonstrate that the integration of the above techniques results in an efficient framework for browsing and searching large video databases.","PeriodicalId":415406,"journal":{"name":"ACM International Workshop on Multimedia Databases","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2004-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117003938","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}
Video classification is an important step towards multimedia understanding. Most state-of-the-art approaches which apply HMM to capture the temporal information of videos have the limitation by assuming that the current state of a video depends only on the immediate previous state. Nevertheless, this assumption may not hold for videos of various categories. In this paper, we present an effective video classifier which employs the association rule mining technique to discover the actual dependence relationship between video states. The discriminatory state transition patterns mined from different video categories are then used to perform classification. Besides capturing the association between states in the time space, we also capture the association between low-level features in spatial dimension to further distinguish the semantics of videos. Experimental results show that the performance of our association rule based classifier is quite promising.
{"title":"VRules: an effective association-based classifier for videos","authors":"Ling Chen, S. Bhowmick, L. Chia","doi":"10.1145/1032604.1032619","DOIUrl":"https://doi.org/10.1145/1032604.1032619","url":null,"abstract":"Video classification is an important step towards multimedia understanding. Most state-of-the-art approaches which apply HMM to capture the temporal information of videos have the limitation by assuming that the current state of a video depends only on the immediate previous state. Nevertheless, this assumption may not hold for videos of various categories. In this paper, we present an effective video classifier which employs the association rule mining technique to discover the actual dependence relationship between video states. The discriminatory state transition patterns mined from different video categories are then used to perform classification. Besides capturing the association between states in the time space, we also capture the association between low-level features in spatial dimension to further distinguish the semantics of videos. Experimental results show that the performance of our association rule based classifier is quite promising.","PeriodicalId":415406,"journal":{"name":"ACM International Workshop on Multimedia Databases","volume":"56 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2004-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122307200","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：