The Experts below are selected from a list of 306 Experts worldwide ranked by ideXlab platform
Hong-yuan Mark Liao - One of the best experts on this subject based on the ideXlab platform.
-
Unsupervised analysis of human behavior based on manifold learning
2009 IEEE International Symposium on Circuits and Systems, 2009Co-Authors: Yu-ming Liang, Sheng-wen Shih, Arthur Chun-chieh Shih, Hong-yuan Mark LiaoAbstract:In this paper, we propose a framework for unsupervised analysis of human behavior based on manifold learning. First, a pairwise human posture distance matrix is calculated from a training Action sequence. Then, the isometric feature mapping (Isomap) algorithm is applied to construct a low-dimensional structure from the distance matrix. The data points in the Isomap space are consequently represented as a time-series of low-dimensional points. A temporal segmentation technique is then applied to segment the time series into subseries corresponding to Atomic Actions. Next, a dynamic time warping (DTW) approach is applied for clustering Atomic Action sequences. Finally, we use the clustering results to learn and classify Atomic Actions using the nearest neighbor rule. Experiments conducted on real data demonstrate the efficacy of the proposed method.
-
ISCAS - Unsupervised analysis of human behavior based on manifold learning
2009 IEEE International Symposium on Circuits and Systems, 2009Co-Authors: Yu-ming Liang, Sheng-wen Shih, Arthur Chun-chieh Shih, Hong-yuan Mark LiaoAbstract:In this paper, we propose a framework for unsupervised analysis of human behavior based on manifold learning. First, a pairwise human posture distance matrix is calculated from a training Action sequence. Then, the isometric feature mapping (Isomap) algorithm is applied to construct a low-dimensional structure from the distance matrix. The data points in the Isomap space are consequently represented as a time-series of low-dimensional points. A temporal segmentation technique is then applied to segment the time series into subseries corresponding to Atomic Actions. Next, a dynamic time warping (DTW) approach is applied for clustering Atomic Action sequences. Finally, we use the clustering results to learn and classify Atomic Actions using the nearest neighbor rule. Experiments conducted on real data demonstrate the efficacy of the proposed method.
-
Learning Atomic Human Actions Using Variable-Length Markov Models
IEEE Transactions on Systems Man and Cybernetics Part B (Cybernetics), 2009Co-Authors: Yu-ming Liang, Sheng-wen Shih, Arthur Chun-chieh Shih, Hong-yuan Mark LiaoAbstract:Visual analysis of human behavior has generated considerable interest in the field of computer vision because of its wide spectrum of potential applications. Human behavior can be segmented into Atomic Actions, each of which indicates a basic and complete movement. Learning and recognizing Atomic human Actions are essential to human behavior analysis. In this paper, we propose a framework for handling this task using variable-length Markov models (VLMMs). The framework is comprised of the following two modules: a posture labeling module and a VLMM Atomic Action learning and recognition module. First, a posture template selection algorithm, based on a modified shape context matching technique, is developed. The selected posture templates form a codebook that is used to convert input posture sequences into discrete symbol sequences for subsequent processing. Then, the VLMM technique is applied to learn the training symbol sequences of Atomic Actions. Finally, the constructed VLMMs are transformed into hidden Markov models (HMMs) for recognizing input Atomic Actions. This approach combines the advantages of the excellent learning function of a VLMM and the fault-tolerant recognition ability of an HMM. Experiments on realistic data demonstrate the efficacy of the proposed system.
-
A Language Modeling Approach to Atomic Human Action Recognition
2007 IEEE 9th Workshop on Multimedia Signal Processing, 2007Co-Authors: Yu-ming Liang, Sheng-wen Shih, Arthur Chun-chieh Shih, Hong-yuan Mark LiaoAbstract:Visual analysis of human behavior has generated considerable interest in the field of computer vision because it has a wide spectrum of potential applications. Atomic human Action recognition is an important part of a human behavior analysis system. In this paper, we propose a language modeling framework for this task. The framework is comprised of two modules: a posture labeling module, and an Atomic Action learning and recognition module. A posture template selection algorithm is developed based on a modified shape context matching technique. The posture templates form a codebook that is used to convert input posture sequences into training symbol sequences or recognition symbol sequences. Finally, a variable-length Markov model technique is applied to learn and recognize the input symbol sequences of Atomic Actions. Experiments on real data demonstrate the efficacy of the proposed system.
Hai Zhou - One of the best experts on this subject based on the ideXlab platform.
-
ASP-DAC - Synthesis of resilient circuits from guarded Atomic Actions
The 20th Asia and South Pacific Design Automation Conference, 2016Co-Authors: Yuankai Chen, Hai ZhouAbstract:With aggressive scaling of minimum feature sizes, supply voltages, and design guard-bands, transient faults have become critical issues in modern electronic circuits. Synthesis from guarded Atomic Actions has been investigated by Arvind et al. to explore non-determinism for hardware concurrency. We show in this work that non-determinism in the guarded Atomic Actions can be further explored for synthesis of resilient circuits. When an error happens in an Atomic Action, the Action may not need to be recomputed if there exist other feasible Actions. Such flexibilities will be increased in the specification and explored in the synthesis for efficient error resiliency. Our synthesis approach expands the solution space and offers the possibility of performance optimization. Experimental results demonstrate the effectiveness and efficiency of our synthesis approach.
-
Synthesis of resilient circuits from guarded Atomic Actions
The 20th Asia and South Pacific Design Automation Conference, 2015Co-Authors: Yuankai Chen, Hai ZhouAbstract:With aggressive scaling of minimum feature sizes, supply voltages, and design guard-bands, transient faults have become critical issues in modern electronic circuits. Synthesis from guarded Atomic Actions has been investigated by Arvind et al. to explore non-determinism for hardware concurrency. We show in this work that non-determinism in the guarded Atomic Actions can be further explored for synthesis of resilient circuits. When an error happens in an Atomic Action, the Action may not need to be recomputed if there exist other feasible Actions. Such flexibilities will be increased in the specification and explored in the synthesis for efficient error resiliency. Our synthesis approach expands the solution space and offers the possibility of performance optimization. Experimental results demonstrate the effectiveness and efficiency of our synthesis approach.
Yu-ming Liang - One of the best experts on this subject based on the ideXlab platform.
-
Human Action segmentation and classification based on the Isomap algorithm
Multimedia Tools and Applications, 2013Co-Authors: Yu-ming Liang, Sheng-wen Shih, Arthur Chun-chieh ShihAbstract:Visual analysis of human behavior has attracted a great deal of attention in the field of computer vision because of the wide variety of potential applications. Human behavior can be segmented into Atomic Actions, each of which indicates a single, basic movement. To reduce human intervention in the analysis of human behavior, unsupervised learning may be more suitable than supervised learning. However, the complex nature of human behavior analysis makes unsupervised learning a challenging task. In this paper, we propose a framework for the unsupervised analysis of human behavior based on manifold learning. First, a pairwise human posture distance matrix is derived from a training Action sequence. Then, the isometric feature mapping (Isomap) algorithm is applied to construct a low-dimensional structure from the distance matrix. Consequently, the training Action sequence is mapped into a manifold trajectory in the Isomap space. To identify the break points between the trajectories of any two successive Atomic Actions, we represent the manifold trajectory in the Isomap space as a time series of low-dimensional points. A temporal segmentation technique is then applied to segment the time series into sub series, each of which corresponds to an Atomic Action. Next, the dynamic time warping (DTW) approach is used to cluster Atomic Action sequences. Finally, we use the clustering results to learn and classify Atomic Actions according to the nearest neighbor rule. If the distance between the input sequence and the nearest mean sequence is greater than a given threshold, it is regarded as an unknown Atomic Action. Experiments conducted on real data demonstrate the effectiveness of the proposed method.
-
Unsupervised analysis of human behavior based on manifold learning
2009 IEEE International Symposium on Circuits and Systems, 2009Co-Authors: Yu-ming Liang, Sheng-wen Shih, Arthur Chun-chieh Shih, Hong-yuan Mark LiaoAbstract:In this paper, we propose a framework for unsupervised analysis of human behavior based on manifold learning. First, a pairwise human posture distance matrix is calculated from a training Action sequence. Then, the isometric feature mapping (Isomap) algorithm is applied to construct a low-dimensional structure from the distance matrix. The data points in the Isomap space are consequently represented as a time-series of low-dimensional points. A temporal segmentation technique is then applied to segment the time series into subseries corresponding to Atomic Actions. Next, a dynamic time warping (DTW) approach is applied for clustering Atomic Action sequences. Finally, we use the clustering results to learn and classify Atomic Actions using the nearest neighbor rule. Experiments conducted on real data demonstrate the efficacy of the proposed method.
-
ISCAS - Unsupervised analysis of human behavior based on manifold learning
2009 IEEE International Symposium on Circuits and Systems, 2009Co-Authors: Yu-ming Liang, Sheng-wen Shih, Arthur Chun-chieh Shih, Hong-yuan Mark LiaoAbstract:In this paper, we propose a framework for unsupervised analysis of human behavior based on manifold learning. First, a pairwise human posture distance matrix is calculated from a training Action sequence. Then, the isometric feature mapping (Isomap) algorithm is applied to construct a low-dimensional structure from the distance matrix. The data points in the Isomap space are consequently represented as a time-series of low-dimensional points. A temporal segmentation technique is then applied to segment the time series into subseries corresponding to Atomic Actions. Next, a dynamic time warping (DTW) approach is applied for clustering Atomic Action sequences. Finally, we use the clustering results to learn and classify Atomic Actions using the nearest neighbor rule. Experiments conducted on real data demonstrate the efficacy of the proposed method.
-
Learning Atomic Human Actions Using Variable-Length Markov Models
IEEE Transactions on Systems Man and Cybernetics Part B (Cybernetics), 2009Co-Authors: Yu-ming Liang, Sheng-wen Shih, Arthur Chun-chieh Shih, Hong-yuan Mark LiaoAbstract:Visual analysis of human behavior has generated considerable interest in the field of computer vision because of its wide spectrum of potential applications. Human behavior can be segmented into Atomic Actions, each of which indicates a basic and complete movement. Learning and recognizing Atomic human Actions are essential to human behavior analysis. In this paper, we propose a framework for handling this task using variable-length Markov models (VLMMs). The framework is comprised of the following two modules: a posture labeling module and a VLMM Atomic Action learning and recognition module. First, a posture template selection algorithm, based on a modified shape context matching technique, is developed. The selected posture templates form a codebook that is used to convert input posture sequences into discrete symbol sequences for subsequent processing. Then, the VLMM technique is applied to learn the training symbol sequences of Atomic Actions. Finally, the constructed VLMMs are transformed into hidden Markov models (HMMs) for recognizing input Atomic Actions. This approach combines the advantages of the excellent learning function of a VLMM and the fault-tolerant recognition ability of an HMM. Experiments on realistic data demonstrate the efficacy of the proposed system.
-
A Language Modeling Approach to Atomic Human Action Recognition
2007 IEEE 9th Workshop on Multimedia Signal Processing, 2007Co-Authors: Yu-ming Liang, Sheng-wen Shih, Arthur Chun-chieh Shih, Hong-yuan Mark LiaoAbstract:Visual analysis of human behavior has generated considerable interest in the field of computer vision because it has a wide spectrum of potential applications. Atomic human Action recognition is an important part of a human behavior analysis system. In this paper, we propose a language modeling framework for this task. The framework is comprised of two modules: a posture labeling module, and an Atomic Action learning and recognition module. A posture template selection algorithm is developed based on a modified shape context matching technique. The posture templates form a codebook that is used to convert input posture sequences into training symbol sequences or recognition symbol sequences. Finally, a variable-length Markov model technique is applied to learn and recognize the input symbol sequences of Atomic Actions. Experiments on real data demonstrate the efficacy of the proposed system.
A Romanovsky - One of the best experts on this subject based on the ideXlab platform.
-
CAA-DRIP: a framework for implementing Coordinated Atomic Actions
2006 17th International Symposium on Software Reliability Engineering, 2006Co-Authors: Alfredo Capozucca, A Romanovsky, N. Guelfi, Patrizio Pelliccione, Avelino F ZorzoAbstract:This paper presents an implementation framework, called CAA-DRIP, that has been defined to allow a straightforward implementation of dependable distributed applications designed using the coordinated Atomic Action (CAA) paradigm. CAAs provide a coherent set of concepts adapted to the design of fault tolerant distributed systems that includes: structured transActions, distribution, cooperation, competition, and forward and backward error recovery mechanisms triggered by exceptions. DRIP (dependable remote interacting processes) is an efficient Java implementation framework, which provides support for implementing "dependable multiparty interActions (DMI)" which includes a general exception handling mechanism. As DMI has a softer exception handling semantics with respect to CAA semantics, a CAA design can be implemented by DRIP. The aim of the CAA-DRIP framework is to provide a set of Java classes that allows programmers to implement only the semantics of CAAs with the same terminology and concepts at the design and implementation levels. The new framework simplifies the implementation phase and at the same time reduces the size of the final system since it requires fewer number of instances for creating a CAA at runtime. Details of these improvements as well as a precise description of the CAAs behaviour in terms of state charts, which is used as a reference model to define the CAA-DRIP framework, are presented in this paper
-
Structuring integrated Web applications for fault tolerance
The Sixth International Symposium on Autonomous Decentralized Systems 2003. ISADS 2003., 2003Co-Authors: A Romanovsky, P. Periorellis, A.f. ZorzoAbstract:This paper shows how modern structuring techniques can be employed in integrating complex web applications such as travel agency systems. The main challenges the developers of such systems face are dealing with legacy web services and incorporating means for tolerating errors. Because of the very nature of such systems, exception handling is the main recovery technique to be applied in their development. We employ coordinated Atomic Actions to allow disciplined handling of such abnormal situations by recursively structuring the integrated system and by associating handlers with such Actions. We use protective wrappers in such a way that each operation on legacy components is transformed into an Atomic Action with a well-defined interface. To accommodate a combined use of several ready-made environments (such as communication packages, services and run-time supports), we employ a multilevel exception handling. We believe that these techniques are generally applicable for both: structuring integrated web applications and providing their fault tolerance.
-
Coordinated forward error recovery for composite Web services
22nd International Symposium on Reliable Distributed Systems 2003. Proceedings., 2003Co-Authors: V. Issarny, A Romanovsky, F. Tartanoglu, N. LevyAbstract:This paper proposes a solution based on forward error recovery, oriented towards providing dependability of composite Web services. While exploiting their possible support for fault tolerance (e.g., transActional support at the level of each service), the proposed solution has no impact on the autonomy of the individual Web services, our solution lies in system structuring in terms of co-operative Atomic Actions that have a well-defined behavior, both in the absence and in the presence of service failures. More specifically, we define the notion of Web Service Composition Action (WSCA), based on the Coordinated Atomic Action concept, which allows structuring composite Web services in terms of dependable Actions. Fault tolerance can then be obtained as an emergent property of the aggregation of several potentially non-dependable services. We further introduce a framework enabling the development of composite Web services based on WSCAs, consisting of an XML-based language for the specification of WSCAs.
-
a distributed coordinated Atomic Action scheme
Computer Systems: Science & Engineering, 2001Co-Authors: A Romanovsky, Avelino F ZorzoAbstract:Coordinated Atomic Actions have proved to be a very general concept which can be successfully applied for structuring complex concurrent systems consisting of elements which both cooperate and compete. The canonical Coordinated Atomic Action is built of several cooperating participants (roles) and a set of local objects which represent the Action state and provide the feature for cooperation. In addition, Coordinated Atomic Actions can compete for external objects which have conventional transActional properties. This paper offers a general approach to designing distributed Coordinated Atomic Action schemes and discusses the problems of Action components partitioning and distribution. The approach proposed relies on using forward error recovery in the form of distributed and concurrent exception handling and resolution. After discussing the general approach, we demonstrate how it can be applied when the standard distributed model of Ada 95 is used. The presentation of the scheme is sufficiently detailed for it to be used in practice. In particular, a thorough description of the Action support and all patterns (skeletons) required for designing application software are given.
-
Looking ahead in Atomic Actions with exception handling
Proceedings 20th IEEE Symposium on Reliable Distributed Systems, 2001Co-Authors: A RomanovskyAbstract:An approach to introducing exception handling into object-oriented N is presented. A novel Atomic Action scheme is developed that does not impose any participant synchronisation on Action exit. In order to use cooperative exception handling at the Action level as the main fault tolerance mechanism, we develop a distributed protocol that finds, for any exception raised, an Action containing all potentially erroneous information, aborts all of its nested Actions, resolves multiple concurrent exceptions and involves all the Action participants into cooperative handling of the resolved exception. In the scheme, no service messages are sent and no service synchronisation is introduced if there are no exceptions raised. This flexible scheme can be applied in a number of emerging areas in which entities of a different nature (including software tasks, people, plants, documents, organisations, etc.) participate in cooperative activities.
Arthur Chun-chieh Shih - One of the best experts on this subject based on the ideXlab platform.
-
Human Action segmentation and classification based on the Isomap algorithm
Multimedia Tools and Applications, 2013Co-Authors: Yu-ming Liang, Sheng-wen Shih, Arthur Chun-chieh ShihAbstract:Visual analysis of human behavior has attracted a great deal of attention in the field of computer vision because of the wide variety of potential applications. Human behavior can be segmented into Atomic Actions, each of which indicates a single, basic movement. To reduce human intervention in the analysis of human behavior, unsupervised learning may be more suitable than supervised learning. However, the complex nature of human behavior analysis makes unsupervised learning a challenging task. In this paper, we propose a framework for the unsupervised analysis of human behavior based on manifold learning. First, a pairwise human posture distance matrix is derived from a training Action sequence. Then, the isometric feature mapping (Isomap) algorithm is applied to construct a low-dimensional structure from the distance matrix. Consequently, the training Action sequence is mapped into a manifold trajectory in the Isomap space. To identify the break points between the trajectories of any two successive Atomic Actions, we represent the manifold trajectory in the Isomap space as a time series of low-dimensional points. A temporal segmentation technique is then applied to segment the time series into sub series, each of which corresponds to an Atomic Action. Next, the dynamic time warping (DTW) approach is used to cluster Atomic Action sequences. Finally, we use the clustering results to learn and classify Atomic Actions according to the nearest neighbor rule. If the distance between the input sequence and the nearest mean sequence is greater than a given threshold, it is regarded as an unknown Atomic Action. Experiments conducted on real data demonstrate the effectiveness of the proposed method.
-
Unsupervised analysis of human behavior based on manifold learning
2009 IEEE International Symposium on Circuits and Systems, 2009Co-Authors: Yu-ming Liang, Sheng-wen Shih, Arthur Chun-chieh Shih, Hong-yuan Mark LiaoAbstract:In this paper, we propose a framework for unsupervised analysis of human behavior based on manifold learning. First, a pairwise human posture distance matrix is calculated from a training Action sequence. Then, the isometric feature mapping (Isomap) algorithm is applied to construct a low-dimensional structure from the distance matrix. The data points in the Isomap space are consequently represented as a time-series of low-dimensional points. A temporal segmentation technique is then applied to segment the time series into subseries corresponding to Atomic Actions. Next, a dynamic time warping (DTW) approach is applied for clustering Atomic Action sequences. Finally, we use the clustering results to learn and classify Atomic Actions using the nearest neighbor rule. Experiments conducted on real data demonstrate the efficacy of the proposed method.
-
ISCAS - Unsupervised analysis of human behavior based on manifold learning
2009 IEEE International Symposium on Circuits and Systems, 2009Co-Authors: Yu-ming Liang, Sheng-wen Shih, Arthur Chun-chieh Shih, Hong-yuan Mark LiaoAbstract:In this paper, we propose a framework for unsupervised analysis of human behavior based on manifold learning. First, a pairwise human posture distance matrix is calculated from a training Action sequence. Then, the isometric feature mapping (Isomap) algorithm is applied to construct a low-dimensional structure from the distance matrix. The data points in the Isomap space are consequently represented as a time-series of low-dimensional points. A temporal segmentation technique is then applied to segment the time series into subseries corresponding to Atomic Actions. Next, a dynamic time warping (DTW) approach is applied for clustering Atomic Action sequences. Finally, we use the clustering results to learn and classify Atomic Actions using the nearest neighbor rule. Experiments conducted on real data demonstrate the efficacy of the proposed method.
-
Learning Atomic Human Actions Using Variable-Length Markov Models
IEEE Transactions on Systems Man and Cybernetics Part B (Cybernetics), 2009Co-Authors: Yu-ming Liang, Sheng-wen Shih, Arthur Chun-chieh Shih, Hong-yuan Mark LiaoAbstract:Visual analysis of human behavior has generated considerable interest in the field of computer vision because of its wide spectrum of potential applications. Human behavior can be segmented into Atomic Actions, each of which indicates a basic and complete movement. Learning and recognizing Atomic human Actions are essential to human behavior analysis. In this paper, we propose a framework for handling this task using variable-length Markov models (VLMMs). The framework is comprised of the following two modules: a posture labeling module and a VLMM Atomic Action learning and recognition module. First, a posture template selection algorithm, based on a modified shape context matching technique, is developed. The selected posture templates form a codebook that is used to convert input posture sequences into discrete symbol sequences for subsequent processing. Then, the VLMM technique is applied to learn the training symbol sequences of Atomic Actions. Finally, the constructed VLMMs are transformed into hidden Markov models (HMMs) for recognizing input Atomic Actions. This approach combines the advantages of the excellent learning function of a VLMM and the fault-tolerant recognition ability of an HMM. Experiments on realistic data demonstrate the efficacy of the proposed system.
-
A Language Modeling Approach to Atomic Human Action Recognition
2007 IEEE 9th Workshop on Multimedia Signal Processing, 2007Co-Authors: Yu-ming Liang, Sheng-wen Shih, Arthur Chun-chieh Shih, Hong-yuan Mark LiaoAbstract:Visual analysis of human behavior has generated considerable interest in the field of computer vision because it has a wide spectrum of potential applications. Atomic human Action recognition is an important part of a human behavior analysis system. In this paper, we propose a language modeling framework for this task. The framework is comprised of two modules: a posture labeling module, and an Atomic Action learning and recognition module. A posture template selection algorithm is developed based on a modified shape context matching technique. The posture templates form a codebook that is used to convert input posture sequences into training symbol sequences or recognition symbol sequences. Finally, a variable-length Markov model technique is applied to learn and recognize the input symbol sequences of Atomic Actions. Experiments on real data demonstrate the efficacy of the proposed system.
