Computer Communication & Collaboration
ISSN:2292-1028 (Print) ISSN:2292-1036(Online)
Vol. 2, Issue 3 (2014.8)
Table of Contents
Human Action Recognition in video has become a very important topic in many computer vision applications such as automated surveillance and human-computer interaction. Various approaches have been developed, but most of them focus on entire video sequence. In this study, we introduce a step to standard action recognition pipeline called "video highlighting". The main contribution of this paper is to propose a new approach, which highlights parts of input video data for recognition task. More informative data can be retained while less important data are eliminated when necessary. First, key-frames are extracted automatically and then, to preserve general dynamics of video sequences, continuous segments are constructed by including frames around the selected key-frames. Finally, a new input video is generated by concatenating these excerpts. This new sequence is used as input for the action recognition framework. The experiments were conducted for two different conventional action datasets: KTH, UCF sports; and the results are reported for various configurations. We give a performance comparison between our highlighted input videos and original video sequences. Furthermore, different low-level features that are used in the key-frame extraction process are tested with different numbers of extracted key-frames. The experiments indicate that the proposed method not only achieves higher performance, but also can maintain and to some extent increase recognition precision.
Human Action Recognition; Key-frame extraction; Unsupervised feature learning; Bag-of-Words (BoW); Spatio-temporal features; Video sequence
Continuance intention is the attitude of an individual who has performed an act, adopted a behavior or used a technology in the past to once again perform the act, adopt the behavior or use the technology. In studies of continuance intentions a key variable is expectation confirmation. Expectation confirmation is the degree to which an individual's attitudes or experiences pre adoption or use match those after adoption or use. Expectation confirmation has routinely been found to be meaningful in explaining continuance intentions. Yet, little is understood about the variables influencing expectation confirmation. The question addressed in this research examines the variables influencing expectation confirmation in the specific environment of students attitudes toward electronic textbooks. A theoretical model grounded in social cognitive theory is developed and estimated using a structural equations approach. The data were provided by a questionnaire distributed to students of a medium-size university in the western United States. All the included respondents had used an electronic textbook in the past. The results showed that electronic textbooks' perceived usefulness and self-efficacy directly influence expectation confirmation in meaningful ways. Furthermore, self-efficacy was found to indirectly influence expectation confirmation through perceived usefulness. The standardized paths from the antecedents of ease of use, arousal, and access speed significantly influence self-efficacy. Similarly, the standardized path coefficients for the antecedents of social norm and arousal had meaningful influences on perceived usefulness.
Expectation Confirmation; Expectation Confirmation Model; Continuance Intentions; Social Cognitive Theory
A well accepted view of competitive advantage is that physical, human, and organizational resources provide a firm with the basis for a sustainable competitive advantage if the resources are rare, durable, imperfectly imitable and non-tradable. While one role of knowledge management processes and systems is to create, store and share explicit knowledge, it may be the ability to use such processes and systems to successfully manage tacit knowledge that provides the greatest opportunity to develop a source of sustainable competitive advantage. However, the very characteristics that make tacit knowledge a potentially sustainable source of competitive advantage are characteristics that also make it hard to manage and use. One way to manage tacit knowledge is to develop and support social networks, in the form of such things as communities of practice as a knowledge management tool.
Communities of Practice; Competitive Advantage; Distributive Networks; Knowledge Management; Tacit Knowledge