Current acoustic speech recognition technology performs well with very small vocabularies m noise or with large vocabularies in very low noise. Accurate acoustic speech recognition in noise with vocabularies over 100 words has yet to be achieved. Humans frequently lipread the visible facial speech articulations to enhance speech recognition. especially when the acoustic signal is degraded by noise or hearing impairment. Automatic lipreading has been found to improve significantly acoustic speech recognition and could be advantageous in noisy environments such as offices, aircraft and factories.noise it fails except for very small vocabularies.S-5.6.7 Even humans have difficulty distinguishing between some consonants when the acoustic signal is degraded.This paper describes an improved automatic lipreading system which incorporates dynamic time warping'? and vector quantization.3 Results are reported from recognition experiments on the alphabet and the digits using four speakers. Recognition was restricted to isolated utterances and was speaker dependent.An improved version of a previously described automatic lipreading system has been developed which uses vector quantization. dynamic time warping, and a new heuristic distance measure. This paper presents visual speech recognition results from multiple speakers under optimal conditions. Results from combined acoustic and visual speech recognition are also presented which show significantly improved performance compared to the acoustic recognition system alone.The speakers faced a solid state camera which sent digitized video date to a microcomputer system with special purpose video processing hardware to perform windowing, grayscale thresholding and contour coding at video frame rate. The video data is sampled during an utterance and then reduced to a template consisting of a binary image sequence of the mouth. Registration of the mouth is achieved by tracking the speaker's nostrils from frame to frame assuming only small variations in distance between the nostrils and mouth.
Web personalization can achieve two business goals: increased advertising revenue and increased sales revenue. The realization of the two goals is related to two kinds of user behavior: item sampling and item selection. Prior research does not provide a model of attitude formation toward a personalization agent nor of how attitudes relate to these two behaviors. This limits our understanding of how web personalization can be managed to increase advertising revenues and/or sales revenues. To fill this gap, the current research develops and tests a theoretical model of user attitudes and behaviors toward a personalization agent. The model is based on an integration of two theories: the elaboration likelihood model (ELM) and consumer search theory (CST). In the integrated model, a user's attitude toward a personalization agent is influenced by both the number of items he/she has sampled so far (from CST) and the degree to which he/she cognitively processes each one (from ELM). In turn, attitude is modeled to influence both behaviors-that is, item selection and any further item sampling. We conducted a lab study and a field study to test six hypotheses. This research extends the theory on web personalization by providing a more complete picture of how sampling and processing of personalized recommendations influence a user's attitude and behavior toward the personalization agent. For online merchants, this research highlights the trade-off between item sampling and item selection and provides practical guidance on how to steer users toward the attitudes and behaviors that will realize their business goals.
Web personalization allows online merchants to customize Web content to serve the needs of individual customers. Using data mining and clickstream analysis techniques, merchants can now adapt website content in real time to capture the current preferences of online customers. Though the ability to offer adaptive content in real time opens up new business opportunities for online merchants, it also raises questions of timing. One question is when to present personalized content to consumers. Consumers prefer early presentation that eases their selection process, whereas adaptive systems can make better personalized content if they are allowed to collect more consumers' clicks over time. A review of personalization research confirms that little work has been done on these timing issues in the context of personalized services. The current study aims to fill that gap. Drawing on consumer search theory, we develop hypotheses about consumer responses to differences in presentation timing and recommendation type and the interaction between the two. The findings establish that quality improves over the course of an online session but the probability of considering and accepting a given recommendation diminishes over the course of the session. These effects are also shown to interact with consumer expertise, providing insights on the interplay between the different design elements of a personalization strategy.
How good is an IR test collection? A series of papers in recent years has addressed the question by empirically enumerating the consistency of performance comparisons using alternate subsets of the collection. In this paper we propose using Test Theory, which is based on analysis of variance and is specifically designed to assess test collections. Using the method, we not only can measure test reliability after the fact, but we can estimate the test collection's reliability before it is even built or used. We can also determine an optimal allocation of resources before the fact, e.g. whether to invest in more judges or queries. The method, which is in widespread use in the field of educational testing, complements data-driven approaches to assessing test collections. Whereas the data-driven method focuses on test results, test theory focuses on test designs. It offers unique practical results, as well as insights about the variety and implications of alternative test designs.
The browsing literature includes numerous typologies, i.e., types of browsing. However, there does not appear to be a clear consensus regarding the definition of browsing. This section reviews some of the literature on browsing, with the modest goal of understanding the various possible definitions. We will then be in a better position to adopt a definition that suits our current purpose. Definitions of BrowsingWe distinguish between definitions of browsing as a kind of behavior and definitions of browsing as a kind of task or information need. Bates (2002a) clearly defines browsing as a behavior that can be defined on its own mechanical terms: "[Browsing] involves successive acts of glimpsing, fixing on a target to examine visually or manually more closely, examining, then moving on to start the cycle over again." Chang and Rice (1993, p. 237) similarly identify strictly behavioral-even biological-definitions in the works of Morse and O'Connor. A study such as that by Qiu (1993), which measures whether users choose to employ browsing or analytical searching for a given task, clearly defines browsing as a behavior, independently of the task. In that study, browsing includes the behaviors of paging through nodes and looking through tables of contents, and analytical search relates to the behavior of submitting string searches.In contrast, numerous definitions of browsing refer to characteristics of the user's goals or task or his or her information need. We will refer to this as the cognitive definition of browsing. For an example of this approach, we refer to another article by Bates (Bates, 2002b), in which browsing is defined as whatever one does in the course of an undirected, active search. Active is behavioral; undirected is a characteristic of the need, not of the method or behavior: "Here we have no special information need or interest, 2 but actively 2 It should be noted that when these definitions refer to "undirected" or "unspecified," the intention is not (only) that these are not known-item searches, but that the searcher does not even know what specific information he or she needs.1 Browsers means "users engaged in browsing"; it does not refer to the World Wide Web navigation software. Relevance for Browsing, Relevance for Searching David BodoffDepartment of Information and Systems Management, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong. E-mail: dbodoff@ust.hk The concept of relevance has received a great deal of theoretical attention. Separately, the relationship between focused search and browsing has also received extensive theoretical attention. This article aims to integrate these two literatures with a model and an empirical study that relate relevance in focused searching to relevance in browsing. Some factors affect both kinds of relevance in the same direction; others affect them in different ways. In our empirical study, we find that the latter factors dominate, so that there is actually a negative correlation between the probability of a document...
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.