This paper presents a strategic approach to tackling trimap-free natural image matting. Specifically, to address the false detection issue of existing trimap-free matting algorithms when the foreground object is not uniquely defined, we design a novel tangled structure (TangleNet) to handle foreground detection and matting prediction simultaneously. TangleNet enables information exchange between foreground segmentation and alpha prediction, producing high-quality alpha mattes for the most salient foreground object based on RGB inputs alone. TangleNet boosts network performance with a frequency-guided attention mechanism utilizing wavelet data. Additionally, we pretrain for salient object detection to aid in the foreground segmentation. Experimental results demonstrate that TangleNet is on par with the state-of-the-art matting methods requiring additional inputs, and outperforms all previous trimap-free algorithms in terms of both qualitative and quantitative results.

Most industries and organizations use collections of tools, devices, and applications that are growing in complexity. New tools or applications may be acquired and old tools may become obsolete over time. They are often running on a variety of platforms, have different bandwidth and QoS requirements, and in most cases they cannot be accessed through a single point of entry. Moreover, some tools may require specific configurations done by technical experts. To address these issues, we propose an extensible, reliable, and simple software architecture that can hide the complexity of provisioning the network and running the tools. This paper introduces a service-oriented approach for creating an agile infrastructure to provision devices, applications, and their underlying networks. The Eucalyptus prototype is developed as an empirical application to test this approach. Eucalyptus is built on a set of generic fine-grained Web services to manage and configure available resources, where new resources can be custom-built or imported from a third party. They can be integrated into Eucalyptus using a set of Web service-enabled APIs. Our user community consists of architects and industrial designers. Eucalyptus can manage and configure the resources needed by geographically distributed groups of architects who need to collaborate in real time on the design of buildings, in a virtual participatory design studio (PDS). Eucalyptus provides a single point of entry for the architects to access a wide variety of tools: videoconference applications, visualization services, rendering services employing parallel computers, etc. Eucalyptus provides a set of upper layer services for users to provision devices and applications running on high-speed broadband networks, as well as the commercial IP networks.

While most service-oriented solutions are developed for system integration or data integration, we propose a Web services-based solution, Eucalyptus, for provisioning applications and networks on demand. Eucalyptus is built on the power of the user-controlled lightpath provisioning (UCLP) tool, which pioneers a user-centric and service-oriented approach for creating and managing a private end-to-end optical network. As a first user of UCLP, Eucalyptus aims to construct an participatory environment for geographically distributed teams of architects and industrial designers with the with the support of configurable broadband switched networks, as well as the traditional routed IP networks. The contribution of Eucalyptus is to provide a proof-of-concept example on how Web service and service-oriented architecture (SoA) can effectively provide on-demand provisioning for heterogenous resources in hybrid networks. These resources can be provisioned, launched, monitored, terminated, and reserved through Web services. Eucalyptus is network and platform neutral. It offers a single point of entry for users to access resources ranging from video conference applications, rendering clusters, to the underlying networks. Each resource is configured through a resource-specific Web service. Eucalyptus also includes a set of generic management Web services to coordinate sessions, to manage resources and users, and to compose workflows, such that the network and the application are properly configured for the users engaging in a participatory design session.

Monotonicity is a simple yet significant qualitative characteristic. We consider the problem of segmenting an array in up to K segments. We want segments to be as monotonic as possible and to alternate signs. We propose a quality metric for this problem, present an optimal linear time algorithm based on novel formalism, and compare experimentally its performance to a linear time top-down regression algorithm. We show that our algorithm is faster and more accurate. Applications include pattern recognition and qualitative modeling.

Abstract The authors present a study of the real‐life information needs of 59 McGill University undergraduates researching essay topics for either a history or psychology course, interviewed just after they had selected their essay topic. The interview's purpose was to transform the undergraduate's query from general topic terms, based on vague conceptions of their essay topic, to an information need‐based query. To chart the transformation, the authors investigate N. J. Belkin, R. N. Oddy, and H. M. Brooks' Anomalous States of Knowledge (ASK) hypothesis (1982a, 1982b), which links the user's ASK to a relevant document set via a common code based on structural facets. In the present study an interoperable structural code based on eight essay styles is created, then notions of structural facets compatible with a high‐impact essay structure are presented. The important findings of the study are: (a) the undergraduates' topic statements and terms derived from it do not constitute an effective information need statement because for most of the subjects in the study the topic terms conformed to a low‐impact essay style; (b) essay style is an effective interoperable structural code for charting the evolution of the undergraduate's knowledge state from ASK to partial resolution of the ASK in an information need statement.

Abstract Domain novice users in the beginning stages of researching a topic find themselves searching for information via information retrieval (IR) systems before they have identified their information need. Pre‐Internet access technologies adapted by current IR systems poorly serve these domain novice users, whose behavior might be characterized as rudderless and without a compass. In this article we describe a conceptual design for an information retrieval system that incorporates standard information need identification classification and subject cataloging schemes, called the INIIReye System, and a study that tests the efficacy of the innovative part of the INIIReye System, called the Associative Index. The Associative Index helps the user put together his or her associative thoughts—Vannevar Bush's idea of associative indexing for his Memex machine that he never actually described. For the first time, data from the study reported here quantitatively supports the theoretical notion that the information seeker's information need is identified through transformation of his/her knowledge structure (i.e., the seeker's cognitive map or perspective on the task for which information is being sought).

This demonstration reports on the current stage of development of a web-based environment to support video sharing and annotation. The initial requirements and their implementation are briefly presented. The application is aimed at supporting the professional development of teachers with a multimedia application over a broadband network. However, work with teachers has pointed us to new requirements, differentiating VSA from prrevious work. Continued work with teachers will lead to further evolution of VSA.

This article presents “Islands,” an interactive graphical interface for construction, modification, and management of queries during a search session on a bibliographic database. The Islands interface is compared to the Dialog Interface™ on a search of the INSPEC database.