Special Offers
- IGI Global’s New Emerging Topic e-Book Collections
  Acquire highly focused and affordable Cutting-Edge Peer-Reviewed Research Content through a selection of 17 topic-focused e-Book Collections discounted up to 90%, compared to list prices. Collection topics include Artificial Intelligence, Data Science, Language Learning, Marketing and Customer Relations, Sustainability, and many more. Hosted on the InfoSci^® platform, these collections feature no DRM, no additional cost for multi-user licensing, no embargo of content, full-text PDF & HTML format, and more.
  Learn More
- Open Access Book (Free Access) - Encyclopedia of Information Science and Technology, Sixth Edition (ISBN: 9781668473665)
  The Encyclopedia of Information Science and Technology, Sixth Edition) continues the legacy set forth by the first five editions by providing comprehensive coverage and up-to-date definitions of the most important issues, concepts, and trends pertaining to technological advancements and information management within a variety of settings and industries. The entire book is being published under open access.
  Read Now
- Open Access Book (Free Access) - Food Sustainability, Environmental Awareness, and Adaptation and Mitigation Strategies for Developing Countries (ISBN: 9781668456293)
  Food Sustainability, Environmental Awareness, and Adaptation and Mitigation Strategies for Developing Countries provides information on the recent technology, mitigation, and environmental protection that must be applied for food sustainability in developing countries. This book is being published under Platinum Open Access through funding from Diponegoro University, Indonesia.
  Read Now
- Open Access Book (Free Access) - New Models of Higher Education: Unbundled, Rebundled, Customized, and DIY (ISBN: 9781668438091)
  The Walmart Corporation and the Lumina Foundation have provided funding to make New Models of Higher Education: Unbundled, Rebundled, Customized, and DIY fully open access, completely removing any paywall between scholars in education and the latest research on new models for the future of higher education.
  Read Now
- Open Access Book (Free Access) - Handbook of Research on the Global View of Open Access and Scholarly Communications (ISBN: 9781799898054)
  Through a collaboration between IGI Global and the University of North Texas, the Handbook of Research on the Global View of Open Access and Scholarly Communications has been published as fully open access, completely removing any paywall between researchers of any field, and the latest research on the equitable and inclusive nature of Open Access and all of its complications.
  Read Now
Books
- - Books by Subject
  - Business, Administration, & Management
  - Scientific, Technical, & Medical (STM)
  - Education
  - Books by Field
Journals
- - Journals
  - OnDemand Journal Articles
  - Journals by Subject
  - Business, Administration, & Management
  - Scientific, Technical, & Medical (STM)
  - Education
  - Journals by Field
e-Collections
OnDemand
Open Access
- View All Open Access Opportunities
  Search across all of IGI Global’s available open access publishing opportunities to unleash your research potential.
  Find an Open Access Journal for Your Next Manuscript
  Search across all of IGI Global’s available open access publishing opportunities to unleash your research potential.
  Submit an Open Access Book Proposal
  Learn more about open access book publishing and how it can propel your research forward in the field.
  Convert Your Work to Open Access
  Already published? You can convert your work to open access to increase its impact through IGI Global’s Restrospective Open Access Program.
  Utilize Open Access Collection Database
  Open up your research potential by utilizing our open access content or integrating the open access collection into your library
  Consider Open Access Agreements
  For Libraries: consider no-cost or investment-level open access agreements with IGI Global to support your faculty's research endeavors.
  Search Funding Resources
  Looking for additional funding resources to support your open accesss endeavors? View industry resources compiled by our open access team.
  Review Open Access Policies & Ethical Guidelines
  Considering IGI Global to publish your work under open access? Review IGI Global’s open access policies and ethical guidelines
Publish with Us
Resources
- - Instructors
  - Course Adoption
  - Teaching Cases
  - K-12 Online Learning Collection
  - Authors and Editors
  - eEditorial Discovery^® System
  - Peer Review Process
  - Ethics and Malpractice
  - COPE Membership
  - Fair Use Policy
  - Open Access Publishing
  - FAQ
Catalogs
About Us

Bregman Hyperplane Trees for Fast Approximate Nearest Neighbor Search

Bilegsaikhan Naidan, Magnus Lie Hetland

Source Title: International Journal of Multimedia Data Engineering and Management (IJMDEM) 3(4)

DOI: 10.4018/jmdem.2012100104

OnDemand:

(Individual Articles)

Available

$37.50

Current Special Offers

No Current Special Offers

Abstract

This article presents a new approximate index structure, the Bregman hyperplane tree, for indexing the Bregman divergence, aiming to decrease the number of distance computations required at query processing time, by sacrificing some accuracy in the result. The experimental results on various high-dimensional data sets demonstrate that the proposed index structure performs comparably to the state-of-the-art Bregman ball tree in terms of search performance and result quality. Moreover, this method results in a speedup of well over an order of magnitude for index construction. The authors also apply their space partitioning principle to the Bregman ball tree and obtain a new index structure for exact nearest neighbor search that is faster to build and a slightly slower at query processing than the original.

Article Preview

Top

2. Preliminaries

Let be our database of vectors. Let be a strictly convex differentiable function. The Bregman divergence (Bregman, 1967) between based on is defined as:

(1)

Here, denotes the gradient of the function at and denotes the dot product between and . In other words, a Bregman divergence measures the distance between and its first-order Taylor expansion. The definition is illustrated in Figure 1.