Loading...

A real-time grid-based method for estimating nearest neighbors in euclidean space

Zamani, Y ; Sharif University of technology | 2018

398 Viewed
  1. Type of Document: Article
  2. DOI: 10.1109/IranianMVIP.2017.8342344
  3. Publisher: IEEE Computer Society , 2018
  4. Abstract:
  5. The problem of finding nearest neighbors in a certain distance is well defined in machine learning area. There are well-known and exact solutions for it. However, in real world problems, especially in machine vision area, where we have a moving sensor and we want to know which objects of the scene are in the measurement range of it, two issues are important. First, usually in these problems, the time cost is more important than accuracy. It means they can tolerate some measurements error if they can do the process in real-time. Second, the location of an object can be described in the three-dimensional space and does not require the higher dimensions. According to these issues, we introduced a parametric method for finding nearest neighbors for a certain distance in a time efficient way. In addition, we defined error functions that help compromise between accuracy and time complexity. © 2017 IEEE
  6. Keywords:
  7. Binary classifier ; K-d tree ; Moving sensor ; Nearest neighbor ; Range tree ; Binary trees ; Computer vision ; Forestry ; Learning systems ; Binary classifiers ; Nearest neighbor search
  8. Source: 10th Iranian Conference on Machine Vision and Image Processing, MVIP 2017, 22 November 2017 through 23 November 2017 ; Volume 2017-November , April , 2018 , Pages 176-181 ; 21666776 (ISSN) ; 9781538644041 (ISBN)
  9. URL: https://ieeexplore.ieee.org/document/8342344