Large scale image search

Please enable javascript in your browser.

Panahi, Soraya | 2014

493 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 45834 (02)
University: Sharif University of Technology
Department: Mathematical Sciences
Advisor(s): Moghaddasi, Reza
Abstract:
Searching for images of the same object or scene in a large number of images is a major problem in computer vision. It has many applications specially in the search engines.For the goal of efficient image search, we need descriptors that are not only discriminative, but also short and need small amount of memory.In this thesis we analyze the image search methods in two categories: The first methods are based on converting the existed descriptors such as gist into a compact binary code. The second methods are based on building short descriptors, especially by some modifications in the framework of bag of features descriptor.Finally we will introduce a novel descriptor "bag of codes" which combines the advantages of the two above categories. The idea is to convert each patch to a binary code by a hashing method such as spectral hashing. Then the codes are using in a framework similar to the bag of words.This descriptor reduce the quantization phase in the BOW method which is done by k-means. In addition the construction of the descriptor is fast, since there is no need to find cluster centers. The general idea is to substitute k-means with hashing
Keywords:
Image Retrieval ; Words Bag Model ; Images Search ; Spectral Hashing ; Large Scale Image Search

No TOC