GPU acceleration-based large-scale image vocabulary tree retrieval method and system

Abstract

The invention discloses a GPU acceleration-based large-scale image vocabulary tree retrieval method and belongs to the field of computer vision. Based on structural features of a GPU platform, a novel vocabulary tree data structure is employed; and traditional SIFT feature points are modularized and organized, so GPU large-scale parallel computing capacity can be fully utilized and large-scale image SIFT feature point mapping efficiency can be greatly improved. After batches of SIFT feature points are mapped into a histogram, a quick and compact method is employed to enhance and improve an original reverse index scheme, so post vision vocabulary peculiarity express, histogram normalization and query image grades can be accelerated to a certain extent. The invention further provides a GPU acceleration-based large-scale image vocabulary tree retrieval system, playing an important acceleration role for image retrieval and having a great effect on a large-scale scene three-dimensional reconstruction algorithm efficiency increase.

Description

technical field [0001] The invention belongs to the field of computer vision, and more specifically relates to a large-scale image vocabulary tree retrieval method and system based on GPU acceleration. Background technique [0002] The vocabulary tree algorithm proposed by David Nistér et al uses a tree structure to quantify visual vocabulary, integrates the quantification and indexing of visual vocabulary, and combines text search technology to speed up the similarity judgment of image descriptors. It has high accuracy and efficiency, and is widely used in target recognition, scene recognition and 3D reconstruction. [0003] The original vocabulary tree algorithm mainly includes three stages: visual vocabulary learning, vocabulary tree construction and vocabulary tree query. [0004] In the visual vocabulary learning stage, hierarchical k-means clustering is employed to generate structured visual vocabulary. First, it is necessary to extract the Scale-Invariant Feature Tr...

AI Technical Summary

Problems solved by technology

[0007] Since the SIFT feature points of all database images need to be mapped to the corresponding visual vocabulary in the construction stage of the vocabulary tree, although the original vocabulary tree algorithm uses a hierarchical tree structure for comparison and query, the efficiency has been improved to a certain extent, but In the face of the current widespread large-scale image retrieval, this stage will still occupy a huge computational overhead of the entire algorithm

Images