Similar image retrieval method based on local feature neighborhood information

Abstract

The invention relates to a similar image retrieval method based on local feature neighborhood information. The method includes the steps of (1) obtaining training pictures, (2) conducting feature detection and description on the pictures in a multi-scale space through a Hessian-Affine feature point detection algorithm and an SIFT local feature descriptor, (3) constructing corresponding shadow features according to features extracted in the step (2), (4) conducting clustering on the features extracted in the step (2) through a k mean value clustering algorithm and generating a visual dictionary including k visual words, (5) mapping all the features to the visual word with the smallest distance with L2 of the visual word one by one and storing the features in an inverted index structure, (6) storing the inverted index structure to form a query database and (7) obtaining the corresponding inverted index for querying the pictures, comparing the inverted index with the query database and obtaining a retrieval result list. Compared with the prior art, the method has the advantages of being high in picture retrieval accuracy.

Description

technical field [0001] The invention relates to a picture retrieval method, in particular to a similar picture retrieval method based on local feature neighborhood information. Background technique [0002] In recent years, computer vision has been developing rapidly. Among them, similar image retrieval is a basic but challenging task, so it has attracted much attention. [0003] Currently, the bag-of-words model based on local features and inverted index structure is one of the most commonly used image retrieval models. Image local features are a type of features used in the field of image processing. Finding extreme points in the scale space, extracting position, scale, and rotation invariants can detect key points in the image. The bag-of-words model is an approximate method for feature matching. In this model, a local feature is quantized to its nearest visual word in a pre-trained dictionary and stored in an inverted index for querying. [0004] However, the above i...

AI Technical Summary

Problems solved by technology

[0005] First, due to the locality of features and quantization errors, the feature matching accuracy is very low, and there are a large number of wrong matches, which affect the final retrieval accuracy

[0006] Second, a large number of algorithms focus on the information of the feature points themselves, ignoring the strong correlation between the neighborhood information of the feature points and the feature points

[0007] Third, some algorithms try to use the spatial relationship between feature points for spatial verification, but such methods generally consume additional computing resources and steps, and the computational efficiency is not high

Images