Unlock instant, AI-driven research and patent intelligence for your innovation.

Image retrieval method based on multi-block bifurcated dictionary tree index structure

A technology of image retrieval and tree indexing, which is used in still image data retrieval, digital data information retrieval, still image data indexing, etc., to achieve the best balance between memory overhead and access speed, and improve retrieval efficiency.

Pending Publication Date: 2022-05-20
青岛图灵科技有限公司
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] In order to overcome the shortcomings and deficiencies of the prior art, the present invention provides an image retrieval method based on a multi-block bifurcated trie index structure, by dividing the binary feature vector generated by each image into m mutually exclusive substrings, For each substring, an efficient block multi-fork dictionary tree index structure is designed to solve the nearest neighbor search problem in Hamming space

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Image retrieval method based on multi-block bifurcated dictionary tree index structure
  • Image retrieval method based on multi-block bifurcated dictionary tree index structure
  • Image retrieval method based on multi-block bifurcated dictionary tree index structure

Examples

Experimental program
Comparison scheme
Effect test

example 1

[0047] Example one, such as Figure 4 As shown, a container is established for each leaf node to store all strings with the same prefix associated with the node. In the specific implementation process, only the first 4 bits of the binary feature vector are used to establish a multi-block bifurcated dictionary tree index structure, every 2 consecutive bits are regarded as a block unit, and the storage container corresponding to each leaf node is a Hash table, thereby reducing the memory overhead caused by feature data storage.

[0048] This embodiment divides all binary feature vectors in the image feature base into m sets of mutually exclusive feature substrings according to a preset collection method under the condition of a given image feature base, query feature vector and retrieval radius. , and construct a corresponding multi-block bifurcated dictionary tree for each feature substring set. The query feature vector is also divided into m query feature substrings accordin...

example 2

[0054] Example 2, on the basis of the previous example 1, such as Figure 5 As shown, the query binary feature vector is 111101, the nearest neighbor retrieval radius is 2, and the target binary feature vector corresponding to the query binary feature vector is retrieved in the aforementioned multi-block bifurcated dictionary tree index structure, and the search traversal path is shown by the dotted line in the figure. , using the Hamming space nearest neighbor search method to retrieve in the constructed multi-block bifurcated dictionary tree index structure, and obtain the image retrieval result.

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention provides an image retrieval method based on a multi-block bifurcated dictionary tree index structure. The method comprises the following steps: S1, obtaining a pre-constructed image feature base library; wherein the image feature base library comprises a binary feature vector of each image in the image library; s2, dividing each binary feature vector in the feature base library to obtain feature substring sets, and constructing a multi-block forked dictionary tree index structure for each feature substring set according to preset dictionary tree parameters; and S3, extracting a query binary feature vector of the to-be-retrieved image, and performing neighbor retrieval in the multi-block bifurcated dictionary tree index structure according to the query binary feature vector to obtain an image retrieval result. According to the scheme, the efficient multi-block bifurcation dictionary tree index structure and the retrieval method are designed, the neighbor search problem in the Hamming space is solved, the storage overhead brought by massive image feature data is reduced, and the retrieval speed is increased.

Description

technical field [0001] The invention relates to the technical field of massive image retrieval, in particular to an image retrieval method based on a multi-block bifurcated dictionary tree index structure. Background technique [0002] A major challenge in large-scale target retrieval tasks for complex scenes is how to achieve fast and efficient target image retrieval under a massive image base. For massive image retrieval problems, a lot of research work is mainly devoted to extracting distinguishing features in images and expressing them. Most of the early implementations are based on scale-invariant and rotation-invariant feature descriptors (such as SIFT, SURF). The computational overhead and response time are of great significance, so a lot of research work is mainly carried out around the hash algorithm. The purpose of the hash algorithm is to map the feature vector from a high-dimensional space to a low-dimensional space, and to maintain the nearest neighbor structu...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06F16/51G06F16/583
CPCG06F16/51G06F16/583Y02D10/00
Inventor 冯栋刘治宇刘浩陈洪伟张永范超
Owner 青岛图灵科技有限公司