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

A method to measure local image similarity based on the l1 distance measure

A technology for measuring local similarity and measuring images, which is applied in the field of image processing and can solve problems such as increased computing overhead

Inactive Publication Date: 2011-04-27
SONY CORP +1
View PDF3 Cites 7 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Clearly, the computational overhead grows rapidly as the shard size increases

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
  • A method to measure local image similarity based on the l1 distance measure
  • A method to measure local image similarity based on the l1 distance measure
  • A method to measure local image similarity based on the l1 distance measure

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0021] The similarity measure used here is based on L 1 gap instead of the common L 2 gap. There are several reasons for this choice. Natural images have heavy tailed distributions, and the noise properties that degrade the image may be non-Gaussian. L 1 Gap is more applicable to such data, as eg P.Howarth and S.Ruger in "Fractional distance measures for content-based image retrieval" (Lecture notes in computer science ISSN 0302-9743, vol. 3408, 2005, pp. 447-456 page, which is hereby incorporated by reference), as described in L 1 Gap is not like L 2 Disparities or other fractional distances are affected by outliers. L 1 Gap gives all components the same weight. Second, with L 2 difference (even if the square root is not taken into account, which is still the sum of squares of the difference), the absolute difference (L 1 gap) is much simpler in terms of calculations.

[0022] Figure 7 Similarity metrics for patch sizes 1x1 and 3x3 are shown. When the fragment s...

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

A method of adaptive local image similarity measurement based on the L1 distance measure is described. A relationship between distance measures is used to estimate appropriate thresholds for various patch sizes. The choice of patch size depends on the degradations contained in the image and the application. The relation between the similarity measures is established using the distribution of L1 distances for various patch sizes. For larger degradations, similarity measure with a bigger patch size is employed. For lesser imperfections, a smaller patch size produces acceptable results. To keep the computational overhead manageable, the smallest patch size that gives the desired image quality is employed.

Description

technical field [0001] The invention relates to the field of image processing. More specifically, the present invention relates to local image similarity measurement. Background technique [0002] Local image similarity estimation is an important problem in image processing. Conceptually, image similarity is Can be classified into 3 categories, including: 1) Low level similarity. Patches are considered similar if some distance measure (e.g., p-norm, EarthMovers, Mahalanobis) is within a certain threshold; 2) Medium-level similarity. Here the fragments share some simple semantic property; and 3) a high level of similarity. In this case, similarity is mainly defined by semantics. The properties that make two fragments similar are not visible, but they can be inferred from visible information such as gesture. [0003] In most single-sensor color imaging systems, only one color is measured per pixel. The remaining components must be estimated to complete the color informa...

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): G06K9/64
CPCG06T7/0002
Inventor 法尔罕·A·巴卡伊西尾研一董晓刚松下伸行松井启贵取二郎
Owner SONY CORP