CN101937558B - Label adding method based on image content - Google Patents

Label adding method based on image content Download PDF

Info

Publication number
CN101937558B
CN101937558B CN2010102512928A CN201010251292A CN101937558B CN 101937558 B CN101937558 B CN 101937558B CN 2010102512928 A CN2010102512928 A CN 2010102512928A CN 201010251292 A CN201010251292 A CN 201010251292A CN 101937558 B CN101937558 B CN 101937558B
Authority
CN
China
Prior art keywords
label
image
color space
cie
pixel
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Fee Related
Application number
CN2010102512928A
Other languages
Chinese (zh)
Other versions
CN101937558A (en
Inventor
冯结青
姜晓希
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhejiang University ZJU
Original Assignee
Zhejiang University ZJU
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Zhejiang University ZJU filed Critical Zhejiang University ZJU
Priority to CN2010102512928A priority Critical patent/CN101937558B/en
Publication of CN101937558A publication Critical patent/CN101937558A/en
Application granted granted Critical
Publication of CN101937558B publication Critical patent/CN101937558B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Image Processing (AREA)
  • Color Image Communication Systems (AREA)

Abstract

The invention discloses a label adding method based on image content, which comprises the steps of: transforming an image to be processed from RGB (Red, Green and Blue) color space to CIE-Lab color space; obtaining the visual importance of each pixel of the image transformed to the CIE-Lab color space; obtaining the mask bitmap information of a label to be added by the rasterization processing of the label; seeking a region with least visual importance loss in the image; adding the label to be added to the sought region with the least visual importance loss; and finishing the adding operation of the label of the image. The invention overcomes the picture sheltering resulted from an image key region of the traditional label adding method and can effectively reduce the visual loss of the image after label adding, thereby enhancing the visual experience of the image. The invention can realize parallel processing in graphic card hardware of a home computer and provides a simple, convenient, intuitive and high-efficiency label adding method of images for ordinary non-professional users.

Description

A kind of label addition method based on picture material
Technical field
The present invention relates to the computer image processing technology field, relate in particular to a kind of label addition method based on picture material.
Background technology
During people tend to unconsciously to image when observing and understanding image become interested in some specific region.The visual quality of entire image depends on the quality of these area-of-interests to a great extent, and the degrading of district of loseing interest in is subtle sometimes.These interesting areas are the zone, image border often.The image border is meant that generally pixel grey scale in the image has the set of the pixel that step changes; The image border is very useful to image recognition and analysis, and it can delineate out the target object profile, and the observer is come into plain view; Comprised abundant information (like direction, step character, shape etc.); Being the important attribute of image, is characteristic the most basic in the image, is the basis of analyzing, understand image.
The change of image edge information means the change of image substance or structure.Simultaneously, the shielding effect of human eye vision makes human eye very sensitive to edge distortion.Therefore, edge of image plays an important role in the understanding of human eye to image information, is the key factor of evaluation map as visual importance.In order better to detect beholder's possibility interesting areas in the image, get access to user's notice than higher content, the researchist has proposed the method that many image vision importance detect.Itti etc. have proposed a kind of algorithm that obtains image importance based on vision, referring to ITTI, and L.; KOCH, C., ANDNIEBUR; E.1998.A model of saliency based visual attention for rapid sceneanalysis.IEEE Transactions on Pattern Analysis and Machine Intelligence.Volume 20, and Issue 11, P1254-1259; But this algorithm speed is slower, and processing procedure is consuming time.
Existing user tag adding technique just simply is filled in certain FX in the image with label.This method is simple, and implementation efficiency is high, but does not consider the rich and varied property of image frame content.But these contents that are blocked user often are quite interested, the structural information of particularly being made up of the image border.And human eye very easily captures the covering that these destroy picture structure property characteristics, causes greatly having influenced and views and admires experience.And do not have special so far to the algorithm of confirming the user tag point of addition according to picture material.
Summary of the invention
The invention provides a kind of label addition method based on picture material, this method can reduce and cause the image vision loss after label adds based on the visual signature importance of image pixel.
A kind of label addition method based on picture material comprises:
(1) with pending image by the RGB color space conversion to the CIE-Lab color space;
(2) obtain the energy value that transforms to each pixel of image behind the CIE-Lab color space;
Because the CIE-Lab color space has separated colour brightness and color change, therefore the energy value of computed image pixel is more accurate than in the RGB color space, calculating in the CIE-Lab color space.
(3) the mask information table of acquisition label to be added;
If the label that adds is a written form, the user can be provided with as the font of the literal of label, style, size, color etc., is provided with according to these, through the rasterization process to the label literal, can obtain the mask information table of this label literal;
If the label that adds is an image format, then the mask information table is identical with the resolution sizes of this image.
(4) search the minimum zone of visual importance loss in the image;
Find out a matrix area identical with the mask information table size of label, make that the energy value sum of this regional interior pixel is minimum, this zone is as the Adding Area of label.
(5) label to be added is added in the zone that finds out, the label of accomplishing image adds operation;
(6) with add image after the label by the CIE-Lab color space conversion to the RGB color space.
The present invention overcomes the shortcoming that the conventional labels adding method causes image critical area picture to be blocked, and can effectively reduce and cause the image vision loss after label adds, thus the visual experience that improves image.The present invention can parallel processing in the graphics card hardware of household PC, for common unprofessional user provides a kind of easy intuitive and efficient image tag adding method.
Description of drawings
Fig. 1 is the technical scheme process flow diagram of the inventive method;
Fig. 2 is the algorithm synoptic diagram that the present invention obtains visual importance.
Embodiment
Below in conjunction with accompanying drawing the inventive method is elaborated.
As shown in Figure 1, a kind of label addition method based on picture material comprises:
(1) with pending image by the RGB color space conversion to the CIE-Lab color space;
The RGB color space conversion is changed according to following formula to the CIE-Lab color space:
L = 0.299 × R + 0.587 × G + 0.114 × B a = 0.713 × ( R - L ) b = 0.564 × ( B - L ) - - - ( 1 )
In the formula (1), R, G, B be the red, green, blue color value of presentation video in the RGB color space respectively; The luminance channel value of L presentation video in the CIE-Lab color space, a and b presentation video two color channel values in the CIE-Lab color space.Image file in the computing machine is generally preserved with the form of RGB color space; Image is handled to the CIE-Lab color space from the RGB color space conversion; Be because the CIE-Lab color space has separated colour brightness and color change; More can describe and reflect the perception of people to color, the visual importance value of calculating is more accurate.And the mutual conversion of two kinds of color spaces is reversible.
(2) obtain the energy value that transforms to each pixel of image behind the CIE-Lab color space;
The energy value of each pixel of computed image is than more accurate in the RGB color space in the CIE-Lab color space.As shown in Figure 2, use P I, jAny pixel in the presentation video, L I, jThe luminance channel value of representing this pixel, a I, j, b I, jTwo color channel values representing this pixel, this pixel energy value S I, jComputing formula be:
S i，j＝|L i，j-L i+1，j|+|a i，j-a i+1，j|+|b i，j-b i+1，j|
(2)
+|L i，j-L i，j+1|+|a i，j-a i，j+1|+|b i，j-b i，j+1|
Wherein, L I+1, j, a I+1, j, b I+1, jRemarked pixel P I, jNeighbor P on the vertical direction I+1, jLuminance channel value and two color channel values, L I, j+1, a I, j+1, b I, j+1Remarked pixel P I, jNeighbor P on the horizontal direction I, j+1Luminance channel value and two color channel values.Can find out that by formula (2) energy value of a pixel is adjacent the absolute value sum of the difference of pixel respective color channel value for this pixel.Therefore, the energy value of pixel is big more, and it is strong to show that this pixel is adjacent each Color Channel graded Shaoxing opera of pixel, and then its edge feature is just strong more, and visual importance is corresponding also just big more.
(3) the mask information table of acquisition label to be added;
Label to be added can be the symbol of passage, the little image of a width of cloth or other various certain information of reception and registration.Label for written form; To its rasterization process; Obtain the mask information table of the M * N of width of cloth two dimension according to the font attribute setting (comprising font style, font size etc.) of literal, it is by 0 and 1 matrix formed, and wherein 0 representes white space; 1 expression color fill area, 0 and 1 assembled arrangement correspondence goes out the image conversion of literal and representes.For the label of image format, its mask information table is the M * N matrix of image tag corresponding resolution, and all elements of matrix are 1.
(4) search the minimum zone of visual importance loss in the image;
In the image of label to be added, find out the identical matrix area of mask information size with label; Promptly find out the matrix area of a M * N; Make that the energy value sum of interior all pixels of this matrix area is minimum, the zone that add as label in the zone that finds out.Because the energy value of pixel has reacted the visual importance of pixel, the label Adding Area that finds out according to the inventive method can guarantee that image is carried out label to be added after the operation, destroys minimum to the sense of vision of original image.
(5) label to be added is added in the zone that finds out, the label of accomplishing image adds operation;
For image tag; Addition manner according to user's appointment; Directly replace the color value of respective pixel on the original image with the pixel color value of image tag; Also can the pixel color value of image tag and the color value of original image respective pixel be pressed preset mixed, as the new color value of this pixel.
For the interpolation process of word tag mask information table based on label.Element 1 pairing original image color of pixel value is directly replaced perhaps by preset mixed, as the new color value of this pixel according to the color value of word tag appointment in the mask information table of label; In the mask information table 0 element, constant in the pixel color value that image is corresponding.
(6) with add image after the label by the CIE-Lab color space conversion to the RGB color space;
Image is changed according to following formula to the RGB color space by the CIE-Lab color space conversion:
R = L + 1.403 × a G = L - 0.714 × a - 0.334 × b B = L + 1.773 × a - - - ( 3 )

Claims (2)

1. the label addition method based on picture material is characterized in that, comprises:
(1) with pending image by the RGB color space conversion to the CIE-Lab color space;
(2) energy value of image each pixel of computational transformation behind the CIE-Lab color space;
(3) obtain the mask information table of label to be added;
If the label that adds is a written form, according to the setting of label literal, the label literal is carried out rasterization process, obtain described mask information table;
If the label that adds is an image format, then the mask information table is identical with the resolution sizes of label image;
(4) in image, find out the minimum zone of visual importance loss, as the Adding Area of label;
The described minimum zone of visual importance loss that in image, finds out makes the energy value sum of all pixels in this matrix area minimum in image, to search a matrix area identical with the mask information table size of label;
(5) label to be added is added in the zone that finds out;
(6) with add image after the label by the CIE-Lab color space conversion to the RGB color space.
2. the label addition method based on picture material according to claim 1 is characterized in that, the computing formula of the energy value of described image pixel is:
S i，j＝|L i，j-L i+1，j|+|a i，j-a i+1，j|+|b i，j-b i+1，j|
+|L i，j-L i，j+1|+|a i，j-a i，j+1|+|b i，j-b i，j+1|
Wherein, S I, jExpression (i, the j) energy value of individual pixel, L I, jPresentation video is (i, j) the luminance channel value of individual pixel, a in the CIE-Lab color space I, j, b I, jDifference presentation video (i, j) two of individual pixel color channel values in the CIE-Lab color space.
CN2010102512928A 2010-08-10 2010-08-10 Label adding method based on image content Expired - Fee Related CN101937558B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN2010102512928A CN101937558B (en) 2010-08-10 2010-08-10 Label adding method based on image content

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN2010102512928A CN101937558B (en) 2010-08-10 2010-08-10 Label adding method based on image content

Publications (2)

Publication Number Publication Date
CN101937558A CN101937558A (en) 2011-01-05
CN101937558B true CN101937558B (en) 2012-04-25

Family

ID=43390872

Family Applications (1)

Application Number Title Priority Date Filing Date
CN2010102512928A Expired - Fee Related CN101937558B (en) 2010-08-10 2010-08-10 Label adding method based on image content

Country Status (1)

Country Link
CN (1) CN101937558B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP6035842B2 (en) 2012-04-25 2016-11-30 ソニー株式会社 Imaging apparatus, imaging processing method, image processing apparatus, and imaging processing system
GB2542133B (en) * 2015-09-08 2020-05-27 Imagination Tech Ltd Graphics processing method and system for processing sub-primitives
CN113988243B (en) * 2021-10-19 2023-10-27 艾斯芸防伪科技(福建)有限公司 Three-dimensional code generation and verification method, system, equipment and medium with verification code

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN100401302C (en) * 2006-09-14 2008-07-09 浙江大学 Image meaning automatic marking method based on marking significance sequence
US7747107B2 (en) * 2007-03-06 2010-06-29 Mitsubishi Electric Research Laboratories, Inc. Method for retargeting images
US8094973B2 (en) * 2008-10-06 2012-01-10 Xerox Corporation Content aware image resizing for text and contone images
CN101489139B (en) * 2009-01-21 2010-11-10 北京大学 Video advertisement correlation method and system based on visual saliency

Also Published As

Publication number Publication date
CN101937558A (en) 2011-01-05

Similar Documents

Publication Publication Date Title
US11601630B2 (en) Video processing method, electronic device, and non-transitory computer-readable medium
US8090196B2 (en) Album creating apparatus, album creating method and computer readable medium storing thereon program therefor
US10372226B2 (en) Visual language for human computer interfaces
Jacobson et al. Design goals and solutions for display of hyperspectral images
RU2680765C1 (en) Automated determination and cutting of non-singular contour of a picture on an image
US9036860B2 (en) Method for editing propagation of video and image content based on local feature structure preservation
Du et al. Saliency-guided color-to-gray conversion using region-based optimization
CN101536078A (en) Improving image masks
JP2012521708A (en) Method and apparatus for correcting an image using a saliency map based on color frequency
Le Moan et al. Saliency for spectral image analysis
CN106791937A (en) The mask method and system of video image
TW201531919A (en) Adding, deleting digital notes from a group of digital notes
Jin et al. Aesthetic photo composition by optimal crop-and-warp
WO2012077766A2 (en) Display format using display device for machine-readable dot patterns
CN101937558B (en) Label adding method based on image content
CN105118076A (en) Image colorization method based on over-segmentation and local and global consistency
CN106997608A (en) A kind of method and device for generating halation result figure
CN108038458B (en) Method for automatically acquiring outdoor scene text in video based on characteristic abstract diagram
CN103106671B (en) Method for detecting interested region of image based on visual attention mechanism
CN109978904A (en) Emergent aquactic plant growth information extracting method based on image technique
CN102184403B (en) Optimization-based intrinsic image extraction method
CN114529909A (en) Sample data set generation method and device and electronic equipment
CN105512663A (en) Significance detection method based on global and local contrast
Cui et al. Research on region selection strategy for visible watermark embedding
CN104240197A (en) Achromatic algorithm capable of maintaining contrast ratio, color consistency and gray pixel characteristics

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20120425

Termination date: 20150810

EXPY Termination of patent right or utility model