AU671500B2 - Refinement of colour images using reference colours - Google Patents

Description

-1- Refinement of Colour Images Using Reference Colours The present invention relates to the correction of colour images for the subsequent display or printing of colour images. In particular the present invention relates to the correction of colours using reference colours.

Background Art A number of aspects of the background art will now be described with reference to the accompanying drawings in which: Fig. 1 illustrates a sample image, Fig. 2 illustrates a first prior art method of correcting colour images.

Fig. 3 illustrates colour correction graph of an ideal image not requiring colour correction.

c:o Fig. 4 illustrates colour correction graphs of a non ideal colour image that has had colour correction applied, illustrating a change in colour brightness.

Fig. 5 illustrates a colour correction graph illustrating a decrease in brightness.

r 'Fig. 6 illustrates colour correction graph illustrating a change in contrast.

15 00Colour ;mages are often stored on a computer system in a certain format, convenient for the display of images on a colour display device. One such format is an RGB additive colour format where, for each pixel to be displayed, integer values of the Red, Green and Blue colour channel components of that particular pixel are stored, with normally the integer values only taking on a certain range of values. For example, a common format used is to store 8-bits of colour information for each of the Red, Green and Blue colour channels, with each channel taking on possible values in the range of 0 to 255, thereby providing 256 separate colour levels, and over 16 million separate displayable colours :when considering all three colour channels.

Other popular data storage fomats, to which the present invention can be applied, by means of transformations, known to those skilled in the art, include Cyan, Magenta, r Yellow and Black (CYMK), Hue, Saturation, Value (HSV), Hue, Lightness, Saturation (HLS) models.

In another popular format for the storing colours (known as an RGBA format) the RGB information is stored in addition to an opacity value 0 being stored which relates to the degree of transparency (or Opacity) that particular image has at that particular pixel.

This method of storage is especially convenient when multiple images are to be 'composited' over one another and transparency effects are to be used in the final composite colour image. A full discussion of the compositing process using an opacity t channel is disclosed in 'Compositing Digital Images' by Thomas Porter and Tom Duff, published in Computer Graphics, Volume 18, Number 3, July 1984, at page 253.

(Page+: 11)(258387) I~ -2- It is often desirable to correct particular colour information in a picture by alteration of the particular colour values that the colour channels of an image take, a process known as colour correction. This colour correction process is often performed with images that have been scanned into a computer storage device using a scanning device designed to produce, for example, RGB formatted colour output. The reasons for performing colour correction are many and various and include: the correction of over- or under exposure problems, to neutralize any colour cast due to lighting, filters, film, etc. and to optimise the reproduction range by adjusting the overall contrast.

Adjustment of tonal clipping of areas near the extremes of the displayable gamut of colours and the correction of out of gamut errors.

:oo I Applying various enhancements to the image, by for example, altering certain colours so that the image has a 'warmer' or 'cooler' appearance.

0Colour information data stored containing opacity information can be used with the i 15 present invention by first recovering the relevant colour information and using the method of the present invention on the recovered colour information.

Referring now to Fig. 1 and Fig. 2, there is shown a first method of correcting colour images. In a software system designed for the colour correction of images, a colour correction panel 1 is provided containing a number of sliders 2, one for each colour channel, from which the user can select and alter the sliders and the effect of the alterations i are transferred to the image 4. These effects are often applied immediately to the image 4 L thereby providing an immediate feed back effect. Alternatively, a manual input means 3 is provided for inputting selected numerical values to which the sliders 2 are automatically set and the changes reflected in the image 4. The values which the sliders can take lie between S 25 a certain maximum and minimum range (generally 0 to 255) and are used to derive a contrast and brightness value for each colour channel as described below.

Referring now to Fig. 3 there is shown colour correction graph of an ideal image not requiring colour correction. In this image the pixel values at each location and for each channel stay the same for the input and output, for example, an input value of 100 will become an output value of 100, and so forth. The contrast value is derived to be a measure of the gradient of the colour correction graph, whereas the brightness value is defined to be a measure of the output value at an input intensity of zero. I Referring now to Fig. 4 there is shown colour correction graphs of a non ideal colouri image that has had colour correction applied, illustrating a change in colour brightness. In this example of a colour correction graph 5 the brightness has been increased to a base level S(Page+: 11)(258387) I, CI 0~*4 6, so that the minimum value of input now has a greater value of output and all other colour values are mapped accordingly. As it is assumed that the maximum colour value displayable using an 8-bit schema is 255, and that input values that would normally give values greater than this maximum limit 7 are mapped to 255.

Referring now to Fig. 5 there is shown a colour correction graph illustrating a decrease in brightness. Again, those values that would fall below a minimum limit 8, must be clippped to zero.

Referring now to Fig. 6 there is shown colour correction graph illustrating a change in contrast. In this graph, the contrast, or gradient of the graph, has been increased so that the difference between neighbouring values has also increased, thereby increasing the contrast of a certain portion of the image. Those values less than a minimum (or maximum) limit 8 are clipped to be at that maximum limit.

Hence for each channel of input colour, a colour correction graph can be described by two variables, being the gradient or contrast c, and the intercept value for zero input being the brightness b. The output colour produced relative to the input colour, apart from any bounds clipping that may take place, is related by the equation: output (input c) b (EQ 1) The values for b and c can be read directly for the slider channel for each colour and the relevant colour corrections applied for each colour channel. Although experienced publishers of colour images can effectively use such a colour slider device as they have a good idea of how such a device works, a novice user is at a disadvantage in the use of such a device and therefore utility of such a colour correction panel is limited.

Summary of the Invention In accordance with one apsect of the present invention there is disclosed a method for colour correcting a predetermined portion of a colour image, said colour image comprising a plurality of pixels and including a plurality of colour information channels for each said pixel, said method comprising the steps of: selecting in said portion a first input reference colour of a first pixel and a corresponding first desired output colour of said first pixel; selecting in said portion a second input reference colour of a second pixel and a corresponding second desired output colour of said second pixel; defining a colour transformation for each said colour information channel based on said input reference colours and said desired output colours; applying said colour transformation to all said pixels in said predetermined portion; and displaying said image.

Page+11 258387 CFPO204AI

~I

-4- Preferably, the deriving step includes prducing a colour correction graph from said selected points and producing a line intersec ing a first point being the intersection of said first input reference colour and said first desired output colour and a second point being the intersection of said second input reference colour and said second desired output colour and deriving a brightness value from the intersection of said line with zero input colour value and deriving a contrast value from the gradient of said line.

Brief Description of the Drawings A preferred embodiment of the present invention will now be described with reference to the remaining drawings in which: Fig. 7 illustrates a method of choosing two original colours and two final colours.and Fig. 8 illustrates the colour correction graph of the preferred embodiment.

000000 o Detailed Description o 90 Referring now to Fig. 7 there is shown a method of choosing two original colour's and 0.

two final coiours.The two original colours comprising a first input colour 16 and a second o 15 input colour 17, are taken from those portions of the image that the user wishes to change.

The selection of a colour of an image which is being displayed, and the subsequent independent display of that colour, is a process known to those skilled in the art of ~..:developing computer graphics packages. The two final colours can be taken from the image itself, but are preferably taken from a colour wheel display. An example, designed specifically to improve image contrast is where blacks in the image appear 'washed out' i and the whites appear murky. In order to improve the picture, the black from the image can form the first input colour 16 and a full black from a colour wheel or like colour selection ii device, known to those skilled in the art, would form the first output colour 18. Similarly, o o the murky white from the image can form the second input colour 17 and a full white from a colour wheel would form the second output colour 19. Upon pressing the apply button the colour correction desired to achieve the desired effect is applied to the image.

Referring now to Fig. 8 there is shown the colour correction graph of the preferred embodiment. This colour correction graph is derived in the following manner upon the Spressing of the apply button 20. The first input colour 16 and first output colour 18 form a first point 22 on a colour correction graph 21, and the second input colour 17 and the second output colour 19 form a seccnd point 23 on the colour correction graph 21. A line drawn through these two points is used, as shown to form the colour correction for the particular graph, with the clipping of colour values taking place if necessary. The colour correction values ofb and c can be derived for this line and used with equation to colour (Page+: 11)(258387) correct the image. This process can be repeated for each input colour channel to produce a colour corrected image, applying the initial desired transformation.

Additionally, the initial image can be stored with the contrast and brightness settings for each channel.

An apparatus implementing the preferred embodiment can comprise a general purpose computer system having a high resolution monitor capable of displaying a graphical user interface.

The computer system could then be programmed to display the image as depicted in Fig. 7. The user would then be instructed to choose the candidate colours 16 19, using an interactive device such as a mouse or the like. Upon choosing the apply button 20, the contrast and brightness values are derived for each primary as hereinbefore described and o subsequently applied to every pixel in the input image to thereby produce a colour corrected input image.

The foregoing describes only one embodiment of the present invention particular to a 15 the RGB model of image storage. The use of other models such as the CYMK model used in colour printing or the RGBA model used in compositing, and modifications, obvious to those skilled in the art, can be made thereto without parting from the scope of the invention.

S i3 (Page+: 1)(258387)

Claims (5)

1. A method for colour correcting a predetermined portion of a colour image, said colour image comprising a plurality of pixels and including a plurality of colour information channels for each said pixel, said method comprising the steps of: selecting in said portion a first input reference colour of a first pixel and a corresponding first desired output colour of said first pixel; selecting in said portion a second input reference colour of a second pixel and a corresponding second desired output colour of said second pixel; defining a colour transformation for each said colour information channel based on said input reference colours and said desired output colours; applying said colour transformation to all said pixels in said predetermined portion; and displaying said image.

2. A method of correcting a colour image as claimed in claim 1 wherein said a 15 information values include contrast and brightness values for each said colour information S° channel of said colour image.

3. A method of correcting a colour image as claimed in claim 1 or 2 wherein said .4 plurality of colour information channels include red, green and blue colour channels.

4. A method of correcting a colour image as claimed in claim 1, 2 or 3 wherein said S 20 colour transformation is a linear colour transformation. il

5. A method of correcting a colour image substantially as described herein with reference to Figs. 7 and 8 of the drawings. I 25 Dated this Twenty-first day of June 1996 Canon Information Systems Research Australia Pty Ltd Canon Kabushiki Kaisha Patent Attorneys for the Applicants S 0 30 Spruson Ferguson :i Sage+ll 258387 CFP0204AU ivr o 1"L I Abstract Refinement of Colour Images Using Reference Colours Methods of correcting colour images often involve a complex interrelation of many colour components and require extensive training to be able to be fully utilized. The present disclosure involves a simplified form of colour correction. From the selection of a series of actual (16,17) and desired (18,19) colours, a series of contrast and brightness values are derived, which can then be applied to the colour image to produce the colour corrected image. In*~~ o o n I o~ rr n J a o o D r r *r Fig. 8. B D PI C-P D 0 r It 1 O (Page+:11)(258387)