WO2008028893A1

WO2008028893A1 - Method of measuring blemishes on skin

Info

Publication number: WO2008028893A1
Application number: PCT/EP2007/059198
Authority: WO
Inventors: Venkatesh Shankar Iyer; Jayalakshmi Subramanyashastry Panyam
Original assignee: Unilever Plc; Unilever N.V.; Hindustan Unilever Limited
Priority date: 2006-09-07
Filing date: 2007-09-03
Publication date: 2008-03-13
Also published as: MX2009002291A; BRPI0714902A2

Abstract

Assessment of the efficacy of compositions to correct imperfections of the skin is an important part of the activity in the development of skin products. One commonly used method for assessment of the blemishes is through use of professional clinical graders. In this method trained clinicians grade the improvement of the skin appearance over extended use of the product. This assessment is subjective and expensive. Thus a method of measuring blemishes on the skin, particularly dark skin, is provided comprising the steps of: (a) acquiring a colour digital image of the skin; (b) selecting either the blue or the green colour channel of the image based on the average colour of the skin; (c) subjecting the selected colour channel to a Gaussian filter operation to obtain an intensity value for each pixel; (d) binarizing the image into pixels of either a black or a white colour using a predetermined threshold intensity value; and (e) counting the number of black pixels.

Description

METHOD OF MEASURING BLEMISHES ON SKIN

The invention relates to a method of measuring blemishes on the skin and more particularly to a method of measuring and tracking blemishes on dark skinned people like the Asian population.

The human skin has many imperfections. The exposed parts of the skin e.g. face, neck and hands are more prone to imperfections since they are exposed to sunlight and atmospheric irritants like pollutants. Imperfections can be due to localised skin pigmentation e.g. sun tan, freckles or could be due to formation of acne which subsequently leads to localized darkening.

Assessment of the efficacy of compositions to correct imperfections of the skin is an important part of the activity in the development of skin products. One commonly used method for assessment of the blemishes is through use of professional clinical graders. In this method trained clinicians grade the improvement of the skin appearance over extended use of the product. This assessment is subjective and expensive. Hence there have been many attempts to use instrument based image analysis methods to objectively grade the skin appearance

EP 516 457 (Unilever, 1992) describes a device for measuring changes in pigmentation of skin which measures the reflection of light from the skin at two wavelengths, one wavelength being within the range of 800 to 1500 nm and the second wavelength being within the range of 600 to 750 nm. This prior art requires the use of more than one range of light wavelengths. It is desirable to develop simpler methods where an image of the skin using normal lighting conditions can be used to measure the imperfections on the skin.

US 6 571 003 (Procter & Gamble, 2003) describes a method for locating one or more visual skin defects of a portion of a person, comprising: acquiring a first digital image of the portion of the person; electronically analyzing the first digital image of the portion of the person to locate an area containing a skin defect; determining a first numerical severity associated with the area containing the skin defect; and generating a comparison between the first numerical severity and a predetermined value associated with a population of people. US 6 761 697 (L'Oreal, 2004) describes a method for predicting evolution of at least one condition of an external body portion of a subject, the method comprising: receiving first data, wherein the first data is representative of at least one condition of an external body portion of a subject in a first time frame; receiving second data, wherein the second data is representative of the at least one condition of the external body portion of the subject in a second time frame occurring after the first time frame; generating, based on the first data and the second data, a forecasted evolution of the at least one condition of the external body portion of the subject; and transmitting to the subject the forecasted evolution.

The above-mentioned prior art publications describe image analysis methods for tracking the overall beauty condition of the skin and comparing/predicting possible scenarios based on the tracking data. None of the prior art have specifically dealt with providing a solution to the problem of analyzing the image of a blemish on the skin and providing a score on the number of blemishes in a highly representative and reproducible manner. Further, there still exists a need for providing a solution to the problem of tracking a blemish to measure whether the severity of the blemish has increased or decreased. This is especially true for measuring and tracking the blemishes on dark skinned people e.g. Asian populations especially in the South Asian region who have a skin colour in the range of 6 to 9 on a scale of 10, ten being the darkest.

It is thus an object of the invention to provide for a method and a system to measure blemishes on skin using a digital image taken under normal lighting conditions.

It is another object of the invention to provide for a method and a system to track the progress of a blemish on the skin over time using a digital image taken under normal lighting conditions.

It is yet another object of the invention to provide a method and a system especially suitable to measuring and tracking blemishes on dark skinned people.

Summary of the invention

According to one aspect of the invention there is provided for a method for measuring skin blemishes comprising the steps of: (a) acquiring a colour digital image of the skin; (b) selecting either the blue or the green colour channel of the image based on the average colour of the skin;

(c) subjecting the selected colour channel to a Gaussian filter operation to obtain an intensity value for each pixel; (d) binarizing the image into pixels of either a black or a white colour using a predetermined threshold intensity value; and (e) counting the number of black pixels.

It is particularly preferred that the method comprises measuring blemishes on dark skinned people having an average skin color in the range of 6 to 9 on a scale of 10, 10 being the darkest.

According to another aspect of the invention there is provided a system for carrying out the method of the invention comprising means to acquire the digital image of the skin operatively connected to a means for carrying out the image processing.

Detailed Description of the Invention

The present invention mainly aims at a method for measurement and analysis of blemishes on the skin. The measurement could be on any part of the skin, mostly on the exposed parts like hands, neck and face. The method involves the use of a simple colour digital camera and an image analysis means like a personal computer. The present invention provides for an objective measurement of the blemish size, shape and growth. The measurement and analysis by the present invention has been well correlated with grading of trained clinical graders while minimizing the subjectivity involved and reducing the cost and uncertainty of availability of such trained personnel.

This and other objects of the invention, its advantages and method of use will now be described with the help of the following detailed description and figures. The figures depicted and described herein form some embodiments of the invention and have been incorporated for illustrative purposes only. These should not be construed so as to limit the invention in any manner, as it would be obvious to the reader that many such embodiments are possible without diverting from the scope of the invention. In light thereof, variations and modifications will be suggested to one skilled in the art, all of which are within the spirit and purview of this invention. - A -

Brief Description of the Figures

Figure 1 is a schematic of the process used for measuring a blemish on the skin.

Figure 2 depicts a digital image of an Asian face with an average skin colour of 7 on a scale of 10, having blemishes.

Figure 2a shows figure 2 subjected to the process of the invention i.e. where the green channel is selected for image processing.

Figure 2b shows figure 2 subjected to the process outside the present invention i.e. where the blue channel is selected for image processing of the dark skinned Asian skin type.

Figure 2c shows figure 2 subjected to the process outside the present invention i.e. where the red channel is selected for image processing of the dark skinned Asian skin type.

Detailed Description of the Figures

Figure 1 is a schematic of the process used for quantification of the number of blemishes on the skin. The method of the invention involves the following main steps.

First a digital image of the skin is taken. This is done using a colour digital camera to get a digital image which is noise free and of good resolution. The position of the camera is adjusted such that all the images are acquired at a fixed lighting condition.The image is usually digitized as a 2048x 1536 matrix, where each pixel in the matrix has a colour intensity value. The data is then preferably transferred to a means for image analysis e.g. a personal computer. The average colour of the image of the skin is then determined by calculating the intensity of the picture. This may be done using any standard algorithm in image analysis software.

An image is mainly composed of three color channels red, green and blue channels. It has been determined through extensive experiments on various skin types that quantification of the number of blemishes on the skin is best when the blue channel is chosen for analysis of skin which is light and the green channel is chosen when the skin is darker. Thus according to the present embodiment of the invention, the blue channel is chosen when the average skin colour is less than 4 on a scale of 1 to 10, ten being the darkest. The blue channel is therefore mainly suitable for Caucasian skin colours. The green channel is chosen when the average skin colour is greater than 5 on a scale of 1 to 10, ten being the darkest. The green channel is therefore mainly suitable for Asian skin colours especially South Asian skin colours

The digital image is in the discrete domain where each pixel has an intensity value associated with it. There could be large discontinuities in the intensity values especially in areas having blemishes. In order to facilitate the image processing computations while at the same time enhancing the colour contrast in making a decision on quantifying the number of blemishes, it has been found that optimum results are obtained when the selected colour channel is subjected to a Gaussian filter operation. The filter operation is explained below.

Suppose an image is digitized as an image in a 20^*20 matrix format. Gaussian blur is basically a tool to blur the fine details of the digital image selectively within the image. This is again a matrix based filter. So if one applies a radius of 3 i.e. a 3^*3 Gaussian filter on a 20^*20 image matrix, the fine image element in every third pixel in the row and the column of the 20^*20 image matrix would be blurred. Thus the larger the radius employed the larger the scale of fine detail that will be removed. Low radius values remove only very fine detail while high radius values remove larger levels of detail. When applied in two dimensions, this produces a surface whose contours are concentric circles with a Gaussian distribution from the center point. Pixels where this distribution is non-zero are used to build a convolution matrix, which is applied to the original image. Each pixel's value is set to a weighted average of that pixel's neighborhood. The original pixel's value receives the heaviest weight (having the highest Gaussian value), and neighboring pixels receive smaller weights as their distance to the original pixel increases. This Gaussian Filter operation results in a blur that preserves boundaries and edges better than other, blurring filter operations.

The Gaussian filter operation is unique in that, in addition to being completely circularly symmetric, it can be applied to a two-dimensional image as two independent one- dimensional calculations. That is, the effect of applying the two-dimensional matrix can also be achieved by applying a series of single-dimensional Gaussian matrices in the horizontal direction, then repeating the process in the vertical direction. It has been observed that optimum results are obtained when an original image is subjected to two selective Gaussian filter operations and the resultant image is the difference between the images obtained after subjecting the original image to the two selective Gaussian filter operations. The radii for the finer filter operation is preferably in the range of 20 to 40, more preferably 20 to 30 and optimally 25. The radii for the course filter operation is in the range of 20 to 50, more preferably 20 to 40 and optimally 30.

Thus according to the present embodiment of the invention applied to Asian skin types, two copies of the selected green channel of the original image are made and named Imagel and Image2. The Imagel is subjected to a fine Gaussian filter operation of r = radius 20 to make Image 1 a. Image 2 is then subjected to a gross Gaussain filter operation of radius R = 30 to make Image 2a. The image 1a is the subtracted from the Image 2a to make Image 3.

The Image 3 is then binarized to make Image 4. This operation involves selecting a threshold colour intensity value. All pixels having the colour intensity value above the selected threshold value are given the black colour while the pixels having colour intensity value below the selected threshold value are given the white colour. The Image 4 is then the final image depicting the blemishes in a quantitative manner. The amount of blemishes may be depicted as a number by counting the number of black pixels.

Figure 2 depicts a colour image of an Asian face. Figure 2a, 2b, and 2c depict the image of the face as in Fig. 2 using green, blue and red channels respectively. The image of Fig. 2 was also graded by trained clinicians for the extent of blemishes. The data from the clinical grading correlated best with the data from figure 2a.

The invention is also suitable as a method which compares the change in the number of skin blemishes over time. The method of the invention also encompasses a method of tracking the progress of a particular blemish of interest. The method comprises the steps of (i) selecting a blemish of interest in the digital image obtained; (ii) measuring the average grey value, maximum grey value and the minimum grey value of the blemish of interest; (iii) calculating the standard deviation of the blemish and (iv) comparing the standard deviation value over time. The grey level of a pixel is defined as the amplitude of the function f(x,y) at any pair of coordinates (x,y) where f(x,y) is a two dimensional function representing an Image. Comparison of the standard deviation values over time indicates the progress of the blemish. If the standard deviation decreases, it indicates that the blemish is decreasing while if the standard deviation increases, it indicates that the blemish is increasing in size.

The invention also provides for a system for carrying out the method of the invention comprising means to acquire the digital image of the skin operatively connected to a means for carrying out the image processing.

The invention thus provides for a method and a system to measure and track the progress of blemishes on skin over time using a digital image taken under normal lighting conditions, especially for dark skinned people.

Claims

1. A method for measuring skin blemishes comprising the steps of: (a) acquiring a colour digital image of the skin; (b) selecting either the blue or the green colour channel of the image based on the average colour of the skin;

(c) subjecting the selected colour channel to a Gaussian filter operation to obtain an intensity value for each pixel;

(d) binarizing the image into pixels of either a black or a white colour using a predetermined threshold intensity value; and

(e) counting the number of black pixels.

2. A method as claimed in claim 1 wherein the blue colour channel is selected if the average skin colour is below a predetermined skin colour and the green channel is selected if the average skin colour is above said predetermined skin colour.

3. A method as claimed in claim 2 wherein the blue colour is selected for Caucasian skins and a green channel is selected for Asian skin.

4. A method as claimed in any one of the preceding claims wherein the intensity value for each pixel is the difference in the intensity value obtained by subjecting the image to two Gaussian filters of predetermined radii.

5. A method for tracking the number of skin blemishes over time comprising the method of measuring skin blemishes according to any one of the preceding claims and the additional steps of:

(a) repeating the foregoing steps at a later moment in time; and

(b) comparing the change in the number of skin blemishes over time.

6. A method for tracking the progress of a blemish comprising the method of measuring skin blemishes according to any one of the preceding claims and the additional steps of:

(a) selecting a blemish of interest in the digital image obtained;

(b) measuring the average grey value, maximum grey value and the minimum grey value of the blemish of interest; (c) calculating the standard deviation of the blemish; and

(d) comparing the standard deviation value over time.

7. A system for carrying out the methods as claimed in any one of the preceding claims comprising means to acquire the digital image of the skin operatively connected to a means for carrying out the image processing.

8. A system as claimed in claim 7 wherein said means to acquire a digital image is a digital camera.

9. A system as claimed in claim 7 or 8 wherein the means for carrying out the image processing is a personal computer.