1. TECHNICAL FIELD OF INVENTION

The present invention relates generally to a methodology for measuring - lesion area of skin pigmentation disorders such as vitiligo and melasma using Principal Component Analysis (PCA) (201), Independent Component Analysis (ICA) (203) and Region Growing (205) as an assessment tool to evaluate the size of lesions.

2. BACKGROUND OF THE INVENTION Vitiligo is the common cutaneous pigmentary disorder characterized by depigmented macules and patches that result from loss of epidermal melanocytes, which is the pigment producing cell. Vitiligo also known as leucoderma or achromia is appeared as white spots on the skin due to lack of melanin and in many cases is thought to occur when the subject's immune cells attack and kill melanocytes in the subject's skin. Vitiligo can occur on any part of the integument with the first manifestations are loss of pigment and it often starts as a single spot or a few spots which gradually grow in size and number to several body parts such as hands, feet, arms, face, lips and eyes. The disease is usually progressive and the spots are often present in a symmetrical pattern on both sides of the body. Even though spontaneous repigmentation may occur, it is usually minor.

The prevalence of vitiligo varies from 0.1% to 2% in various global populations. It affects persons of all ethnic origins and both sexes. FIG. 1 shows the examples of vitiligo lesion on Asian patient in which the vitiligo lesions are paler than normal skin due to lack of melanin.

There are various therapeutic options available for vitiligo, ranging from a wide variety of topical drugs to phototherapy. The aim of treatment is to arrest disease progression and to repigment the vitiligo skin lesions. Currently, there is no objective method to assess the therapeutic response to vitiligo treatment. In a standard clinical practice, the degree of repigmentation is assessed subjectively by the physician by comparing the extent of vitiligo lesions before and after treatment using a scoring system based on an ordinal scale, the physician's global assessment (PGA). This assessment reflects the overall visual impression of the treatment response. The PGA uses nominal binary scales and reports the proportion of treated patients who achieve a specified degree of repigmentation as opposed to absolute values of repigmentation achieved by the individual patient. Therefore, it is a highly subjective method of assessment as the scales can vary from one study to another and are subjected to inter and intra-observer variations. Moreover, it may take months in order to able to observe repigmentation within the lesions. Treatment efficacy thus cannot be determined in a short term period. Typically, the degree of repigmentation that defines success is set somewhat arbitrarily at more than 50-75%.

While Melasma is an acquired, symmetric, irregular hypermelanosis on sun- exposed areas of the face. Many factors have been implicated in the pathogenesis of melasma with the most important ones remain UV radiations, hereditary predisposition and hormonal dysfunction. Melasma has historically been difficult to treat and therapy remains a challenge for this chronic condition. The prevalence of melasma is around 4% of world population. There are various therapeutic options available for melasma. The aim of treatment is to stop disease progression and to depigment the lesion areas. In a standard clinical practice, the assessment of treatment efficacy is assessed subjectively by the physician using the PGA. The PGA assessment is based on the overall visual impression of the treatment response. This assessment is subjective and subjected to inter and intra- observer variations. It is therefore advantageous if the above shortcoming is alleviated by having a methodology for measuring lesion area of skin pigmentation disorders such as vitiligo and melasma using digital imaging technique. Said digital imaging technique employing the digital analysis technique of Principal Component Analysis, Independent Component Analysis and Region

Growing to extract skin lesion features which may help physician to perform an objective follow up study of skin lesion progression and test the efficacy of therapeutic procedures.

3. SUMMARY OF THE INVENTION Accordingly, it is the primary aim of the present invention to provide a methodology for measuring lesion area of skin pigmentation disorders using digital imaging to evaluate treatment efficacy objectively and more effectively.

It is yet another objective of the present invention to provide a methodology for measuring lesion area of skin pigmentation disorders using digital imaging to enable physician to assess effectively the efficacy of treatment by comparing the area of lesion before and after treatment.

It is yet another objective of the present invention to provide a methodology for measuring lesion area of skin pigmentation disorders using digital imaging wherein the assessment of lesions can be done in a short term period and hence the progression of repigmentation/ depigmentation can be monitored effectively.

It is yet another objective of the present invention to provide a methodology for measuring lesion area of skin pigmentation disorders using digital imaging that has the potential to minimize variations due to inter-rater and intra-rater during assessment, thus capable of providing greater accuracy, sensitivity and specificity.

Other and further objects of the invention will become apparent with an understanding of the following detailed description of the invention or upon employment of the invention in practice.

According to a preferred embodiment of the present invention there is provided,

A methodology for measuring lesion area measurement of skin pigmentation disorders using digital imaging comprising:- i: capturing digital colour image of the subject area to acquire skin colour image; ii. executing feature analysis towards said acquired skin colour image; characterised in that said step of executing feature analysis towards said acquired skin colour image is done by the sub-steps of: a. converting the sample skin colour data of said acquired skin colour image by using Principal Component Analysis (PCA) to reduce the dimension from three RGB colour channels to two principal components; b. performing alignment to said two principal component axes by using Independent Component Analysis (ICA) to represent pure density vectors of melanin and haemoglobin; c. segmenting lesion areas in the pigment melanin image using Region Growing method.

In a second embodiment of the present invention there is provided,

An apparatus for measuring lesion area of skin pigmentation disorders using digital imaging comprising: i. at least one image capturing device for capturing digital colour image of the subject area; ii. at least one electronic processing means for processing and analysing said captured digital colour image; characterised in that further comprises at least a reference means with constant and known size attached to the subject area as reference or indicator to define the size of said digital colour image; the use of lighting setup with standardized conditions

4. BRIEF DESCRIPTION OF THE DRAWINGS

Other aspects of the present invention and their advantages will be discerned after studying the Detailed Description in conjunction with the accompanying drawings in which:

FIG. 1 shows examples of vitiligo lesion on Asian patient in which the vitiligo lesions are paler than normal skin due to lack of melanin.

FIG. 2 shows a flow chart outlining the general steps for measuring lesion area of skin pigmentation disorders using digital imaging. FIG. 3 shows the Independent Component Analysis (ICA) process.

FIG. 4 A to 4C shows examples of the ICA output.

FIG. 5A to 5D shows the Region Growing segmentation process.

FIG. 6 A to 6C shows examples of the original image and the manual segmentation output.

FIG. 7 shows the average performance of the present invention for different body parts.

5. DETAILED DESCRIPTION OF THE DRAWINGS

In the following detailed description, numerous specific details are set forth in order to provide a thorough understanding of the invention. However, it will be understood by those or ordinary skill in the art that the invention may be practised without these specific details. In other instances, well known methods, procedures and/ or components have not been described in detail so as not to obscure the invention. The invention will be more clearly understood from the following description of the methods thereof, given by way of example only with reference to the accompanying drawings. In the descriptions that follow, like numerals represent like elements in all figures. For example, where the numeral (2) is used to refer to a particular element in one figure, the numeral (2) appearing in any other figure refers to the same element.

Referring now to FIG. 2, there is shown a flow chart outlining the general steps for measuring lesion area of skin pigmentation disorders using digital imaging. Said skin pigmentation disorders includes vitiligo, melasma and other pigmentation disorder issue. The digital colour image of the subject area is first being captured by at least one image capturing device to acquire skin colour image. Said skin colour image is generated by using RGB colour model, which is combining three spectral bands, i.e. red, green and blue. Researchers have reported that skin colour is due to the combination of skin choromophores namely, melanins and haemoglobins.

Said image capturing device used for capturing said digital colour image of the subject area includes a digital single-lens reflex camera (DSLR camera) or any other digital camera which is able to provide accurate view for the captured image. Said image capturing device is connected to at least one electronic processing means which may be a software for processing and analysing said captured digital colour image. When said digital colour image is being captured, the use of lighting setup is standardized and preferably at least 1000 lumens of illumination is used. Typically the subject area is attached with at least a reference means (101) with constant and known size as reference or indicator to define the size of said digital colour image. Said reference means (101) includes a marker, a sticker or a label to be attached to the subject area and preferably it is green in colour

According to Tsumura, the spatial distribution of melanin and haemoglobin in digital images of skin can be separated by employing Independent Component Analysis (ICA) (203) of a skin colour image. Said analysis is based on three assumptions namely; linearity in the optical density domain of RGB channels, spatial variations of skin colour are caused by two skin choromophores (melanin and haemoglobin) and their quantities are mutually independent.

In the present invention, Principal Component Analysis (PCA) (201) is used to reduce the dimensions from three (RGB) colour channels to two principal components. Next, Independent Component Analysis (ICA) (203) is used to align the two principal component axes to represent pure density vectors of melanin and haemoglobin. As a result, skin image data that represent skin areas due to melanin and haemoglobin is obtained. The skin pigmentation disorders areas in the pigment melanin image are then segmented using Region Growing method (205). Principal Component Analysis (PCA)

After said skin colour image is being captured, some feature analysis is carried out towards said acquired skin colour image. This is carried out by first using Principal Component Analysis (PCA) (201) which is a dimensional reduction tool, and having skin colour subspace to be represented by its first and second principal components. To find the principal components of an image, the mean value is initially subtracted from each of image spectral bands (RGB) to obtain zero mean variable data.

R = R _Q - ^ _Rg , G = G _Q ~ ^ _Go , Β = Β _ϋ - μ _Βΐί ^

where Ro, Go and Bo denote the image spectral band before subtraction and μ _Κα μ _α and μ _Β denote the mean value of image spectral bands.

The covariance matrix of the three dimensional RGB dataset is then calculated. The covariance matrix of an image in the RGB dataset is defined as follows:

CRR CGR BR

cor

(2) where,

Where X, Y e {R, G, B} and N and μ denote number of pixels in the image and means value, respectively.

The eigenvectors are then determined from the covariance matrix by solving the following equation:

CO V = γλγ ^τ ₍₆₎

Where λ is a diagonal matrix representing eigenvalues of covariance matrix, COV and γ is a matrix of eigenvectors of covariance matrix, COV, arranged as a columns.

The eigenvectors are used as linear transform of original (R, G, B) values. It is reported that the resulting vectors have uncorrelated components, or in other words, the primary axis of the data has been aligned where the

maximal. The vectors in new space [Xi X ₂ X ₃ ] ^T are obtained by

n i . 12 Π 3

7l\ 722 V23

Where 31 732 733 are the eigenvectors of the covariance matrix.

It has been shown that the RGB skin images can be adequately represented by using two principal components with an accuracy of 99.3%.

Independent Component Analysis (ICA)

Skin colour distribution can be formulated as functions of vector of melanin and haemoglobin. Let c, (x, y) and c ₂ (x, y) represent the quantities of the two colour pigments on image coordinate (x, y), which are independent. The colour vectors of the two pigments per unit quantity are denoted as at and a ₂ , respectively. It is also assumed that the compound colour vector, v(x, y) can be calculated using linear combination of the colour vectors as follows:

v(x,y) = a _l c _l (x,y) + a ₂ c ₂ (x,y) (8) This equation can be rewritten as:

v{x,y) = Ac(x,y) (9) The above formula can be estimated using Independent Component Analysis (ICA) (203) as illustrated in FIG. 3, in which A is a mixing matrix and c the sources (independent components). The goal of ICA is to find a linear transformation, Wwith output, u that estimates the sources.

The outputs from the ICA process (203) are images that represent skin areas due to melanin (melanin image) and haemoglobin (haemoglobin image). FIG. 4A to FIG. 4C shows examples of the ICA output, whereby FIG. 4A is the original skin image, FIG. 4B is the skin area image due to haemoglobin (haemoglobin image) and FIG. 4C is the skin area image due to melanin (melanin image). The PC A/ ICA analysis on skin pigmentation disorders lesion is able to detect skin repigmentation areas whose sizes are only 1-by-l pixel.

Region Growing

Referring now to FIG. 5A to FIG. 5D, there is shown the Region Growing (205) segmentation process. Region Growing (205) is an image segmentation procedure that groups pixels or sub regions based on predefined criteria. In this approach, a set of seed points are determined. From these seeds, segmentation regions grow by appending to each seed those neighbouring pixels that have properties similar to the seed or specific ranges of intensity level. Samples representing skin pigmentation disorders lesion and normal skin in the melanin image are first obtained as shown in FIG. 5 A. FIG. 5B shows the histogram of both lesion and normal skin samples. Euclidean distances between intensity of every pixel in the image and mean intensities of lesion samples and normal skin samples are then calculated. The intensity value, Imax, is determined as the intensity value that maximizes the separation of lesion and normal skin. The seed points are then taken from every pixel on lesion sample region as shown in FIG. 5C. Said lesion area is then expanded from seed points by adding to each seeds all neighbouring pixels that have intensity lower than l _m ax. FIG. 5D shows the complete lesion area due to said Region Growing segmentation method.

A total of 41 RGB digital images of skin pigmentation disorders lesions are taken from 18 patients to evaluate the performance of the present invention and its use in clinical setting. Said lesion images are taken from different body areas, in which 9 from the patients' head, 8 from the upper limbs, 15 from the trunks and 9 from the lower limbs. The size of each image is 2240 x 1488 pixels. To remove artefacts due to specular reflections, a low-pass Gaussian smoothing filter is applied after said digital colour image of the subject area is captured in order to obtained higher quality pictures without affecting the size and shape of lesion. The performance of the present invention is evaluated by comparing manually segmented lesion area of an image with the segmented lesion area from present invention. The manual segmentation of lesions is performed to ensure the objectivity of the result. FIG. 6A to FIG. 6C shows examples of the original image (FIG. 6A), the manual segmentation output (FIG. 6B) and the segmentation using present methodology (FIG. 6C). The accuracy, sensitivity and specificity parameters for each image are determined to evaluate the performance of the algorithm and to validate its use in a clinical setting. FIG. 7 shows the results of average performance for different body parts. From the evaluation, the present invention achieved sensitivities of 0.9105 ± 0.0161, specificities of 0.9973 + 0.0009 and accuracies of 0.9901 ± 0.0028 at 95% confidence level. This shows that the present invention consistently exhibited high levels of accuracy, sensitivity and specificity for images taken from different body areas. Therefore, the present methodology is able to provide physicians with an objective and accurate assessment tool to evaluate the changing size of lesions, either over time or with treatment. This will also enable the physicians to assess effectively the efficacy of treatment by comparing the area of lesion before and after treatment.

While the preferred embodiment of the present invention and its advantages has been disclosed in the above Detailed Description, the invention is not limited there to but only by the scope of the appended claim.