DIABETIC IDENTIFICATION USING FINGER NAIL

STRUCTURE
 Abstract: Digital image processing plays critical role in each and every field. Similarly in the field
of medical it can be used to find disorder and organ failure with the help of finger nails structure.
Previously nails structure are used to find whether person have consumed alcohol or not and to find
organ failure. Each person possesses unique nail structure. Even twins have unique nail structure. By
processing these nail structure using various algorithms and technique, the disease can be identified. This
paper explains about identification of disease namely Diabetes, Kidney failure and dermatitis.
Keyword Finger nail structure; unique nail structure; Diabetes; Kidney failure; Dermatitis.

1. INTRODUCTION

A nail is a horn-like envelope covering the tips of the fingers and toes in humans, most non-
human primates, and a few other mammals. The nail consists of the nail plate, the nail matrix and the
nail bed below it, and the grooves surrounding it. The matrix, sometimes called] the matrix unguis,
keratogenous membrane, nail matrix, or onychostroma, is the tissue (or germinal matrix) which the
nail protects. It is the part of the nail bed that is beneath the nail and contains nerves, lymph and
blood vessels. The matrix is responsible for producing cells that become the nail plate. The width and
thickness of the nail plate is determined by the size, length, and thickness of the matrix, while the
shape of the fingertip itself shows if the nail plate is flat, arched, or hooked. The nail bed is the skin
beneath the nail plate. [1]

The nail sinus is where the nail root is i.e. the base of the nail underneath the skin. It originates
from the actively growing tissue below, the matrix The nail plate is the hard part of the nail, made of
translucent keratin protein. Several layers of dead, compacted cells cause the nail to be strong but
flexible. Healthcare and pre-hospital-care providers (EMTs or paramedics) often use the fingernail
beds as a cursory indicator of distal tissue perfusion of individuals who may be dehydrated or in
shock. However, this test is not considered reliable in adults. This is known as the CRT or blanch test.
The fingernail bed is briefly depressed to turn the nail-bed white. When the pressure is released, the
normal pink color should be restored within a second or two. Delayed return to pink color can be an
indicator of certain shock states such as hypovolemia. Nail growth record can show the history of
recent health and physiological imbalances, and has been used as a diagnostic tool since ancient
times. Deep transverse grooves known as Beau's lines may form across the nails (not along the nail
from cuticle to tip) and are usually a natural consequence of aging, though they may result from
disease. Discoloration, thinning, thickening, brittleness, splitting, grooves, lines, small white spots,
receded lunula, clubbing , flatness, and spooning can indicate illness in other areas of the body,
nutrient deficiencies, drug reaction or poisoning, or merely local injury . [14]

Nail disease can be very subtle and should be evaluated by a dermatologist with a focus in this
particular area of medicine. However, most times it is a nail technician who will note a subtle change
in nail disease. Some nail conditions that show signs of infection or inflammation may require
medical assistance. Deformity or disease of the nails is referred to as onychosis.[4]

Nail inspection can give hints to the internal condition of the body as well. Nail disease can
be very subtle and should be evaluated by a dermatologist with a focus in this particular area of
medicine. A nail technician may be the first to note a subtle change in nail health. [3][12][13]Figure
I show finger nail structure.
 Figure I Nail structure

2. PROPOSED WORK

The main aim of disease identification using finger nail structure is to identify the disease
without producing damage to any of the internal structure. For example while taking X-ray it will
emit radiation which in turn causes some trouble to the patients. This technique overcomes all the
disadvantage of the previous technology. Next is that it is economical and doesnt consume much
time. Figure 2 shows the data flow diagram of the process. [5]

PRE- SEGMENTATI FEATURE

PROCESSING ON EXTRACTION

MATCHING CLASSIFICATI
ON

Figure 2

First we have to remove the noise using filters namely mean, median and Gabor filter. The
output of these will give noise less and improved quality image and also ROI i.e.., Region of
Interest. After pre processing the image will undergo several images ,processing technique. Here
we use canny edge detection algorithm for detecting the edge. [11]
The Canny Edge Detector design looks quite similar to the ideal canny edge detector built in
function provided by Matlab. The deviation is seen in terms of the response for each pixel based on
the edge direction & its neighbors. The response is much thinner than the built in function and to
some extent is also two pixels are more disconnected than the original. This may be a result of
the Scaling. Different Scales produce a different result depending on the image properties. The
Canny Edge Detector does many iterations of the nrXnc image matrix. So the order of the
Algorithm is O(nr * nc). Among all the edge detectors, Canny Edge detector gives the best
response. It is the closest to the edges that can be perceived by a human. The goal of any image
processing is to build a system which mimics the human vision system to the closest, canny edge
detectors does a good job at it.[6]
After detecting the edge we will be extracting common parameter among the data set. That
common parameter is responsible for particular disease. For example number of strips or depth of
 particular color in the nail structure is some of the parameter which will be used to detect the
disease using finger nail.

3.FEATURE EXTRACTION

A collection of particular disease affected nail image known as Data set is being processed by
several algorithm for example Canny edge detection algorithm . We use this algorithm because it
will retain certain feature of the image without much data or feature loss. Figure 3 shows feature
extraction. [2]

Figure 3
HSV image is hyper spectral imaging, like other spectral imaging, collects and processes
information from across the electromagnetic spectrum. The goal of hyper spectral imaging is to
obtain the spectrum for each pixel in the image of a scene, with the purpose of finding objects,
identifying materials, or detecting processes. The following function creates the auto-correlogram
vector for an input mage of any size. The different distances which are assumed apriority can be
user-defined in a vector. It implements the algorithm as defined in Huang et al. paper 'Image
Indexing using color auto correlogram.

Output
correlogram_vector=this is a straight vector representing the probabilities of occurrence of 64 quantized
colors. Its total dimension is given by the equation, 64n X 1; where n is the number of
different in f-norm distance.

Usage

To create the auto-correlogram vector for user-defined distances

I=imread('peppers.png');distance_vector=[13];
correlogram_vector=color_auto_correlogram(I,distance_vector);

In general there are two phases in digital image processing namely training phase and testing phase [9].
Training phase includes first four process mentioned in Figure 2 and the testing phase comprises of
training phase and matching. We intend to get the image from the user who needs to test for particular
disease it is shown in figure 4.
 Figure 4 IV CLASSIFICATION

The classification of hyper spectral remote sensing images by support vector machines (SVMs). First,
we propose a theoretical discussion and experimental analysis aimed at understanding and assessing the
potentialities of SVM classifiers in hyper dimensional feature spaces. Then, we assess the effectiveness of
SVMs with respect to conventional feature-reduction-based approaches and their performances in hyper
subspaces of various dimensionalities. To sustain such an analysis, the performances of SVMs are
compared with those of two other nonparametric classifiers (i.e., radial basis function neural networks and
the K-nearest neighbor classifier). Finally, we study the potentially critical issue of applying binary SVMs
to multiclass problems in hyper spectral data. In particular, four different multiclass strategies are analyzed
and compared: the one-against-all, the one-against-one, and two hierarchical tree-based strategies.
Different performance indicators have been used to support our experimental studies in a detailed and
accurate way, i.e.,
the classification accuracy, the computational time, the stability to parameter setting, and the complexity of
the multiclass architecture. The results obtained on a real Airborne Visible/Infrared Imaging Spectral
radiometer hyper spectral dataset allow to conclude that, whatever the multiclass strategy adopted, SVMs
are a valid and effective alternative to conventional pattern recognition approaches (feature-reduction
procedures combined with a classification method) for the classification of hyper spectral remote sensing
data.

Usually, medical experts evaluate NC images manually. NC is also commercially available, mainly used to
detect CTD, RP, and SSC. The advantages are inexpensive, noninvasive, high sensitivity, good specificity
and easy interpretation of the results. Now-a-days researchers propose image-processing methods to
recognize the abnormalities of the capillaries [10] [14].

V RESULT AND CONCLUSION

The normal nail fold capillaroscopic pattern shows a regular disposition of the capillary loops along with
the nail bed. The abnormalities of capillaries (giant capillaries, bushy capillaries, capillaries loop thickness,
its length, density, number of capillaries) can signals the systemic/severe diseases. The capillaries size and
count may vary due to age factor.

Chaikan uses Nail code to increase the accuracy of fingerprint recognition [7]. CCD camera
captures a top-view finger image while the user is touching a fingerprint sensor, and the acquired gray scale
image is preprocessed to enhance the edges, the skin furrows, and the nail shape before a bank of oriented-
filters filters the image. A square tessellation is applied to create a feature map, called a Nail Code [8],
employed in the matching process by employing a Euclidean distance computation. It is in conjunction
with fingerprint for multimodal biometric identification accuracy. The scanned image is cropped manually
for index, middle, ring and little fingers. Gaussian low pass filter is applied, and then converted into binary
image by threshold. Nail is extracted on the basis of orientation-based algorithms to decide Region of
Interest (ROI). To improve the quality of the images, enhancement has been done. Wavelet decomposition
method is used to extract the nail features: longitudinal ridges & lunula. Pre-processed fingernails
decompose using wavelet up to second level. Then, concatenate the both feature in single scale feature
using mean operation. The difference in detail level features as well as approximation level feature. The
metric level features are fusion and classified by Multi-Layer Preceptor (MLP) with a back propagation
learning algorithms. Since the
feature may vary as per age group so, but Authors focusing the base level of decomposition and
concentrates on nails only, this is crucial.
 Figure 5

Figure 5 shows that the disease identified is diabetes and the infected area is shown in the figure. In
Medical Imaging, lots of paper and research are published but they are not applied in the real life . This
paper over come all the challenges and results are discussed.

REFERENCES

[1] Dr. M. Renuka Devi Study of Nail Unit using Image Processing Methods

[2] Pragnesh Parmar, Gunvanti B. Rathod, Forensic Onychology: An Essential Entity against Crime, J
Indian Acad Forensic Med. October-December 2012, Vol. 34, No. 4 ISSN 0971-0973, pp-355-
357

[3] Kavitha Jaba Malar R, Joseph Raj V,1 Geometric Finger Nail Matching using
FuzzyMeasures,IJITEE, ISSN: 2278-3075,
Volume4 Issue-4, September 2014

[4] Shruti Garg, Amioy Kumar, and M. Hanmandlu,Finger Nail Plate: A New Biometric Identifier
IJCISIM,2014, pp. 126 138

[5]Karbhari V. Kale, Yogesh Rode, Majharoddin M. Kazi, Shrinivas Chavan, and Siddharth Dabhade and
Prapti Deshmukh,Multimodal Biometric System Using Finger Knuckle and Nail: A Neural Network
Approach, International Journal of Electrical Energy,
Vol. 1, No. 4, 2013, pp.222-227

[6] Mark C. Hou Sheng-Chieh Huang Hui-Min Wang, A computerized system of nail-fold
capillaroscopy for dry eye disease
diagnosis, Feb 2012, Springer plus, Multidim Syst Sign Process (2012) pp:515524

[7] Doshi, N.P., Schaefer, G. ; Shao Ying Zhu, An Improved Binarisation Algorithm for Nailfold Capillary
Skeleton Extraction,2013, IEEEXplore, SMC, pg. 2565 2569

[8] Chaikan P, Songkhla ; Karnjanadecha, M. The Use of Top-View Finger Image for Personal
Identification Image and Signal Processing and Analysis, 2007. ISPA 2007. 5th International Symposium
on 2007, pg: 343 - 346

[9] Ximena Wortsman Common Applications of Dermatologic Sonography,, J Ultrasound Med

2012,31:97111

[10] Igor Barros Barbos,a Theoharis Theoharis, Christian Schellewald & Cham Athwal,Transient
Biometrics using Finger Nails,2010

[11] Stamatis Gregoriou, George Argyriou, George Larios, Dimitris Rigopoulos, Nail disorders and
systemic disease: What the nails tell us J Fam Pract. 2008 August;57(8):509-514.

[12] M.Cutolo, M.Matucci Cerinic, Nailfold Capillaroscopy and classification criteria for systemic
sclerosis, Clinical and Experimental Rheumatology 2007, pp. 663-665

[13] P Dolezalova, S P Young, P A Bacon, T R Southwood, Nailfold capillary microscopy in healthy

children and in childhood rheumatic diseases: a prospective single blind observational study,
Ann Rheum Dis 2003;62:444449

[14] Lambova S, Mller-Ladner U.,Capillaroscopic pattern inparaneoplastic Raynaud's phenomenon

Rheumatol Int. 2013 Jun;33(6):1597-9

[15] Robert Baran, Rodney PR Dawber, Eckart Haneke, Antonella Tosti, Ivan Bristow, &Luc Thomas,,A
Text Atlas of Nail Disorders Techniques in Investigation and Diagnosi, third edition, Martin Dunitz,
Taylor & Francis group.

[16] Kim HS1, Park MK, Kim HY, Park SH. Capillary dimension measured by computer-based
digitalized image correlated with plasma endothelin-1 levels in patients with systemic sclerosisClin
Rheumatol. 2010 Mar;29(3):247-54

[17] Wu PC, Huang MN, Kuo YM, Hsieh SC, Yu C,. Clinical applicability of quantitative nailfold
capillaroscopy in differential
diagnosis of connective tissue diseases with Raynaud's phenomenon
, J Formos Med Assoc. 2013 Aug;112(8):482-8.