Pandit et al., International Journal of Advanced Engineering Technology E-ISSN 0976-3945
Research Paper
A SYSTEM FOR PALM COLOR ANALYSIS IN HEALTHCARE
Dr. Hardik B. Pandit and Prof. Dipti Shah
Address for Correspondence
Department of Computer Science, Sardar Patel Univeristy, Vallabh Vidyanagar. Gujarat. 388 120 (India)
ABSTRACT
This research paper explains a prototype based on digital image processing and analysis in field of healthcare. In medical
science, color of human palm is carefully observed by doctors to get primary idea about health of the patient. Different
colors of palm specify certain diseases. Human eye has limitation in identification of colors and resolution. Computer can
help here to analyze color of human palm using digital image processing techniques. The working prototype is fully
automatic, i.e. there is no human intervention in the process of color analysis. The paper also discusses results achieved by
this prototype.
KEY WORDS Digital image processing, color image processing, palm color analysis, healthcare
INTRODUCTION
The field of medical science is widely growing.
There are so many tests which help medical
practitioners to diagnose patient’s disease. One of
these techniques is palm color analysis. Different
colors of human palm give idea about health status of
the person. The color of a normal palm is a light red
or pinkish red with a shiny, smooth texture. If the
color appears either darker or lighter than normal,
this may indicate that the condition of health is
abnormal. Following are examples of abnormal palm
colors [2]:
(i) Pale White:
• A palm appearing pale white in colour
indicates anaemia or possibly occult bleeding.
[2].
• If the palm looks white, this usually indicates
lung disease or inflammation in the body [2].
(ii) Blue:
• A Blue palm usually indicates intestinal
obstruction [2].
(iii) Green:
• A Palm with a dark green colour usually
indicates obstruction in the circulation of the
blood [2].
• A greenish palm that is not dark may indicate
anaemia or spleen /stomach disease [2].
(iv) Yellow:
• A sallow yellow palm usually indicates
chronic disease. This is because chronic
disease typically affects the spleen and
stomach. Yellow is the colour of earth and
the spleen and stomach pertain to earth. In
this case, the spleen and stomach are vacuous
and weak [2].
• A palm with a bright, golden yellow colour is
often seen in the liver disease accompanied
by jaundice. In this case, there is
liver/gallbladder damp heat [2].
• If the palm skin grows thicker, stiffer and is
dry with a light yellow, shiny, smooth surface
this is called palm calcar keratosis [2].
• A palm that looks yellowish brown and has
no sheen indicates the possibility of cancer
[2].
(v) Red:
• A palm with Red, net like capillaries often
appears with Vitamin C deficiency.
• If the whole palm is covered with dark red or
purple spots, this is usually seen in liver
disease [2].
Int J Adv Engg Tech/Vol. V/Issue I/Jan.-March,2014/30-32
• When the skin on the surface of the palm and
especially the thenar and hypothenar
eminences and the fingertips appears
congested red, this is commonly due to
cirrhosis of the liver or cancer of the liver [2].
• The palm that first appears red and gradually
changes to dark purple is usually a sign of
heart disease. It is an indication that the
disease is worsening. An excessively red
palm indicates a tendency for apoplexy [2].
• If the whole palm of a hypertensive patient
appears black like tea, this is forewarning of
cerebral haemorrhage. If the skin of red palm
is soft as satin, this indicates a tendency
towards rheumatic fever. In general in
Chinese medicine, red indicates the presence
of heat. As the red becomes darker, this
indicates that heat is being complicated by
stagnation and statis [2].
(vi) Purple:
• When the subcutaneous tissue of the palm
shows prune coloured purple, it indicates
serious infectious shock [2].
(vii) Grey:
• Thin Cigarette-ash like spots on the palm are
sign of heart disease in a heavy smoker [2].
(viii) Black:
• A palm that looks black is often seen in
kidney disease [2].
• If the central part of the palm looks brownish
–black, this often indicates gastrointestinal
disease [2].
• A dark purple of black colour appearing from
the wrist to the hypothener eminence is a sign
of low back disease due to wind. The same
colour may also appear on the medical side of
the foot and ankle [2].
Following are two sample images which show
abnormal colouring and surfaces of palm. The
diseases associative with each palm are named to
give idea about relationship between palm and
diseases. Figure 1 shows symptoms of Raynaud’s
syndrome in human palm.
(a) The acute phase,
showing severe
blanching of the tip of
one finger
(b) Primary Raynaud’s
syndrome, occasionally
progresses to fingerprint
ulceration or even gangrene
Figure 1: Raynaud’s Syndrome in Human Palm [1]
Pandit et al., International Journal of Advanced Engineering Technology E-ISSN 0976-3945
Figure 2 shows cases of Hypercarotenaemia. In this
case, a yellowish discoloration is seen but not like
jaundice.
Figure 2: Hypercarotenaemia [1]
Thus, one can understand the importance of palm in
identification and prediction of diseases in medical
science.
A prototype is developed which analyzes human
palm for its color without human intervention. The
algorithm for the same is discussed in this paper.
Work Contribution
To start the process of palm color analysis, the palms
of left and right hands are scanned through flat bed
scanner. The images are uploaded to the prototype for
further processing. An algorithm for extraction of
palm from rest of the image [3] is executed to
separate the palm from its background.
Since the analysis is to be done for only palm region,
to reduce the pixel to be processed and to simplify
the processing, the resulting images after applying
palm extraction algorithm [3] are cropped using
cropping algorithm. The goal of cropping is to get the
four points shown in figure 3 named topmost point,
bottommost point, leftmost point, and rightmost
point. Using these four points, image is cropped. To
understand the cropping algorithm properly, consider
the image displayed in following figure 3 for
reference.
Fig 3: Points and Directions of interest for
CropAlgorithm
Cropping algorithm is as follows:
Algorithm 1: Cropping of Palm from rest of the
area
Input: Output image of palm extraction algorithm.
Steps:
1. Define max_x = max_y = (0,0) and min_x =
min_y = (image_width, image_height) as initial
values for rightmost point, bottommost point,
leftmost point, and topmost point respectively.
2. For each row of pixels follow the direction
shown in figure. The first point of the palm
region encountered would be the topmost point
(the point with minimum value of ycoordinate). Save this point as “min_y”.
Proceeding in the same fashion, the last point
Int J Adv Engg Tech/Vol. V/Issue I/Jan.-March,2014/30-32
(the point with highest value of y-coordinate) of
palm region in vertical direction is also
obtained. Save this point as “max_y”.
3. For each column of pixels follow the direction
shown in figure. The first point of the palm
region encountered would be the leftmost point
(the point with minimum value of xcoordinate). Save this point as “min_x”.
Proceeding in the same fashion, the last point
(the point with highest value of x-coordinate) of
palm region in vertical direction is also
obtained. Save this point as “max_x”.
4. Get four end points of palm region leftmost
point = x-coordinate of min_x; rightmost point
= difference between x-coordinates of max_x
and min_x; topmost point = y-coordinate of
min_y; bottommost point = difference between
y-coordinates of max_y and min_y.
Crop the rectangle using these four points, which will
touch the four boundaries of palm.
Result of algorithm-1
The result of algorithm for cropping of palm is shown
in figure 4.
Fig 4: Result of Cropping Algorithm
Palm color analysis
To analyze the color of palm, the prototype analyzes
color of each pixel on the region of palm. In the
cropped image of palms, there are only two sets of
pixels:
(i) Yellow pixels with RGB component value (255,
255, 0) and
(ii) Non-yellow pixels which are the pixels of palm.
The prototype analyzes only non-yellow pixels.
Following algorithm is designed and implemented for
palm color analysis.
Algorithm-2: Palm Color Analysis
Input: Output image of algorithm 1.
Steps:
1. Get the color of pixel in terms of RGB
component values.
2. Compare the value of each component with the
value stored in knowledge base. While
comparing value of each component consider
the deviation for each component as stored in
knowledge base as mentioned in section 4.2 of
chapter 4.
3. If the match is found, consider this pixel as the
pixel of mentioned color and increase the
counter for this color by one (the counter for
each color is initialized by 0). The value of
counter for each color will give number of
pixels with respective color.
4. Count the percentage of pixels with each color.
If the percentage of pixels found with given
color is less than 5%, the case is not considered.
If the percentage of pixels found with given
color is between 5% to 10%, then the level of
Pandit et al., International Journal of Advanced Engineering Technology E-ISSN 0976-3945
disease is 0, which shows probability of
disease. If the percentage of pixels found with
given color is between 11% to 30, then the level
of disease is 1, which shows initial stage of
disease related to respective color. If the
percentage of pixels found with given color is
more than 30%, then there are strong chances of
materialization of the disease associated with
respective color.
Result of algorithm 2
Figure 5 shows the output of the process of palm
color analysis. There are three major components of
output. (i) The input and output images, (ii) the
statistical analysis, and (iii) textual prediction.
Figure 5 (a) and (b): Result of Palm Color Analysis
• The input and output images: the output of palm
color analysis is shown in pictorial format for
both the palms. The images of left and right palms
are colored according to the color of pixels at
respective places. The displayed color in output
image and original color of respective pixels is
mentioned in neighboring table on the same page.
This component gives visual effect of any
abnormality found in palm. Simply, if in output
image any region of palm is found colored
differently, user should refer the neighboring table
to get the meaning.
• The statistical analysis: the table displayed in the
right side shows statistical description of palm
color analysis. The table explains meaning of each
color painted in output image by giving actual
color name of that region. The table also shows
number of pixels found with each color. In the
end, the table shows percentage of pixels found
with each color. Based on these values only, the
predictions are made by the prototype.
• Textual Prediction: the output screen also has a
section named “Conclusion of Analysis”, which
displays the prediction made by the prototype in
textual format. The predictions contain the name
of the disease if found, its stage, and if necessary
it gives advice.
CONCLUSION
Conclusion of this research work is shown in figure 5
(a) and (b). The prototype is working successfully to
analyze palm color. It gives results which are easy to
understand by the user. Thus, using this prototype it
Int J Adv Engg Tech/Vol. V/Issue I/Jan.-March,2014/30-32
is easy to overcome the limitations of human eye for
color identification and getting better assistance in
decision making activity by medical practitioners.
REFERENCES
1. Graham Douglas, Fiona Nicol, Colin Robertson –
“Macleod’s Clinical Examination”, twelfth edition, Elsevier
publication, ISBN: 9780443068485
2. Xiao-Fan Zong, Gary Liscum – “Chinese Medical
Palmistry-Your Health in Your Hand”, Blue Poppy Press,
ISBN: 0-938185-64-3
3. Hardik Pandit, Dipti Shah, “The Model for Extracting a
Portion of a Given Image Using Color Processing”,
International Journal of Engineering Research &
Technology (IJERT) ISSN: 2278-0181 Vol. 1 Issue 10,
December- 2012.
4. Rafael C. Gonzalez, Richard E. Woods - “Digital Image
Processing” second edition, Pearson Education, ISBN: 817808-629-8
5. B. Chanda, D. Datta Majumder – “Digital Image Processing
and Analysis” second edition, PHI Learning Private
Limited, ISBN: 978-81-203-4325-2
6. Kenneth A. Kozar – “Humanized Information Systems
Analysis and Design”, McGraw-Hill Book Company,
ISBN: 0-07-035600-9
7. Bill Evjen, Scott Hanselman, Devin Rader – “Professional
ASP.NET 3.5 (SP1) in C# and VB”, Wiley India Pvt. Ltd,
ISBN: 978-81-265-2104-3