4. 3/19/2016 4
Introduction(cont.)
Brain writing
Conscious Mind
Control-WHAT we write
SUB-Conscious Mind
Controls-HOW we write
Governs our Moods ,feelings ,
behaviors and
A significant part of our personality.
Act of writing involves Conscious and Sub-conscious mind,
Nerves, Muscles and Fingers
The strokes we make while
writing , slant , loops , spacing ,
margins , pressure and many
other are takes care of by the
subconscious mind.
5. Handwriting occurs through
the interactions of many
structures and circuits in the
brain.
When one portion of the brain
is damaged, handwriting is
affected in a way that reflects
the function of that structure
or circuit.
3/19/2016 5
Introduction(cont.)
Brain writing
6. Graphology is a word originated from Greek language.
The first person that carried out systematic observations on the manner of
handwriting was Camillo Baldi in 1622 AD.
3/19/2016 6
2 Greek words
Graphein
(writing)
Logos
(science)
Introduction(cont.)
• Graphology is a scientific method of identifying, evaluating and
understanding personality through the strokes and patterns
revealed by handwriting.
• It is a study of any graphic movements, such as
hand writing, drawings, scribbling and doodles.
• Professional handwriting examiners called graphologist.
7. Graphology reveals insights into the mental, physical of the
writer.
3/19/2016 7
Introduction(cont.)
Habits , Likes and Dislikes
Relationship patterns
Intelligence
Your handwriting develops right from childhood,
adolescence and adulthood.
Emotions ,Feelings and
Temperament
Intuition and Instincts
Creativity and Talents
8. Common Features of Graphology
3/19/2016 8
Introduction(cont.)
Size Baseline Pressure
11. 3/19/2016 11
personality Arabic English
/اصرار
Persistence
/عدوانية
Aggressive
فكر /سيولة
Fluidity of
thoughts
Introduction(cont.)
Personality analysis in Arabic and English
12. 3/19/2016 12
personality Arabic English
مسحوبعاطفيا /
Emotionally
withdrawn
الشخصية /مزدوج
Dual
personality
/دبلوماسي
diplomacy
Introduction(cont.)
Personality analysis in Arabic and English
13. 3/19/2016 13
personality Arabic English
/مجادل
argumentative
وسيطرة /هيمنة
dominant
عالي /تركيز
concentration
Introduction(cont.)
Personality analysis in Arabic and English
14. 3/19/2016 14
personality Arabic English
/غامض
secretive
/الكذب
laying
المشاعر /اتزان
ambivert
Introduction(cont.)
Personality analysis in Arabic and English
18. Handwriting analysis on-line vs. Off-line.
On-line Off-line
Low noise
High recognition
(Automatic conversion of text)
Written on a special digitizer or
PDA.
Elements
digital pen or stylus .
Touch sensitive surface.
Software application.
High noise
Low recognition
(scanned image)
Written on papers
Elements
Fountain pen
A4 paper
Scanner or Digital camera
3/19/2016 18
21. Human behavior
Database Features Classifiers Accuracy
Multiple samples Baseline
pen pressure
Height of the T-bar
ANN(Artificial
Neural Network).
100 writers (70-80 words)
most of them are cursive
, few of them are printed
Size of letters.
Slant of letters and words.
Baseline.
pen pressure.
Spacing between letters.
Spacing between words.
SVM(support
vector machine)
30 writer of
Age between
(20-24) 100 words
size of letters
slant of letters and words
baseline
pen pressure
spacing between letters and
words
Breaks(connected&disconnecte
d)
Margins
Speed
AHWAS
(Automated
Handwriting
Analysis System)
calibrated with
manual analysis.
883 writers (404men ,479
women) age from 20 to 30
years
Size
Width of middle zone letters
Slant
Size of margins
The way of ending the verse
Angularity
Stability of pressure
SVM(Support
Vector machine)
3/19/2016 21
22. Human behavior
3/19/2016 22
Database Features Classifiers Accuracy
50 samples Margins - Baseline
Size - Zonal ratio
Slant - Space
Degree of connection
Myer Briggs
dichotomies
Based on
Keirsey’s
temperament sorter.
handwriting samples Slant - size
Pressure - word spacing
line spacing - Baseline
Least Squares Linear
Regression
100 data set for signature
and 156 type of 26
characters
Curved start - End Streak
Shell - middle streaks
Underline - Extreme margin
Dot structure - Separate
Streaks disconnected
Learning Vector
Quantization (LVQ) for
letters,
ANN and
multi-structure for
signature
10 signatures Curved start - End Streak
Shell - middle streaks
Underline - Extreme margin
Dot structure - Separate
Streaks disconnected
ANN and
multi-structure
24. Forensic
3/19/2016 24
Database Features Classifiers Accuracy
5,600 signatures
(genuine, random and
simulated forgeries).
Static features (caliber ,
proportion , spacing , alignment
to baseline)
Pseudo-dynamic features
(progression ,distribution of
pixels, Form , Slant)
HMM (hidden Markov
models).
Offline signature
1-QU online
signature
database
(194 persons)
2-ICDAR 2009 data
sets
Pressure
Distances
Angles
Speed
Angular speeds
Using multiple
classifiers
1-Random Forest
2-logistic regression
3-linear regression
4-MARS(Multivariate
Adaptive Regression
Spline)
5-Neural Network with
(2,5,10) hidden neuron.
online signature
verification for
both forgeries
and disguised
signatures
29 writers by 10
sample/writer, 34
image/sample (9860
images)
Enlarge to 70 users
2 auxiliary database
final vowel "a"
final vowel “o“
First group(writer and his/her
writing)
Skew ,Slant, Pressure
VowelinfoA,
VowelinfoO Second group
(written words and writer)
Correlation, Length
,Union of letters
,Thinning area
SVM
,
NN+MVA(Most Voted
Algorithm)
Brazilian forensic letter
database(BFL)
(315 writers) 945 images
Texture Features:
Caliber , Progression
Proportion , Pressure
Entry/Exit points , Slant
GLCM descriptors
SVM(support
Vector machine)
dissimilarity
representation
25. Forensic
3/19/2016 25
Database Features Classifiers Accuracy
(BFL) 315 writers
, IAM database 650 writers
texture descriptor
local binary patterns (LBP)
local phase quantization (LPQ)
SVM(support vector
machine)
Brazilian forensic letter
database(BFL) (20 writers)
Brazilian forensic letter
database(BFL) (200
writers)
Number of lines
Proportion of black pixels
Right margin position.
The lower left margin position.
Upper margin position
Bottom margin position
Height of the first word
Axial slant
SVM(support vector
machine)
lAM English handwriting
dataset(657 different
writers )
Directions
Curvatures
Tortuosity
Chain code
Edge based directional
Random forest
lAM English handwriting
dataset
Multi-scale Local Binary
Patterns Histogram texture
features
(MLBPH)
Edge-hinge distribution
Spectral regression(SR-
KDA) for dimensionality
reduction , K-nearest
neighbor classifier
(K-NN)
27. Graphologists have determined that certain breaks in writing, slight
interruptions in the upstroke and in the downstroke , especially in
letters with loops, can point to heart disease. (En) [19]
1-The “Heart Tick”
3/19/2016 27
[2008] Joel Engel , Early Cancer Detection through Graphology Analysis.
Variations of
normal handwriting
Down Strokes
Up Strokes
28. Earlier detecting cancer(cont.)
Finding Cancer in Its Early Stages
Samples of microphotographs of Mrs. B’s handwriting.
3/19/2016
28
Age 28
Age 33
Age 40
First Sample
Second Sample
Third Sample
Smooth, continuous
flow of movement
The writing spreads out
widely
clear interruptions
between descending
and ascending
strokes
29. Graves’ Disease(Manual analysis)
Objectives:
Evaluate handwriting characteristics before and after therapy for
hyperthyroid Graves’ disease (GD).(En)[20]
3/19/2016 29
Database Features Classifier
22patients (15
women, 7 men) with
untreated GD
(median age: 44
years; range: 20–70
years)
write slandered text
before and 12
months after
euthyroid
size of letters(mm)
distance between letters
width of letters
distance between words
extension of
letters(assessed in the
letters l, t, g, and p)
angles(The presence of the
letters a, d, g, and q)
groove depth
Stereoscopic
microscope
Magnifying glass.
Giampaolo Papi,1,2 Cristina Botti,3 Salvatore Maria Corsello,2 Anna Vittoria Ciardullo,1
Alfredo Pontecorvi,2 and Laszlo Hegedu¨s ( 2014) 'The Impact of Graves’ Disease and
Its Treatment on Handwriting Characteristics', Mary Ann Liebert,
Inc., 24,[Online].(Accessed: Number 8, 2014).
30. Graves’ Disease(cont.)
3/19/2016 30
(A) During
hyperthyroidism الغدة نشاط فرط
,الدرقية
handwriting is hypertrophic
and contracted with several
angles.
(B) Post treatment, in the
euthyroid State العادية الحالة ,ف
the handwriting is
characterized by an
increased fluidity.
Standard text written by Seventy-year-old female with Graves’ disease
31. Graves’ Disease(cont.)
3/19/2016 31
In the euthyroid state (B) the size
of the letters (dotted line)
increases compared to the
hyperthyroid state (A).
whereas extensions of letters
(white and gray arrows)
and angles (black arrows)
are reduced
32. Graves’ Disease(cont.)
3/19/2016 32
Thirty-six-year-old female with Graves’
disease. Following recovery from
hyperthyroidism
the distance between the
words (black dotted line)
and the distance between
the letters (gray line) are
reduced,
whereas the width of the
letters (arrow) increased.
34. Arabic Handwriting analysis
3/19/2016 34
Database Features Classifiers Accuracy
Printed text
20 different characters
fonts(320 text images
printed)Handwritten text
22 persons (132
handwriting )
Texture features using (16 Gabor
filters)
WED(weighted Euclidian
Distance)
10 writers , 20 Arabic
images
multi-scale edge-hinge features
grapheme features
K-NN
AHDB Dataset
100 writer
(32,000 Arabic word)
Edge-direction distribution
Moment invariants
Word measurements
(Area , Height,
length from baseline to upper
edge,
length from baseline to the
lower edge )
K-NN
QUWI database that
contains both Arabic and
English handwritings
($commercially)
1017 WRITERS
Directionsاتجاه
Curvaturesتقوس
Tortuosity تعرج
chain codes
edge-based directional
K-NN
35. Arabic Handwriting analysis
3/19/2016 35
Database Features Classifiers Accuracy
QUWI (Arabic and English
handwritings
($commercially)
Directionsاتجاه
curvaturesتقوس
Tortuosityتعرج
chain codes
edge-based directional
Random forest
,Kernel discriminant
analysis using spectral
regression
120 Farsi handwriting
samples
Left and right margins
Word expansion
Letter size
Line and word spacing
Line skew
The ratio of vertical to horizontal
elongation of words
Slant
SVM
36. Summary
# English Arabic
Common
Features
size of letters
slant of letters and
words
baseline
pen pressure
spacing between letters
and words
Breaks(connected
disconnected)
Margins
Speed
Edge-direction
distribution
Moment invariants
Word measurements
Directionsاتجاه
curvaturesتقوس
Tortuosityتعرج
chain codes
Classifier
s
SVM(7) K-NN(3),Random
forest
Database IAM,BFL AHDB(100 WRITERS)3/19/2016 36
38. Research Plan
3/19/2016 38
Building Android Application For Online Arabic Graphology .
We will work on available Database Arabic and English for
writer identification with an improved set of features and
classification methods.
After that we will work on forgery signatures with real
Arabic dataset.
We aspires to work on Diseases diagnoses in Early Stages with
Arabic dataset ,It will required building a database of real
patients .
Goal
First
Second
Future
work
39. References(English)
1. Champa H N,Dr. K R AnandaKumar (2010) 'Artificial Neural Network for Human
Behavior Prediction through Handwriting Analysis', International Journal of
Computer Applications(0975 – 8887), 2(2), pp. 36-41 ,(Accessed: May 2010).
2. Shitala Prasad,Vivek Kumar Singh,Akshay Sapre (2010) Handwriting Analysis
based on Segmentation Method for Prediction of Human Personality using
Support Vector Machine, International Journal of Computer Applications (0975 –
8887), pp. 25-29 ,8(12), (Accessed: October 2010).
3. Vikram Kamath, Nikhil Ramaswamy, P. Navin Karanth, Vijay Desai and S. M.
Kulkarni (2011) 'DEVELOPMENT OF AN AUTOMATED HANDWRITING
ANALYSIS SYSTEM', ARPN Journal of Engineering and Applied Sciences , 6(9),
pp. 135-140 [Online]. Available at: www.arpnjournals.com (Accessed:
SEPTEMBER 2011).
4. UZANNA GÓRSKA,ARTUR JANICKI (2012) 'RECOGNITION OF
EXTRAVERSION LEVEL BASED ON HANDWRITING AND SUPPORT VECTOR
MACHINES1',Perceptual and Motor Skills 114, 3, 857-869, pp. 858-869 [Online].
Available at:(Accessed: May 31, 2012.).
5. Rashi Kacker and Hima Bindu Maringanti, (2012) 'Personality Analysis Through
Handwriting', GSTF Journal on Computing (JoC), 2(1), pp. 858-869 [Online].
(Accessed: April 2012).
6. Abdul Rahiman M,Diana Varghese,Manoj Kumar G (2013) 'HABIT: Handwritten
Analysis Based Individualistic Traits Prediction', International Journal of Image
Processing (IJIP), 7(2), pp. 209-218 [Online]. Available at: (Accessed: 2013).
3/19/2016
39
40. 6-Abdul Rahiman M,Diana Varghese,Manoj Kumar G (2013) 'HABIT: Handwritten
Analysis Based Individualistic Traits Prediction', International Journal of Image
Processing (IJIP), 7(2), pp. 209-218 [Online]. Available at: (Accessed: 2013).
7-Esmeralda C Djamal, Sheldy Nur Ramdlan, Jeri Saputra (2013) 'Recognition of
Handwriting Based on Signature and Digit of Character Using Multiple of Artificial
Neural Networks in Personality Identification , Information Systems International
Conference (ISICO), 2(4), pp. 411-415 [Online]. (Accessed: December 2013).
8-Sandeep Dang,Prof. Mahesh Kumar, Mahesh (2014) 'Handwriting Analysis of
Human Behaviour Based on Neural Network', International Journal of Advanced
Research in Computer Science and Software Engineering, 4(9), pp. 227-232 [Online].
Available at:www.ijarcsse.com (Accessed: September 2014).
9-Luiz S. OLIVEIRA a , Edson JUSTINO a , Cinthia FREITAS a and Robert
SABOURINb (2005) 'The Graphology Applied to Signature Verification', ,(Retrieved
on:10 December2015).
10-Abdelâali Hassaïne,Somaya Al-ma'adeed (2012) 'An Online Signature Verification
System for Forgery and Disguise Detection', [Online]. : (Accessed: NOVEMBER
2012). Retrieved on: 07 October 2015
11-Omar Santana, Carlos M. Travieso, Jesus B. Alonso, Miguel A. Ferrer (2010) 'Writer
Identification Based on Graphology Techniques', IEEE A&E SYSTEMS MAGAZINE,,(),
pp. [Online]. Available at: (Accessed: JUNE 2010).
12-R. K. Hanusiak · L. S. Oliveira · E. Justino · R. Sabourin (2011) 'Writer verification
using texture-based features', Springer, (), pp. 214 -226,[Online]. (Accessed: 24 May
2011). 3/19/2016 40
References(English)
41. 13-D. Bertolini a, L.S. Oliveira a,⇑, E. Justino b, R. Sabourin c ( 2012) 'Texture-based
descriptors for writer identification and verification ', Elsevier Ltd, 40(6), pp. 2069–2080
[Online]. Available at: 18 October 2012 (Accessed: May 2013).
14-A. M. M. M. Amaral, C. O. A. Freitas, F. Bortolozzi. “The Graphometry applied to
writer identification”. In Proceedings of the 2012 International Conference on Image
Processing, Computer Vision, and Pattern Recognition, Las Vegas, USA, vol.1, pp.10-
16, 2012.
15-Aline Maria M. M. Amaral1,2, Cinthia O. A. Freitas2, and Flavio Bortolozzi1.
“2013)Multiple Graphometric Features for Writer Identification as part of Forensic
Handwriting Analysis”. In Proceedings of the 2013 International Conference on Image
Processing, Computer Vision, and Pattern Recognition, Las Vegas, USA, vol.1, pp.10-
16, 2013.
16-A. Hassa¨ıne, S. Al-Maadeed, and A. Bouridane, “A set of geometrical features for
writer identification,” Neural Information Process. Berlin Heidelberg: Springer,, vol. 45,
pp. 584–591,2012.
17-E Khalifa S Al-Maadeed2, M A Tahir3, F Khelifil and A Bouridane1 ( 2013) 'OFF-
LINE WRI TER I DENTIF ICATI ON U S ING MULTI- SCALE LOCAL BINARY
PATTERNS AND SR-KDA', IEEE, [Online].
18-Shweta Hegade1, Gargee Hiray2, Prajkta Mali3, Prof. Punam Raskar4 (2015)
'FODEX: Forensic Document Examiner –Using Graphology Science', IJETST, 2(3), pp.
2042-2045 [Online]. Available at: (Accessed: March 2015).
19-[2008] Joel Engel , Early Cancer Detection through Graphology Analysis.
20-Giampaolo Papi,1,2 Cristina Botti,3 Salvatore Maria Corsello,2 Anna Vittoria
Ciardullo,1 Alfredo Pontecorvi,2 and Laszlo Hegedu¨s ( 2014) 'The Impact of Graves’
Disease and Its Treatment on Handwriting Characteristics', Mary Ann Liebert,3/19/2016 41
References(English)
42. 21-FEDDAOUI Nadia, HAMROUNI Kamel (2006) 'Personal identifi'cation
based on texture analysis of Arabic handwriting text', IEEE, (), pp. 1302-1307
[Online].
22-Somaya Al-Ma’adeed, Amat-AlAleem Al-Kurbi, Amal Al-Muslih, Reem Al-
Qahtani, Haend Al Kubisi (2008) 'Writer Identification of Arabic Handwriting
Documents Using Grapheme Features', IEEE, (), pp. 923-924 [Online].
23-Somaya Al-Ma’adeed, Eman Mohammed, Dori Al Kassis, Fatma Al-Muslih,
(2008) 'Writer Identification using Edge-based Directional Probability
Distribution Features for Arabic Words', IEEE, (), pp. 582-590 [Online].
24-Somaya Al-Maadeed (2012) 'Text-DependentWriter Identification for Arabic
Handwriting', Journal of Electrical and Computer Engineering, 2012(), pp. 8
[Online].
25-Somaya Al Maadeed, Wael Ayouby, Abdelˆaali Hassa¨ıne, Jihad Mohamad
Aljaam (2012) 'QUWI: An Arabic and English Handwriting Dataset for Offline
Writer Identification', IEEE, (), pp. 746-751 [Online].
26-Somaya Al–Maadeed, Fethi Ferjani, Samir Elloumi, Abdelaali Hassaine
and Ali Jaoua (2013) 'Automatic Handedness Detection from Off-Line
Handwriting', IEEE, (), pp. 119-124 [Online].
27-Al Maadeed and Hassaine: Automatic prediction of age, gender, and
nationality in offline handwriting. EURASIP Journal on Image and Video
Processing 2014 2014:10.
28-Somayeh Hashemi1, Behrouz Vaseghi2, Fatemeh Torgheh3 (2015)
'Graphology for Farsi Handwriting Using Image Processing Techniques', IOSR
Journal of Electronics and Communication Engineering (IOSR-JECE), 10(3),
pp. 01-07 [Online]. Available at:(Accessed: May - Jun.2015).3/19/2016 42
References(Arabic)
Smooth, continuous flow of movement
The strokes have an oval shape, the turns
from descending to ascending strokes
are narrow, curved, and show continuity of
movement throughout.
A regular pattern of heavier (wider and darker)
descending strokes and lighter ascending
strokes prevails throughout the sample.
Heavier descending strokes and
lighter ascending strokes is still preserved
The narrow turns have disappeared.
The writing spreads out widely
The strokes are much weaker and highly
unstable غير مستقر
Clear interruptions between descending
and ascending strokes are also visible.
Breakdown of every phase of the writing process.
The strokes are stiff or formless.
The pressure is uneven, sometimes too heavy,
and in other strokes too light.
There are clear interruptions between
descending and ascending strokes
The study of Arabic handwriting identification is limited
The recognition of Arabic characters is also important for
certain non-Arabic-speaking languages, such as Farsi, Kurd,
Persian, and Urdu.