ÖZET
Amaç:
Adli el yazısı incelemelerinde cinsiyet tahmini yıllar boyunca ilgi çekici bulunmuş ve bilim insanları tarafından araştırılmıştır. Aidiyet kararı verilirken adli el yazısı incelemesinde diğer tanı unsurlarının yanında cinsiyet parametresinin destekleyici bir unsur olarak yer alması sonuçların güvenilirliğini artıracaktır. Ayrıca hem erkek hem de kadınlardan oluşan geniş bir şüpheli grubu için araştırılacak kişi sayısının azaltılmasına da yardımcı olacaktır. Bu nedenle adli el yazısı incelemesinde cinsiyet tahmini üzerine araştırmalar yapılmaktadır. Bu çalışmada belirlenen harflerin alçalan ve yükselen kısımlarının ölçülerek cinsiyet tahminine katkısının araştırılması amaçlanmıştır.
Yöntem:
Bu amaç doğrultusunda Yükseköğretim öğrencisi ve mezunu 50 kadın ve 50 erkek katılımcıdan, Türk dilinde başta ortada ve sonda olacak şekilde “b, d, f, g, h, k, t, y, p” harflerini içeren 11 adet cümle yazdırılmak sureti ile yazı örnekleri toplanmıştır. Bu harflerin yükselen ve alçalan kısımları milimetre cinsinden ölçülmüştür. Makine öğrenmesi çalışmalarında en başarılı dört yöntem olan logistik regresyon, K en yakın komşuluk (KNN), destek vektör makineleri (DVM) ve yapay sinir ağları (YSA) seçilerek bu verilere uygulanmış ve yazıyı yazan kişinin yazı karakterine göre erkek ve kadın olup olmadığı belirlenmeye çalışılmıştır.
Bulgular:
Harflerin alçalan ve yükselen kısımlarının ölçülmesi ile kadın ve erkek katılımcılar arasında istatistiksel olarak anlamlı farklılıklar bulunmuştur. Erkeklerde “b, d, h, k, t’nin” yükselen kısımlarının istatistiksel olarak anlamlı derecede daha uzun olduğu belirlenmiştir. Doğruluk oranları lojistik, KNN, DVM ve YSA için sırasıyla 0,65, 0,60, 0,71 ve 0,82’dir.
Sonuç:
Kanaatimizce elde edilen sonuçlar ümit vericidir. Bu konudaki çalışmaların artırılması halinde daha yüksek başarı oranları elde edilebilir ve adli el yazısı incelemesine daha fazla katkı sağlanabilir.
References
1
Bandi KR, Srihari SN. Writer demographic classification using bagging and boosting. In Proceedings of the twelfth International Graphonomics Society Conference. 2005; 26-29; Salerno, Italy, 2005;133-137.
2
Huber RA, Headrick AM. Editors. Handwriting Identification: Facts and Fundamentals. Science, Scientific Method, and Writing Identifications. Florida, Boca Roton: CRC Press LLC. 1999;362-398.
3
Kumar S, Saran V, Vaid, BA, Gupta AK. Handwriting and gender: A statistical study. Z Zagadnień Nauk Sądowych, 2013;95:620-626.
4
Al Maadeed S, Hassaine A. Automatic prediction of age, gender, and nationality in offline handwriting. J Image Video Proc. 2014;10:1-10. https://doi.org/10.1186/1687-5281-2014-10
5
Akbari Y, Nouri K, Sadri J, Djeddi C, Siddiqi I. Wavelet-based gender detection on off-line handwritten documents using probabilistic finite state automata. Image and Vision Computing. 2017;59:17-30. https://doi.org/10.1016/j.imavis.2016.11.017
6
Goodenough FL. Sex differences in judging the sex of handwriting. J Soc Psychol. 1945;22(1):61-68. http://doi.org/10.1080/00224545.1945.9714182
7
Marzinotto G, Nunez JCR, Yacoubi ME, Garcia-Salicetti S. Age and Gender Characterization through a Two Layer Clustering of Online Handwriting. In: Proceeding of sixteenth International Conference on Advanced Concepts for Intelligent Vision Systems. ACIVS; 2015 26-29 Oct; Catania, Italy, 2015;428-439. http://doi.org/10.1007/978-3-319-25903-1_37.
8
Hamid S, Loewenthal KM. Inferring gender from handwriting in Urdu and English. J Soc Psychol. 1996;136(6):778-782. http://doi.org/10.1080/00224545.1996.9712254
9
Binet A. Lea Efivelations de 1 ‘Ecriture d ‘Apres un Controle Scientiftque, Paris. 1906:1-22.
10
Young PT. Sex differences in handwriting. J Appl Psychol. 1931;15(5):486-498. http://doi.org/10.1037/h0072627
11
Tomai CI, Kshirsagar DM, Srihari SN. Group discriminatory power of handwritten characters. In: proceeding of the seventeeth ICPR; 2004 26 Aug; Cambridge, UK, 2004;638-641. http://doi.org/10.1109/ICPR.2004.1334329
12
Topaloglu M, Ekmekci S. Gender detection and identifying one’s handwriting with handwriting analysis. Expert Syst Appl. 2017;79(1):236-243. http://doi.org/10.1016/j.eswa.2017.03.001
13
Siddiqi I, Djeddi C. Raza A, Souici-Meslati L. Automatic analysis of handwriting for gender classification. Pattern Anal Appl 2015;18(4):887-899. http://doi.org/10.1007/s10044-014-0371-0
14
Sokic E, Salihbegovic A. Ahic-Djokic M. Analysis of off-line handwritten text samples of different gender using shape descriptors. In: proceeding of the nineteenth International Symposium on Telecommunications (BIHTEL); 2012 Oct 25-27; Sarajevo, Bosnia and Herzegovina, IEEE publications 2012; 1-6.
15
Liwicki M, Schlapbach A, Loretan P, Bunke H. Automatic detection of gender and handedness from on-line handwriting. In: Proceeding of the thirteenth Conference of the International Graphonomics Society; 2007 Nov 11-14; Melbourne, Australia, 2007;179-183.
16
Ahmed M, Rasool AG, Afzal H, Siddiqi I. Improving handwriting based gender classification using ensemble classifiers. Expert Syst Appl. 2017;85(1):158-168. http://doi.org/10.1016/j.eswa.2017.05.033
17
Morera Á, Sánchez Á, Vélez JF, Moreno AB. Gender and handedness prediction from offline handwriting using convolutional neural networks. Complexity, 2018. http://doi.org/10.1155/2018/3891624
18
Youssef AE, Ibrahim AS, Abbott AL. Automated gender identification for Arabic and English handwriting. In: Proceeding of fifth International Conference on Imaging for Crime Detection and Prevention (ICDP); 2013 16-17 Dec; London, UK, IET publications 2013;1-6. http://doi.org/10.1049/ic.2013.0274
19
Riza LS, Zainafif A, Rasim SN. Fuzzy rule-based classification systems for the gender prediction from handwriting. Telkomnika, 2018;16(6):2725-2732. http://doi.org/10.12928/telkomnika.v16i6.9478
20
Cha SH, Srihari SN. A priori algorithm for sub-category classification analysis of handwriting. In: Proceedings of Sixth International Conference on Document Analysis and Recognition; 2001 Sep 13-13; Seattle, WA, USA, IEEE publications 2002;1022-1025. http://doi.org/10.1109/ICDAR.2001.953940
21
Sesa-Nogueras, E, Faundez-Zanuy, M, Roure-Alcobé, J. Gender classification by means of online uppercase handwriting: a text-dependent allographic approach. Cogn Comput. 2016;8(1):15-29. http://doi.org/10.1007/s12559-015-9332-1
22
Ibrahim AS, Youssef AE, Abbott AL. Global vs. local features for gender identification using Arabic and English handwriting. In: Proceeding of International Symposium on Signal Processing and Information Technology (ISSPIT). 2014:15-17 Dec; Noida, India, IEEE publications 2015;155-160. http://doi.org/10.1109/ISSPIT.2014.7300580.
23
Bouadjenek N, Nemmour H, Chibani Y. Age, gender and handedness prediction from handwriting using gradient features. In: Proceeding of thirteenth International Conference on Document Analysis and Recognition (ICDAR);2015 Aug 23-26; Tunis, Tunisia, IEEE publications 2015;116-1120. http://doi.org/10.1109/ICDAR.2015.7333934
24
Guerbai Y, Chibani Y, Hadjadji B. Handwriting gender recognition system based on the one-class support vector machines. In: Proceeding of Seventh International Conference on Image Processing Theory, Tools and Applications (IPTA); 2017 28 Nov- 1 Dec; Montreal, QC, Canada, IEEE publications 2018; 1-5. http://doi.org/10.1109/IPTA.2017.8310136
25
Dreiseitl S, Ohno-Machado L. Logistic regression and artificial neural network classification models: a methodology review. J Biomed Inform. 2002;35(5-6):352-359. https://doi: 10.1016/S1532-0464(03)00034-0
26
Noble, WS. What is a Support vector machine? Nat Biotechnol. 2006;24(12):1564-1567. http://doi.org/10.1038/nbt1206-1565
27
Skapura DM. “Building Neural Networks” Addison-Wesley, New York; 1996; 29-64 (ch 2).
28
Haykin S. Neural Networks: A Comprehensive Foundation” MacMillan. New York. 2008;22-24 (ch:1).
29
Oztemel E. Artificial neural networks. Papatya Publishing. Istanbul. 2003;29-31 (ch:2) (Turkish Translate)
30
Chaudhuri BB, Bhattacharya U. Efficient Training and Improved Performance of Multilayer Perceptron in Pattern Classification. Neurocomputing. 2000;34(1-4):11-27. http://doi.org/10.1016/S0925-2312(00)00305-2
31
Agresti A. An Introduction to Categorical Data Analysis. Logistic regression. John Wiley and Sons. Inc., 2019;89-92 (ch:4), Third Edition.
32
Lemeshow S, Hosmer D. Applied Logistic Regression (Wiley Series in Probability and Statistics). The Multiple Logistic Regression Model. Wiley-Interscience; 2013; 35-36 (ch:2) Third Edition.
33
Bhatia N, Vandana. Survey of nearest neighbor techniques, IJCSIS. 2010;8(2):302-305.
34
Qiu XY, Kang K, Zhang HX. Selection of kernel parameters for K-NN. In: Proceeding of International Joint Conference on Neural Networks (IJCNN), 2008 1-8 June; Hong Kong, China, IEEE publications. 2008;61-65. http://doi.org/10.1109/IJCNN.2008.4633767
35
Batista Gustavo. EAPA, Silva, DF. How k-nearest neighbor parameters affect its performance. In: Proceeding tenth Argentine Symposium on Artificial Intelligence (ASAI), 2009 24-25 Aug; Mar Del Plata, Argentina. 2009;95-106.
36
Hassaïne A, Al Maadeed S, Aljaam J, Jaoua A. competition on gender prediction from handwriting. In: Proceeding of twelfth International Conference on Document Analysis and Recognition (ICDAR). 2013:25-28 Aug; Washington, DC, USA, IEEE publications 2013;1417-1421. http://doi.org/10.1109/ICDAR.2013.286
37
Demir I. Statistics Guide with SPSS. Regression analysis for categorical data. Istanbul 2020;438-439. (ch:14) (Turkish Translate).
38
Erbilek M, Fairhurst M, Li C. Exploring gender prediction from digital handwriting. In: Proceeding of twenty forth Signal Processing and Communication Application Conference (SIU). 2016;16-19 May; Zonguldak, Turkey, IEEE publications 2016;789-792. http://doi.org/10.1109/SIU.2016.7495858
39
Mirza A, Moetesum M, Siddiqi I, Djeddi C. Gender classification from offline handwriting images using textural features. In: Proceeding fifteenth International Conference on Frontiers in Handwriting Recognition (ICFHR), 2016 23-26 Oct; Shenzhen, China, IEEE publications. 2017;395-398. http://doi.org/10.1109/ICFHR.2016.0080
40
Gattal A, Djeddi C ,Siddiqi I, Chibani Y. Gender classification from offline multi-script handwriting images using oriented basic image features (oBIFs). Expert Syst Appl. 2018;99(1):155-167. http://doi.org/10.1016/j.eswa.2018.01.038
