IJCCS (Indonesian Journal of Computing and Cybernetics Systems) Vol.17, No.2, April 2023, pp. 115~126 ISSN (print): 1978-1520, ISSN (online): 2460-7258 DOI: 10.22146/ijccs.74627 ◼ 115 Received May 13 th ,2022; Revised April 29 th , 2022; Accepted April 30 th , 2022 Siamese-Network Based Signature Verification using Self Supervised Learning Muhammad Fawwaz Mayda* 1 , Aina Musdholifah 2 1 Undergraduate Program of Computer Science, FMIPA UGM, Yogyakarta, Indonesia 2 Departement of Computer Science and Electronics, FMIPA UGM, Yogyakarta, Indonesia e-mail: * 1 muhammad.fawwaz.mayda@mail.ugm.ac.id, 2 aina_m@ugm.ac.id Abstrak Penggunaan tanda tangan sangat sering kita jumpai dalam berbagai dokumen publik mulai dari dokumen akademik hingga dokumen bisnis yang menjadi tanda bahwa keberadaan tanda tangan sangatlah krusial dalam berbagai proses administrasi . Seringnya penggunaan tanda tangan bukan berarti sebuah prosedur tanpa celah, tetapi kita harus tetap waspada terhadap pemalsuan tanda tangan yang dilakukan dengan berbagai motif dibelakangnya. Oleh karenannya dalam penelitian ini dikembangkan sistem verifikasi tanda tangan yang bisa mencegah terjadinya pemalsuan tanda tangan dalam dokumen publik dengan menggunakan citra digital dari tanda tangan yang ada. Dalam penelitian ini digunakan jaringan syaraf tiruan dengan arsitektur berbasis jaringan kembar yang juga memberdayakan teknik self supervised learning untuk meningkatkan akurasi pada ranah data yang terbatas. Evaluasi akhir terhadap metode pembelajaran mesin yang digunakan mendapatkan akurasi maksimal sebesar 83% dan hasil ini lebih baik daripada model pembelajaran mesin yang tidak melibatkan metode self supervised learning. Kata kunci— pembelajaran mesin, siamese network, self supervised learning, tanda tangan Abstract The use of signatures is often encountered in various public documents ranging from academic documents to business documents that are a sign that the existence of signatures is crucial in various administrative processes. The frequent use of signatures does not mean a procedure without loopholes, but we must remain vigilant against signature falsification carried out with various motives behind it. Therefore, in this study, a signature verification system was developed that could prevent the falsification of signatures in public documents by using digital imagery of existing signatures. This study used neural networks with siamese network-based architectures that also empower self-supervised learning techniques to improve accuracy in the realm of limited data. The final evaluation of the machine learning method used gets a maximum accuracy of 83% and this result is better than the machine learning model that does not involve self-supervised learning methods. Keywords—machine learning, siamese network, self-supervised learning, signature 1. INTRODUCTION In so many kinds of scenarios that confidentiality can be considered a very important aspect of it, biometrics technology deemed very useful to be used. The main purpose of biometrics technology is to provide a means of identifying an individual based on certain merit such as behavioral or physiological attributes. In many cases, this can be achieved by measuring certain aspects such as fingerprint, iris, palm, voice, facial expression, etc [1]. Amongst them, the handwritten signature also can be used to verify certain individuals and is probably still the most commonly used since the ancient times for so many things such as bank checks, forms, insurance,