Proceeding of the Faculty of technical Sciences, Novi Sad

Mechatronics

Vol. 41 No. 01 (2026): Proceedings of the Faculty of Technical Sciences

IMPLEMENTATION OF A SIAMESE NEURAL NETWORK FOR FACE VERIFICATION BASED ON ONE-SHOT PREDICTION

  • Petar Surla
DOI:
https://doi.org/10.24867/33IH02Surla
Submitted
August 23, 2026
Published
2026-08-23

Abstract

This paper presents the implementation of a Siamese neural network for face verification. It describes the process of dataset creation, image preprocessing, as well as the design and training of the network. The objective of this work is to apply one-shot prediction to train the system using a small amount of data. Experimental results and comparisons with other models are provided, along with final conclusions.

References

  1. [1] Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep Learning. MIT Press.
  2. [2] Hesaraki, S. (2023, Novembar 15). One-shot learning. Preuzeto sa: https://medium.com/@saba99/one-shot-learning-312e385993bc
  3. [3] Koch, G., Zemel, R., & Salakhutdinov, R. (2015). Siamese neural networks for one-shot image recognition.
  4. [4] Labeled Faces in the Wild (LFW) Dataset. (2018). Preuzeto sa: https://vis-www.cs.umass.edu/lfw/
  5. [5] Raschka, S., Liu, Y., & Mirjalili, V. (2022). Machine Learning with PyTorch and Scikit Learn. Packt Publishing.