Přístupnostní navigace
E-application
Search Search Close
Bachelor's Thesis
Author of thesis: Vojtěch Hora
Acad. year: 2025/2026
Supervisor: Mgr. Jana Procházková, Ph.D.
Reviewer: Ing. Pavel Mikuláček
This bachelor's thesis deals with the automated registration of microscopic images, a crucial step in image analysis and reconstruction. The work analyzes classical algorithms, such as SIFT, ORB, and ECC, alongside state-of-the-art deep learning pipelines. As the main contribution, a custom Convolutional Neural Network is proposed to directly estimate similarity transformation parameters, providing a fast coarse alignment. In the practical part, the proposed network and baseline methods are experimentally evaluated in terms of accuracy, robustness, and computational complexity. Based on the results, an ideal registration pipeline suited for microscopic images is proposed.
Image registration, microscopic images, neural networks, deep learning, transformation matrix, feature extraction, correlation, computer vision.
Date of defence
09.06.2026
Result of the defence
Defended (thesis was successfully defended)
Grading
A
Process of defence
Student přednesl prezentaci svojí bakalářské práce na téma Moderní metody registrace mikroskopických obrazů a dále odpověděl na otázky dané oponentem. V následné diskuzi se doc. Vašík zeptal na zvýšení robustnosti vůči rotaci obrazu.
Language of thesis
English
Faculty
Fakulta strojního inženýrství
Department
Institute of Mathematics
Study programme
Mathematical Engineering (B-MAI-P)
Composition of Committee
doc. Mgr. Petr Vašík, Ph.D. (předseda) doc. Mgr. Zuzana Hübnerová, Ph.D. (místopředseda) doc. Mgr. Zdeněk Opluštil, Ph.D. (člen) Mgr. Jitka Zatočilová, Ph.D. (člen) Ing. Pavel Loučka, Ph.D. (člen)
Supervisor’s reportMgr. Jana Procházková, Ph.D.
Grade proposed by supervisor: A
Reviewer’s reportIng. Pavel Mikuláček
Grade proposed by reviewer: A
Responsibility: Mgr. et Mgr. Hana Odstrčilová