Závěrečná práce: Bc. Adam Panáček: AI-based Detection of Anthropological Landmarks on 3D Facial Scans
Diplomová práce
AI-based Detection of Anthropological Landmarks on 3D Facial Scans
Anotace
Manuálna anotácia antropologických bodov na 3D skenoch tváre, ktorá je nevyhnutná pre rôzne typy analýz, je pracná a časovo náročná. Táto diplomová práca sa zaoberá riešením tohto problému prostredníctvom integrácie systému na automatickú detekciu antropologických bodov založeného na strojovom učení (ML) do softvéru FIDENTIS Analyst II. Systém využíva metódu Deep-MVLM, ktorá kombinuje viacnásobné 2D …více
Abstract
Manual annotation of anthropological landmarks on 3D facial scans, essential for various facial analyses, is a labor-intensive and time-consuming task. This thesis addresses this challenge by integrating a machine learning (ML)-based automated landmark detection system into FIDENTIS Analyst II software. Leveraging the Deep-MVLM method, the system combines multi-view 2D rendering of 3D scans, convolutional …více
Zadání práce
Study existing methods for automatic detection of landmarks on 3D facial scans. Consider both geometric and ML-based approaches. Consider detection directly from 3D triangular meshes and also from 2D images taken as snapshots from 3D data. Design and implement services for the automatic detection of selected anatomical regions.
The thesis will describe the principles of using landmarks in FIDENTIS Analyst II and review existing solutions for their automatic detection. The selection of a suitable method and technology will be justified.
The proposed automatic ML-based detection system will be implemented and deployed as external REST services, which will be integrated into the desktop application via remote calls. The services and their deployment will be documented. The system will be trained on the dataset provided to the student to increase the precision of detection. The effectiveness and precision of the detection will be experimentally evaluated for different categories of landmarks.
22. 5. 2025 09:08, doc. RNDr. Radek Ošlejšek, Ph.D., učo 3636
Přílohy
Práce na příbuzné téma
Seznam prací, které mají shodná klíčová slova.
-
AI-based Detection of Anatomical Areas on 3D Facial Scans
Ing. Lukáš Kovařík -
Superimposition of 3D face scans to body scans
Ing. Ján Popeláš, učo 484972 -
Automated Pose Estimation of 3D Facial Scans
Ing. Markéta Schlemmerová -
Interactive 3D Mask for Analyzing Human Faces
Bc. Mário Chromík -
3D Masks for the Similarity Measurement of Human Faces
Bc. Martin Bjaloň -
Interactive 3D Mask for Analyzing Human Faces
Bc. Mário Chromík -
Development of Data Deduplication Model for National Oncology Registry
Bc. Jozef Kraus -
Využití umělé inteligence ve výuce zeměpisu
Mgr. Kristýna Michalcová




