|
|
The impact of facial expressions on 3D face recognition.
Kováč, Peter ; Goldmann, Tomáš (oponent) ; Pleško, Filip (vedoucí práce)
The thesis "The impact of facial expressions on 3D face recognition" focuses on studying the effect of facial expressions on the accuracy of 3D face recognition. The first chapters explore various 3D modeling techniques and their applications, including 3D Morphable Models (3DMMs), blendshape models, and neural network-based methods like FLAME (Faces Learned with an Articulated Model and Expressions). The thesis then presents a new approach for reconstructing 3D face models from 2D images and proposes an evaluation framework to measure the impact of facial expression changes on 3D face recognition. Experimental results demonstrate how different expressions, such as anger, happiness, and fear, influence recognition accuracy. The findings of this work contribute to a deeper understanding of facial expressions’ role in 3D face recognition and propose potential improvements for enhancing recognition systems.
|
|
|
Deep Neural Networks for Landmark Detection in Images
Stolárik, Samuel ; Španěl, Michal (oponent) ; Kubík, Tibor (vedoucí práce)
Landmark detection in medical images is essential for the diagnosis of various medical conditions. However, it remains laborious, time-consuming, and therefore costly. Deep neural networks have proved themselves to be an applicable solution for the creation of fully automatic landmark detection systems. Automatic landmark detection in the medical domain faces the obstacle of insufficient training data; because of this, using the most efficient representation of the data is important. In this work, three different approaches to the representation of data during the training of deep neural networks were tested. The first two use the U-Net model for heatmap regression and differ in the way how the heatmaps are created. One uses 2D Gaussian heatmaps, which, however, ignore the underlying anatomical information in the input image, and the second one addresses this problem by using novel contour-hugging heatmaps. The third approach uses a graph neural network to process the hierarchical graph representation of the input image. Experiments demonstrated that the combination of the U-Net and the novel contour-hugging heatmaps could detect landmarks in cephalometric images with a mean error of 1.8 mm. This represents a 3.34% improvement over the 2D Gaussian heatmaps. Additionally, the method predicted 70.21% of landmarks with an error of less than 2 mm.
|
|
| |
|
|
Deep Learning Model Uncertainty in Medical Image Analysis
Drevický, Dušan ; Kolář, Martin (oponent) ; Kodym, Oldřich (vedoucí práce)
This thesis deals with quantifying uncertainty in the predictions of deep learning models. While they achieve state of the art results in many areas of computer vision, their outputs are usually deterministic and provide by themselves little information about how certain the model is about its prediction. This is important especially in the domain of medical image analysis where mistakes are costly and the ability to filter uncertain predictions would allow a supervising physician to review the relevant cases. This work applies several different uncertainty measures developed in recent research to deep learning models trained on a cephalometric landmark localization task. They are then evaluated and compared in a set of experiments which aim to determine whether each of the uncertainty measures provides us with useful information about the model's confidence in its predictions.
|
|
|
Deep Learning Model Uncertainty in Medical Image Analysis
Drevický, Dušan ; Kolář, Martin (oponent) ; Kodym, Oldřich (vedoucí práce)
This thesis deals with quantifying uncertainty in the predictions of deep learning models. While they achieve state of the art results in many areas of computer vision, their outputs are usually deterministic and provide by themselves little information about how certain the model is about its prediction. This is important especially in the domain of medical image analysis where mistakes are costly and the ability to filter uncertain predictions would allow a supervising physician to review the relevant cases. This work applies several different uncertainty measures developed in recent research to deep learning models trained on a cephalometric landmark localization task. They are then evaluated and compared in a set of experiments which aim to determine whether each of the uncertainty measures provides us with useful information about the model's confidence in its predictions.
|
|
| |
|
|
Fixing of Facial Triangle Meshes
Nekvinda, Tomáš ; Pelikán, Josef (vedoucí práce) ; Krajíček, Václav (oponent)
Hlavním cílem této práce je navržení algoritmu pro automatické čištění a ořezávání obličejových skenů. Za tímto účelem jsme vyvinuli algoritmus, který sestává ze tří hlavních částí. Jednou z těchto částí je nový algoritmus pro detekci obličejových landmarků, který je založený na poznatcích diskrétní diferenciální geometrie a na metodách strojového učení. Další část se zabývá rozpoznáváním a odstraňováním geometrických a topologických defektů. Poslední část si klade za cíl vyplňování děr v trojúhelníkových sítích, k čemuž je využíváno vylepšení jistého stávajícího algoritmu. Výsledkem této práce je program, který dokáže automaticky ořezávat a odstraňovat typické nedostatky obličejových skenů. Navíc dokáže detekovat sadu výrazných obličejových bodů -- špičku a kořen nosu, ústní a vnitřní oční koutky. Při testování se ukázalo, že si program dokáže bez potíží poradit s daty produkovanými optickým skenerem Vectra3D.
|
host ::
RSS.