|
|
Panorama stitching regularly spaced photos
Douša, Marek ; Šikudová, Elena (advisor) ; Goliaš, Matúš (referee)
This thesis addresses the problem of stitching photos into a panorama in two ways - using features and uniformly distributed photos. The mathematical basis is introduced in the form of transformations used in both approaches, followed by the currently mainstream approach to stitching, and finally the theoretical part describes the evenly spaced image approach, for which a demonstrative program is also provided. Both approaches have their advantages and disadvantages on which the thesis emphasizes. This bachelor thesis is designed to serve as a stepping stone for anyone who would like to study the topic of stitching photos into panoramas along with the current state of the art, which has its pros and cons, so that the reader can avoid dead ends or possible surprises during the creation of panoramas.
|
|
|
Neural Cell Segmentation from Fluorescent Microscopy Images of Mouse Brains
Studna, Martin ; Šikudová, Elena (advisor) ; Bída, Michal (referee)
Our objective is to propose a neural cell segmentation algorithm for fluorescent mi- croscopic images of mice brains. We received a dataset from the Laboratory of Neu- rochemistry, Institute of Physiology of the Czech Academy of Sciences in Prague. We conducted various image segmentation experiments to identify a method that can most accurately segment neural cells. Our thesis will present the challenges linked with the segmentation of biological images. Then, we will describe the details of our experiments and the evaluation metrics used for measuring our methods' accuracy. 1
|
|
|
HDR image semantic segmentation
Fadeev, Rustam ; Šikudová, Elena (advisor) ; Mirbauer, Martin (referee)
This work attempts to create a pipeline that accepts the high dynamic range (HDR) input in the .exr format, processes it, and feeds it to the deep neural network, which can perform a semantic segmentation task, detecting the sky. Currently, to our knowledge, available semantic segmentation models cannot accept HDR input in the .exr format. Several models that are trained on HDR input are presented and analyzed here. 1
|
|
|
Corneal neovascularization assesment using machine learning methods
Mačák, Adrián ; Šikudová, Elena (advisor) ; Holeňa, Martin (referee)
In our work, we tried to help ophthalmologists with their research on treating - corneal neovascularization. The massive improvement of hardware and algorithms in machine learning opens new ways to solve many medical imaging problems. During this work, we created a unique dataset and AI-powered solution which quantifies this disease. This solution consists of the computational and user interface part. The computational part uses the deep convolutional neural network with customized U-Net architecture to detect and segment corneal vessels. The user interface provides a toolkit for ophthalmologists to quantify patients' images. Experimentally, we deployed this solution to the hospital FNKV Prague for research purposes. 1
|
|
|
Iris segmentation
Ramesh, Vishal ; Šikudová, Elena (advisor) ; Rittig, Tobias (referee)
Accurate iris image segmentation is crucial to a range of proposed medical diagnosis and treatment systems. Previous models have worked well with healthy eye images but do not generalize to diseased eye images. We work with a dataset where many subjects have eye diseases or deformities. We analyse the performance of the U-Net, a deep learning architecture for semantic segmentation. Our model was trained on a hand-annotated dataset and tuned to generalise on unseen images. Our model achieves a pixel accuracy of 0.8913 on the test set with a relatively short training time. 1
|
|
|
Number recognition on digital displays using Axis cameras' CPU
Kaifer, Jan ; Šikudová, Elena (advisor) ; Kruliš, Martin (referee)
Práce se zabývá implementací aplikace pro kamery AXIS, která rozpoznává digitální čísla na tabulích ve sportovních halách. Existuje mnoho algoritmů pro rozpoznávání čísel, ale velmi málo z nich je dostatečně efektivních, aby je bylo možné spustit přímo na kameře bez dedikovaného výpočetního akcelerátoru. Aplikace provádí detekci pomoci neuronových sítí a s nimi se nám povedlo dosáhnout velmi dobrých výsledků, díky kterým je možné aplikaci využít v produkčním prostředí. Mimo neuronové sítě práce také popisuje naivní algoritmus postavený na technikách počítačového vidění. 1
|
|
|
Comparison of deep learning and classical methods for traffic signs detection
Geiger, Petr ; Šikudová, Elena (advisor) ; Mirbauer, Martin (referee)
The goal of this thesis is to explore and evaluate classic and deep neural network computer vision methods in the task of detection position of a level crossing barrier. This thesis is based on an initial detection algorithm using a Stable Wave Detector. The initial algorithm is optimized both in performance and quality of the results. Both is crucial, because the best method should be suitable as a component of the real-time level crossing safety system. Then an another approach is implemented using deep neural networks and optimized in the same manner. Throughout the work several datasets are created for both training and testing of the algorithms. Both approaches are finally evaluated on the same test datasets and the results are compared.
|
|
|
Differentiable Depth Estimation for Bin Picking
Černý, Marek ; Klusáček, David (advisor) ; Šikudová, Elena (referee)
The goal of this thesis was to investigate the neural 3D surface reconstruction from multiple views with the intent to use the resulting depth maps for bin picking. Survey of papers from 2014 to 2018 showed that none of the state of the art methods would be used to control a robot arm in our setup. Therefore we decided to create our low-level neural approach which we called the EmfNet. The network is based on a pyramidal resolution refining approach. At each pyramid's layer, there are three separate networks that take part in the computation. Each of them has a definite goal, which gives us almost complete understanding of what is going on inside the network. The EmfNet model was partially usable, but we nevertheless extended it to EmfNet-v2. First, another measuring layer was added, which freed EmfNet from depending on an unnecessary hyperparameter. Second, we used constraints on geometry for the network not to be confused by occlusions (cases where a certain part of the surface is visible only from a single camera). Both networks were implemented and tested on a corpus that was created as a part of this thesis. A corpus containing rendered as well as real data. The process of correspondence pairing inside the network can be observed using the visualization tool. We designed a way how to use a robotic arm...
|
|
|
Learning to solve geometric construction problems from images
Macke, Jaroslav ; Šivic, Josef (advisor) ; Šikudová, Elena (referee)
Geometric constructions using ruler and compass are being solved for thousands of years. Humans are capable of solving these problems without explicit knowledge of the analytical models of geometric primitives present in the scene. On the other hand, most methods for solving these problems on a computer require an analytical model. In this thesis, we introduce a method for solving geometrical constructions with access only to the image of the given geometric construction. The method utilizes Mask R-CNN, a convolutional neural network for detection and segmentation of objects in images and videos. Outputs of the Mask R-CNN are masks and bounding boxes with class labels of detected objects in the input image. In this work, we employ and adapt the Mask R- CNN architecture to solve geometric construction problems from image input. We create a process for computing geometric construction steps from masks obtained from Mask R- CNN and describe how to train the Mask R-CNN model to solve geometric construction problems. However, solving geometric problems this way is challenging, as we have to deal with object detection and construction ambiguity. There is possibly an infinite number of ways to solve a geometric construction problem. Furthermore, the method should be able to solve problems not seen during the...
|
|
|
Automatic On-Line Calibration and Calibration Monitoring of Cameras and Lidars
Moravec, Jaroslav ; Šikudová, Elena (advisor) ; Obdržálek, Štěpán (referee)
Title: Automatic On-Line Calibration and Calibration Monitoring of Cameras and Lidars Author: Jaroslav Moravec Department: Department of Software and Computer Science Education Supervisor: doc. RNDr. Elena Šikudová, Ph.D., Department of Software and Computer Science Education Abstract: Cameras and LiDARs are important devices in the automotive indus- try as their combination provides useful information (3D coordinates of a point, its colour and intensity) for perception, localization, mapping and prediction. Successful data fusion and interpretation requires accurate calibration of intrin- sic parameters of the sensors and their 6D relative pose. In this thesis, we present a target-less calibration method on three different calibration tasks. The solu- tion is based on a robust likelihood function constructed over the reprojection error of LiDAR edges relative to image edges. When the calibration slowly wears off, our online recalibration procedure can jointly follow the extrinsic calibration drift with an average error of 0.13◦ in rotation and 4 cm in translation. Based on this recalibration tool, we are also able to monitor the calibration and detect an abrupt decalibration in a couple of seconds. And we also present a fully automatic calibration routine that estimates both the extrinsic and intrinsic...
|
guest ::
