|
| |
|
|
Analysis of Microscopic Images of Cancer Cells
Vičar, Tomáš ; Matula,, Petr (oponent) ; Sladoje, Natasa (oponent) ; Kolář, Radim (vedoucí práce)
This dissertation focuses on the analysis of various forms of microscopic image data of cancer cells (static 2D images, static 3D stacks, 2D timelapse live cell imaging). The main focus is on data acquired with a~coherence controlled holographic microscope, which is a~relatively new modality capable of contrast imaging of live cells without staining (label-free) and provide quantitative information (Quantitative Phase Imaging - QPI). In this thesis, the basic procedure for the analysis of cell images is described, where new methods for the individual steps are developed and refined. The largest part of the thesis is devoted to cell segmentation, where classical and deep learning-based methods are summarized. New methods suitable specifically for QPI data are also developed. A~part of the thesis is devoted to the segmentation of 3D fluorescence nuclei and the detection of DNA breaks using deep learning. The thesis also deals with further processing in the form of cell tracking, feature extraction and subsequent analysis, where cell death is detected and suitable interpretable features are developed to classify cell death into apoptotic and lytic. Overall, this thesis contributes to the development of different steps of image analysis of cancer cells and reflects current advances in the image analysis field, deep learning approaches in particular, which is also demonstrated in several research applications.
|
|
|
Analysis of Microscopic Images of Cancer Cells
Vičar, Tomáš ; Matula,, Petr (oponent) ; Sladoje, Natasa (oponent) ; Kolář, Radim (vedoucí práce)
This dissertation focuses on the analysis of various forms of microscopic image data of cancer cells (static 2D images, static 3D stacks, 2D timelapse live cell imaging). The main focus is on data acquired with a~coherence controlled holographic microscope, which is a~relatively new modality capable of contrast imaging of live cells without staining (label-free) and provide quantitative information (Quantitative Phase Imaging - QPI). In this thesis, the basic procedure for the analysis of cell images is described, where new methods for the individual steps are developed and refined. The largest part of the thesis is devoted to cell segmentation, where classical and deep learning-based methods are summarized. New methods suitable specifically for QPI data are also developed. A~part of the thesis is devoted to the segmentation of 3D fluorescence nuclei and the detection of DNA breaks using deep learning. The thesis also deals with further processing in the form of cell tracking, feature extraction and subsequent analysis, where cell death is detected and suitable interpretable features are developed to classify cell death into apoptotic and lytic. Overall, this thesis contributes to the development of different steps of image analysis of cancer cells and reflects current advances in the image analysis field, deep learning approaches in particular, which is also demonstrated in several research applications.
|
|
|
Cell And Sub-Cellular Segmentation In Quantitative Phase Imaging Using U-Net
Majerčík, Jakub ; Špaček, Michal
The ability to automatically segment images, especially microscopy images of cells, opensnew opportunities in cancer research or other practical applications. Recent advancements in deeplearning enabled for effective single-cell segmentation, however, automatic segmentation of subcellularregions is still challenging. This work describes an implementation of a U-net neural networkfor label-free segmentation of sub-cellular regions on images of adherent prostate cancer cells,specifically PC-3 and 22Rv1. Using the best performing approach, out of all that have been tested,we have managed to distinguish between objects and background with average dice coefficients of0.83, 0.78 and 0.63 for whole cells, nuclei and nucleoli respectively
|
|
| |
|
| |
host ::
RSS.