Master Degree / Yüksek Lisans Tezleri
Permanent URI for this collectionhttps://hdl.handle.net/11147/3008
Browse
3 results
Search Results
Now showing 1 - 3 of 3
Master Thesis Development of a Machine Learning Platform for Analysis of Mitochondrial Features in Live-Cell Images(01. Izmir Institute of Technology, 2022) Tarkan, Yalçın; Tuğlular, Tuğkan; Tuğlular, TuğkanIt is a laborious and error-prone manual process to mark the organelles in 2D and 3D images of living cells and identify the behavioral feedback to stimulations under measured conditions. This manual process can be simplified by being largely automated with machine learning techniques. We created a machine learning-based software platform named MitoML, which extracts sub-cellular structures, specifically mitochondria, and helps to identify the effects of external factors or changes under natural conditions. We investigate appropriate machine learning techniques for these objectives. Image processing and segmentation techniques with neural networks, enable researchers to carry out experiments with much better accuracy and a larger scale by automatically segmenting and counting the mitochondria, calculate the energy potentials based on region brightness. This way, analysis of mitochondria feedback in healthy and cancer cells under various conditions, such as nanomedicine and different treatment therapies, can be performed using MitoML. As a result of our work, we proposed a cascaded neural network architecture that can identify and count mitochondria, quantify its energy levels in fluorescence and other microscopy images, fast and at a standard reliable accuracy. Our test results outperformed the classical image processing algorithms provided in segmentation tools and software for medical image segmentation which was taken as a base line. Achieved accuracy rates 93.4% and %89.6 according to Dice and IoU metrics respectively are also better than any other related work encountered during the research. The proposed method can be improved to cover other sub-cellular structures relieving the researchers from non-standardized and laborious manual work which is prone to human error.Master Thesis Hierarchical Image Classification With Self-Supervised Vision Transformer Features(Izmir Institute of Technology, 2022) Karagüler, Caner; Özuysal, MustafaThere are lots of works about image classification and most of them are based on convolutional neural networks (CNN). In image classification, some classes are more difficult to distinguish than others because of non-even visual separability. These difficult classes require domain-specific classifiers but traditional convolutional neural networks are trained as flat N-way classifiers. These flat classifiers can not leverage the hierarchical information of the classes well. To solve this issue, researchers proposed new techniques that embeds class-hierarchy into the convolutional neural networks and most of these techniques exceed existing convolutional neural networks' success rates on large-scale datasets like ImageNet. In this work, we questioned if a hierarchical image classification with self- supervised vision transformer features can exceed hierarchical convolutional neural networks. During this work, we used a hierarchical ETHEC dataset and extract attention features with the help of vision transformers. Using these attention features, we implemented 3 different hierarchical classification approaches and compared the results with CNN alternative of our approaches.Master Thesis Localization of Certain Animal Species in Images Via Training Neural Networks With Image Patches(Izmir Institute of Technology, 2017) Orhan, Semih; Baştanlar, YalınObject detection is one of the most important tasks for computer vision systems. Varying object size, varying view angle, illumination conditions, occlusion etc. effect the success rate. In recent years, convolutional neural networks (CNNs) have shown great performance in different problems of computer vision including object detection and localization. In this work, we propose a novel training approach for CNNs to localize some animal species whose bodies have distinctive pattern, such as speckles of leopards, black-white lines of zebras, etc. To learn characteristic patterns, small patches are taken from different body parts of animals and they are used to train models. To find object location, in a test image, all locations are visited in a sliding window fashion. Crops are fed to CNN, then classification scores of all patches are recorded. To illustrate object location, heat map is generated by the classification scores of the patches. Afterwards, heat maps are converted to binary images and end up with bounding box estimates of objects. The localization performance of our Patch-based training is compared with Faster R-CNN – a state-of-the-art CNN-based object detection and localization algorithm. While evaluating the performances, in addition to the standard precision-recall metric, we use area-precision and area-recall which represent the potential of Patch-based Model better. Experiment results show that the proposed training method has better performance than Faster R-CNN for most of the evaluated classes. We also showed that Patch-based Model can be used with Faster R-CNN to increase its localization performance.