S. Korovin
info
Please Note
<p>This page displays the records of the person named above and is not linked to a unique person identifier. This record may need to be merged to a profile.</p>
3 records found
1
Master thesis
(2024)
-
G.M. Rijpkema, C.S. Smith, D. Kalisvaart, S. Korovin, Jaap van der Weerd, Holger Caesar, A. Jakobi
Forensic microtrace investigation relies on a time- and labour-intensive process of manually analysing samples via microscopy. To aid forensic experts in their investigations, an image recognition model for microtrace localisation and classification is needed. This work investigates the trace recognition accuracy that can be achieved by analysing images captured with automated microscopy through deep learning. Fibres, hairs, skin, glass and sand are pixel-wise classified in microscopy scans of tape-lift samples. As deep learning requires extensive amounts of annotated training data, pretraining is investigated to minimise the required annotation workload. ImageNet pretraining, pretraining with self-supervised learning and a sequential application of these approaches are tested. It is found that pretrained models are able to reduce the required annotated data twofold compared to models trained from scratch, while retaining the prediction accuracy. While the ImageNet-pretrained models outperform the self-supervised-pretrained models, the highest accuracy is achieved by combining the two approaches. With this, traces are recognised with a mean intersection over union of 0.56 when training on only 2.2 dm2 of annotated tape lift scans.
...
Forensic microtrace investigation relies on a time- and labour-intensive process of manually analysing samples via microscopy. To aid forensic experts in their investigations, an image recognition model for microtrace localisation and classification is needed. This work investigates the trace recognition accuracy that can be achieved by analysing images captured with automated microscopy through deep learning. Fibres, hairs, skin, glass and sand are pixel-wise classified in microscopy scans of tape-lift samples. As deep learning requires extensive amounts of annotated training data, pretraining is investigated to minimise the required annotation workload. ImageNet pretraining, pretraining with self-supervised learning and a sequential application of these approaches are tested. It is found that pretrained models are able to reduce the required annotated data twofold compared to models trained from scratch, while retaining the prediction accuracy. While the ImageNet-pretrained models outperform the self-supervised-pretrained models, the highest accuracy is achieved by combining the two approaches. With this, traces are recognised with a mean intersection over union of 0.56 when training on only 2.2 dm2 of annotated tape lift scans.
Self-supervised Learning for Tumor Microenvironment Analysis
Addressing Label Scarcity in Multiplexed Immunofluorescence Imaging with Novel Feature Extraction Techniques
The study of tumor microenvironments (TMEs) and immune cell composition in cancer, a disease characterized by uncontrolled growth and spread of tumor cells, has become increasingly important for understanding tumor progression and patient outcomes. Tools such as the TME-Analyzer enable this kind of research, but their manual workflows highlight a common problem in medical imaging: the scarcity of labeled data. This limits the efficiency and applicability of supervised learning algorithms to improve such medical image analysis tools. Self-supervised learning algorithms offer a promising alternative by learning feature representations without requiring labeled data. This thesis aims to address the issue of label scarcity by exploring the potential of self-supervised learning models for TME analysis involving the classification of individual cells in multiplex immunofluorescence (MxIF) microscopy images of triple-negative breast cancer (TNBC) tissue.
To enable the learning of feature representations from MxIF images with an arbitrary number of color channels, this thesis proposes to pre-train an encoder network on every image channel separately according to the SimCLR algorithm and perform classification of multi-channel images by feeding the concatenated feature representation outputs of every channel to a classifier network — referred to as the Siamese configuration. A hyperparameter search is conducted to optimize the SimCLR encoder’s ability to learn high-quality feature representations of individual cells in MxIF images of TNBC tissue. Upon obtaining an optimal set of hyperparameters, the effectiveness of the learned feature representations in improving label-efficiency for individual cell classification is assessed.
The results demonstrate that the proposed Siamese configuration improves the accuracy of classifying the inflammation status of TNBC tumor sections by 2.63%. Additionally, the optimal set of hyperparameters identified through the search include the use of the normalized temperature cross-entropy loss function with low temperature and an added image intensity thresholding term, as well as zoom and brightness/contrast augmentations. Furthermore, the optimized self-supervised learning model improves label-efficiency for individual cell classification, maintaining performance with only 40% of labeled data, while performance drops only when the label percentage is reduced below this threshold.
...
To enable the learning of feature representations from MxIF images with an arbitrary number of color channels, this thesis proposes to pre-train an encoder network on every image channel separately according to the SimCLR algorithm and perform classification of multi-channel images by feeding the concatenated feature representation outputs of every channel to a classifier network — referred to as the Siamese configuration. A hyperparameter search is conducted to optimize the SimCLR encoder’s ability to learn high-quality feature representations of individual cells in MxIF images of TNBC tissue. Upon obtaining an optimal set of hyperparameters, the effectiveness of the learned feature representations in improving label-efficiency for individual cell classification is assessed.
The results demonstrate that the proposed Siamese configuration improves the accuracy of classifying the inflammation status of TNBC tumor sections by 2.63%. Additionally, the optimal set of hyperparameters identified through the search include the use of the normalized temperature cross-entropy loss function with low temperature and an added image intensity thresholding term, as well as zoom and brightness/contrast augmentations. Furthermore, the optimized self-supervised learning model improves label-efficiency for individual cell classification, maintaining performance with only 40% of labeled data, while performance drops only when the label percentage is reduced below this threshold.
...
The study of tumor microenvironments (TMEs) and immune cell composition in cancer, a disease characterized by uncontrolled growth and spread of tumor cells, has become increasingly important for understanding tumor progression and patient outcomes. Tools such as the TME-Analyzer enable this kind of research, but their manual workflows highlight a common problem in medical imaging: the scarcity of labeled data. This limits the efficiency and applicability of supervised learning algorithms to improve such medical image analysis tools. Self-supervised learning algorithms offer a promising alternative by learning feature representations without requiring labeled data. This thesis aims to address the issue of label scarcity by exploring the potential of self-supervised learning models for TME analysis involving the classification of individual cells in multiplex immunofluorescence (MxIF) microscopy images of triple-negative breast cancer (TNBC) tissue.
To enable the learning of feature representations from MxIF images with an arbitrary number of color channels, this thesis proposes to pre-train an encoder network on every image channel separately according to the SimCLR algorithm and perform classification of multi-channel images by feeding the concatenated feature representation outputs of every channel to a classifier network — referred to as the Siamese configuration. A hyperparameter search is conducted to optimize the SimCLR encoder’s ability to learn high-quality feature representations of individual cells in MxIF images of TNBC tissue. Upon obtaining an optimal set of hyperparameters, the effectiveness of the learned feature representations in improving label-efficiency for individual cell classification is assessed.
The results demonstrate that the proposed Siamese configuration improves the accuracy of classifying the inflammation status of TNBC tumor sections by 2.63%. Additionally, the optimal set of hyperparameters identified through the search include the use of the normalized temperature cross-entropy loss function with low temperature and an added image intensity thresholding term, as well as zoom and brightness/contrast augmentations. Furthermore, the optimized self-supervised learning model improves label-efficiency for individual cell classification, maintaining performance with only 40% of labeled data, while performance drops only when the label percentage is reduced below this threshold.
To enable the learning of feature representations from MxIF images with an arbitrary number of color channels, this thesis proposes to pre-train an encoder network on every image channel separately according to the SimCLR algorithm and perform classification of multi-channel images by feeding the concatenated feature representation outputs of every channel to a classifier network — referred to as the Siamese configuration. A hyperparameter search is conducted to optimize the SimCLR encoder’s ability to learn high-quality feature representations of individual cells in MxIF images of TNBC tissue. Upon obtaining an optimal set of hyperparameters, the effectiveness of the learned feature representations in improving label-efficiency for individual cell classification is assessed.
The results demonstrate that the proposed Siamese configuration improves the accuracy of classifying the inflammation status of TNBC tumor sections by 2.63%. Additionally, the optimal set of hyperparameters identified through the search include the use of the normalized temperature cross-entropy loss function with low temperature and an added image intensity thresholding term, as well as zoom and brightness/contrast augmentations. Furthermore, the optimized self-supervised learning model improves label-efficiency for individual cell classification, maintaining performance with only 40% of labeled data, while performance drops only when the label percentage is reduced below this threshold.
Micromachined membrane deformable mirrors (MMDMs) are commonly utilized in Adaptive Optics (AO) systems due to their relatively good performance and cost-effectiveness. However, these deformable mirrors often exhibit nonlinearity at high control magnitudes and a response that is dependent on external factors such as temperature and humidity. In order to overcome these nonidealities, AO controllers typically implement a linear proportional-integral closed-loop control. Nevertheless, if the model is inaccurate, multiple wavefront (WF) measurements are required, which slow down operations. To address these issues, this thesis proposes a novel approach based on the Direct Inverse Control (DIC) framework, which involves modeling and controlling the AO system using shallow neural networks. Specifically, the specialized learning DIC framework is employed. This approach consists of first identifying a forward model of the plant using a neural network, then placing the controller network in series with the plant one, and finally training the controller to make the overall system resemble an identity transfer function. Since the analyzed system is underdetermined, the controller loss function is augmented with a Lagrangian multiplier. This additional term also enables the regularization of the inversion process, which helps to reduce the risk of saturating actuators. The results of this study show that the proposed approach provides better modeling accuracy
than benchmarks, especially in the working ranges where nonlinearities are present. As a result, it enables faster control convergence than the state-of-the-art method when generating large-phase wavefronts. Moreover, when operating online, the DIC-based method demonstrates better stability and similar tracking abilities to Recursive Least Squares. Overall, the proposed approach provides a promising solution to the challenges associated with using MMDMs in AO systems. ...
than benchmarks, especially in the working ranges where nonlinearities are present. As a result, it enables faster control convergence than the state-of-the-art method when generating large-phase wavefronts. Moreover, when operating online, the DIC-based method demonstrates better stability and similar tracking abilities to Recursive Least Squares. Overall, the proposed approach provides a promising solution to the challenges associated with using MMDMs in AO systems. ...
Micromachined membrane deformable mirrors (MMDMs) are commonly utilized in Adaptive Optics (AO) systems due to their relatively good performance and cost-effectiveness. However, these deformable mirrors often exhibit nonlinearity at high control magnitudes and a response that is dependent on external factors such as temperature and humidity. In order to overcome these nonidealities, AO controllers typically implement a linear proportional-integral closed-loop control. Nevertheless, if the model is inaccurate, multiple wavefront (WF) measurements are required, which slow down operations. To address these issues, this thesis proposes a novel approach based on the Direct Inverse Control (DIC) framework, which involves modeling and controlling the AO system using shallow neural networks. Specifically, the specialized learning DIC framework is employed. This approach consists of first identifying a forward model of the plant using a neural network, then placing the controller network in series with the plant one, and finally training the controller to make the overall system resemble an identity transfer function. Since the analyzed system is underdetermined, the controller loss function is augmented with a Lagrangian multiplier. This additional term also enables the regularization of the inversion process, which helps to reduce the risk of saturating actuators. The results of this study show that the proposed approach provides better modeling accuracy
than benchmarks, especially in the working ranges where nonlinearities are present. As a result, it enables faster control convergence than the state-of-the-art method when generating large-phase wavefronts. Moreover, when operating online, the DIC-based method demonstrates better stability and similar tracking abilities to Recursive Least Squares. Overall, the proposed approach provides a promising solution to the challenges associated with using MMDMs in AO systems.
than benchmarks, especially in the working ranges where nonlinearities are present. As a result, it enables faster control convergence than the state-of-the-art method when generating large-phase wavefronts. Moreover, when operating online, the DIC-based method demonstrates better stability and similar tracking abilities to Recursive Least Squares. Overall, the proposed approach provides a promising solution to the challenges associated with using MMDMs in AO systems.