Examensarbeten för masterexamen
Länka till denna samling:
Browse
Senast publicerade
Visar 1 - 5 av 401
- PostBrain Volume and Cortical Thickness in Type 2 Diabetes: Software Implementation and Comparative Analysis(2024) Stark, Isac; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Lundh, Torbjörn; Schain, Martin; Wilander Björk, MarcusThis thesis aimed to investigate the effects of Type 2 Diabetes Mellitus (T2DM) on brain volumetrics, specifically in terms of changes in cortical thickness and grey matter volume. Using the surface based morphometry software FreeSurfer and Fast- Surfer, a comparative analysis was conducted to evaluate the reliability, effectiveness, and runtime of these software in assessing brain morphology in individuals with T2DM compared to healthy controls. MRI data from two datasets, OASIS and MIND, were processed and analysed, focusing on regions of interest included in the Desikan–Killiany–Tourville atlas using both software. Intra-software reliability was assessed through three metrics, Pearson Correlation Coefficient, Intraclass Correlation Coefficient, and Test-Retest Variability. This was done to determine if both software are consistent in their estimates of brain volumetrics. The inter-software reliability was assessed through the Intraclass Correlation Coefficient. The inter-software reliability was performed to determine if FastSurfer gives a similar result to FreeSurfer, which is more extensively validated. Additionally, the runtime performance of FreeSurfer and FastSurfer was compared to determine their efficiency. The results demonstrated high intra-software reliability for both FreeSurfer and FastSurfer in measuring brain volumes and cortical thickness. The two software also demonstrated high inter-software reliability, demonstrating that FastSurfer has a similar accuracy to FreeSurfer. FastSurfer also exhibited a significantly faster runtime. All of this combined highlights FastSurfer’s potential for large-scale studies and clinical applications. However, contrary to expectations based on prior literature, significant differences in brain volumes between the T2DM group and healthy controls were not found. In conclusion, while this study validates the use of FreeSurfer and FastSurfer for neuroimaging analysis, it also highlights the complexity of detecting brain volume changes associated with T2DM, pointing towards the necessity for further investigations. The improved runtime, as well as the high intra- and inter-software reliability of FastSurfer suggests it as a preferable software for this purpose.
- PostEfficient Evaluation of Target Tracking Using Entropic Optimal Transport(2024) Nevelius Wernholm, Viktor; Wärnsäter, Alfred; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Ringh, Axel; Ringh, Axel; Andersson, Adam; Ljung, PerMultiple target tracking deals with the task of estimating targets which appear, disappear, and move within a scene, given data from noisy measurements. To solve this task, a wide range of algorithms can be employed. In order to assess the performance of such algorithms, the so-called GOSPA metric for trajectories can be applied. This metric is formulated as an optimization problem, which has proven computationally demanding for large problem instances. In this thesis, we reformulate this metric in two different ways to obtain optimization problems with optimal transport structure. Following a recent breakthrough in computational optimal transport, we introduce entropic regularization into these formulations. For the regularized problems, we derive and present two numerical algorithms for finding approximate solutions. We test the performance of each algorithm on simulated data with regards to accuracy and computational efficiency. The numerical results suggest that the regularization can be made small enough to allow for an adequate approximation of the GOSPA metric for trajectories while simultaneously allowing a satisfactory convergence rate. Lastly, we compare the running time of our most efficient algorithm with that of a conventional linear programming solver. If a small approximation error is allowed, we find that our algorithm scales better both when the number of trajectories in the data increases, and when the number of considered time steps in the data increases.
- PostA Deep Learning Method for Nonlinear Stochastic Filtering: Energy-Based Deep Splitting for Fast and Accurate Estimation of Filtering Densities(2024) Rydin, Filip; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Larsson, Stig; Andersson, Adam; Bågmark, KasperIn filtering the problem is to find the conditional distribution of a dynamically evolving state given noisy measurements. Critically, designing accurate filters for nonlinear problems that scale well with the state dimension is exceedingly difficult. In this thesis, a novel filtering method based on deep learning solutions to the Fokker–Planck partial differential equation is treated. Training can be performed offline, which results in a computationally efficient algorithm online, even in high dimensions. This is promising for applications which require good real-time performance, such as target-tracking. The filtering method, referred to as Energy-Based Deep Splitting (EBDS), is presented in detail and implemented. The performance of EBDS on different example problems is then investigated and compared to benchmark filters, such as variants of the Kalman filter and particle filters. In one dimension EBDS seems to perform superbly, especially considering how fast the filter is at evaluation. In higher dimensions the method performs worse in comparison to the benchmarks, although it still yields sensible density estimates in most cases. Additionally, convergence for EBDS in the number of prediction steps is investigated empirically for two of the example problems. The results in both examples indicate strong convergence of order 1/2. Lastly, a neural network architecture based on Long Short-Term Memory (LSTM) encoders is proposed for EBDS. This architecture yields reduced errors compared to standard fully-connected networks. In summary, the results indicate that the method is promising and should be examined further. This thesis can be viewed as a reference for future works that aim to apply EBDS in more specific settings or that aim to improve the method further.
- PostImproving Genome Scale Metabolic Models using Gene Regulatory Networks(2024) Davidsson, Liam; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Cvijovic, Marija; Brunnsåker, DanielThis thesis aims to enhance Genome Scale Metabolic Models (GEMs) by integrating Gene Regulatory Networks (GRNs) to improve modeling capabilities for the yeast Saccharomyces cerevisiae. The project involved constructing a pipeline to incorporate gene expression data into GEMs, resulting in a constrained metabolic model with a more diverse and characteristic flux distribution. By utilizing transcription data and transcription factor interactions, the integration aims to provide a more comprehensive understanding of cellular processes. The research demonstrates that incorporating GRNs can enhance the predictive accuracy of GEMs, despite the challenges associated with data complexity and integration methodologies. Potential future work includes upgrading the GRN to a dynamic Bayesian network and exploring the effects of network size on model outcomes.
- PostCharacterization & Anomaly Detection in Radio Sensors: Algorithms for real-time fault detection and classification(2024) Liljenzin, Oscar; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Jonasson, Johan; Liu B, Jing; Winston, Garcia-gabin; Jonasson, JohanIn the ever-evolving field of radio technology, accurate sensor readings are a necessity. Faults in sensors can be a costly endeavour. Thus, it is interesting to explore the possibility of automatic fault detection using sensor readings. First different types of faults that occur in radion sensors are modeled. Then different algorithms for fault detection were implemented and evaluated. From this evaluation, it was found that algorithms based on simple heuristics and statistics can in most cases, perform similarly or even better than more advanced machine learning methods.