Voltage imaging is a powerful technique for observing fast neural activity, but it often produces images with a high level of noise, making analysis difficult. Deep learning methods have shown promise in denoising such data, but most require large datasets containing both clean and noisy image pairs, which are hard to obtain in real-world settings. To this end, several self-supervised approaches that rely solely on noisy images have been proposed in the literature. In this paper, three self-supervised denoising models, Noise2Void, AP-BSN, and DeepVID v2, are evaluated on both synthetic and real voltage imaging datasets. For the synthetic data, the performance is assessed using PSNR and SSIM, while for the real data, the temporal signal-to-noise ratio (tSNR), a metric well-suited to voltage imaging, is used. Results show that the self-supervised models are effective at denoising both synthetic and real image datasets. In particular, models which use the temporal information of the videos, such as DeepVID v2, obtain the best results.

Voltage imaging using genetically encoded voltage indicators (GEVIs) enables high-speed, population-scale monitoring of neural activity, but it suffers from significant noise due to low photon yield and high frame rates. Effective denoising is essential to recover meaningful signals from such data. In this study, we present a comparative evaluation of three state-of-the-art denoising methods, SUPPORT, DeepCAD-RT, and PMD, on both synthetic and real voltage imaging datasets. Our analysis considers spatial and temporal signal quality, as well as computational efficiency. We find that each method has distinct advantages, and the most suitable choice depends on the specific requirements of the imaging application. SUPPORT is well-suited for tasks requiring spatial detail, PMD offers strong temporal stability and speed, and DeepCAD-RT provides an efficient, balanced alternative. These insights aim to support researchers in selecting and refining denoising tools for real-world neuroscience workflows.

Voltage imaging enables high resolution recordings of neuronal activity but suffers from low signal-to-noise ratios (SNR), primarily due to photon shot noise. Traditional denoising methods like VST-GAT and Penalized Matrix Decomposition (PMD) offer effective noise reduction but often trade off temporal and spatial resolution. Recently, deep learning-based denoising methods, such as CellMincer, have emerged as promising alternatives due to their ability to learn complex signal models without requiring clean training data. This paper compares the performance of traditional and deep learning methods for denoising voltage imaging data using both synthetic and in vivo datasets. Metrics such as SNR, PSNR, and tSNR were used to evaluate performance. The results show that CellMincer outperforms traditional methods on synthetic data and performs competitively on real in vivo recordings, suggesting the viability of self-supervised deep learning for voltage imaging denoising. PMD remains a strong baseline with robust performance across datasets. This comparative study highlights both the potential and current limitations of deep learning approaches and suggests directions for future improvement.

Engineered heart tissues (EHTs) provide a promising platform for modeling cardiac physiology, but their dense and heterogeneous fiber organization makes quantitative analysis highly challenging. This thesis presents an automated pipeline for fiber segmentation and structural analysis of confocal EHT images. The framework integrates frequency based preprocessing using FFT bandpass filtering, state of the art deep learning segmentation models (U-Net, Attention U-Net, and U-Net++), and post-processing refinement through a secondary U-Net. Evaluation was conducted on a synthetic labeled dataset and on real EHT slices with sparse annotations. The results highlight clear trade-offs between model architectures. U-Net produced the most complete and connected fibers but introduced substantial hallucinations. Attention U-Net generated clean outputs but with fragmented fibers, and U-Net++ achieved a balance by capturing directionality and coherence with reduced continuity. Refinement networks were effective at reducing thickness and noise in some cases, but they often removed true fibers and fragmented long structures, providing limited overall benefit. Fiber level metrics and human inspection confirmed these findings, showing that orientation is captured reliably across models, while continuity and connectivity remain major challenges. Overall, the pipeline demonstrates the feasibility of automated structural analysis of EHTs and establishes a foundation for future work with improved datasets, advanced refinement strategies, and broader use of pretrained or structurally informed models.