- Library Home /
- Search Collections /
- Open Collections /
- Browse Collections /
- UBC Theses and Dissertations /
- Deep learning based image segmentation for applications...
Open Collections
UBC Theses and Dissertations
UBC Theses and Dissertations
Deep learning based image segmentation for applications in neuronal two-photon imaging Fung, Jun Shen (Jason)
Abstract
Brain neurons display considerable plasticity during early circuit formation and throughout life. Advancements in fluorescence labeling, microscopy techniques, and computer vision algorithms enable high-resolution imaging and analysis of brain neurons, tracking sensory stimuli-evoked signaling across entire dendritic arbors. SLAP2 two-photon microscopy allows in vivo 3D brain imaging with subcellular resolution at millisecond rates. Our goal is to use SLAP2 for tracking sensory stimuli-evoked signaling by sampling fluorescent biosensors of neural activity across dendritic arbors of Xaenopus laevis tadpoles. To optimize SLAP2's speed, a computational microscopy software pipeline was developed, incorporating computer vision-based machine learning for rapid neuron segmentation and automated assignment of regions of interest (ROIs) for fast activity sampling. Large 4D datasets resulting from this process are challenging to quantify due to complex tracking of structural changes across time-series of 3D image stacks. Dynamo, an open-source Python application, tackles this issue by streamlining arbor reconstruction, registration of structures across time, and quantitative analyses of growth behavior. This thesis project focuses on training supervised deep learning for semantic segmentation models in SLAP2 microscopy software development, achieving fast 3D imaging sampling rates and upgrading the current Dynamo version for automated dynamic morphometrics in post-hoc analysis. Enhancing both tools will facilitate experiments linking functional and structural plasticity, ultimately improving our understanding of the mechanisms driving these processes.
Item Metadata
Title |
Deep learning based image segmentation for applications in neuronal two-photon imaging
|
Creator | |
Supervisor | |
Publisher |
University of British Columbia
|
Date Issued |
2023
|
Description |
Brain neurons display considerable plasticity during early circuit formation and throughout life. Advancements in fluorescence labeling, microscopy techniques, and computer vision algorithms enable high-resolution imaging and analysis of brain neurons, tracking sensory stimuli-evoked signaling across entire dendritic arbors. SLAP2 two-photon microscopy allows in vivo 3D brain imaging with subcellular resolution at millisecond rates. Our goal is to use SLAP2 for tracking sensory stimuli-evoked signaling by sampling fluorescent biosensors of neural activity across dendritic arbors of Xaenopus laevis tadpoles. To optimize SLAP2's speed, a computational microscopy software pipeline was developed, incorporating computer vision-based machine learning for rapid neuron segmentation and automated assignment of regions of interest (ROIs) for fast activity sampling. Large 4D datasets resulting from this process are challenging to quantify due to complex tracking of structural changes across time-series of 3D image stacks. Dynamo, an open-source Python application, tackles this issue by streamlining arbor reconstruction, registration of structures across time, and quantitative analyses of growth behavior. This thesis project focuses on training supervised deep learning for semantic segmentation models in SLAP2 microscopy software development, achieving fast 3D imaging sampling rates and upgrading the current Dynamo version for automated dynamic morphometrics in post-hoc analysis. Enhancing both tools will facilitate experiments linking functional and structural plasticity, ultimately improving our understanding of the mechanisms driving these processes.
|
Genre | |
Type | |
Language |
eng
|
Date Available |
2023-05-29
|
Provider |
Vancouver : University of British Columbia Library
|
Rights |
Attribution-NonCommercial-NoDerivatives 4.0 International
|
DOI |
10.14288/1.0432733
|
URI | |
Degree | |
Program | |
Affiliation | |
Degree Grantor |
University of British Columbia
|
Graduation Date |
2023-11
|
Campus | |
Scholarly Level |
Graduate
|
Rights URI | |
Aggregated Source Repository |
DSpace
|
Item Media
Item Citations and Data
Rights
Attribution-NonCommercial-NoDerivatives 4.0 International