This section contains an outline for the structure of the final thesis. It is not intended as a final version, rather a draft to discuss.
Parts
This thesis uses a combination of mathematical notation, machine learning/image processing/biological jargon, as well as sections that utilize this jargon extensively to explain the models, ideas, pipelines, … My personal idea is to divide the thesis into (at least) two parts. The first part would contain chapters outlining the jargon, while the second part would be responsible for the explanation of the pipelines, models and software utilizing the information provided in the first part.
Chapters
Each part should contain well-defined chapters. A chapter should contain all but not more information than it is responsible for (e.g. a chapter on image processing should not consider mathematical methods such as OT). It is, however, important to create a well-defined structure in these chapters, starting from a strong foundation and getting more niche the further into the thesis we go.
- A chapter should not be long. I prefer having an extra “chapter”, keeping all chapters well-defined, rather than cramming a new topic somewhere in another chapter.
Subchapters
Each chapter considers one broadly but well-defined topic (e.g. image processing techniques). The chapter can contain multiple subchapters to explain various methodologies.
Example Outline
This is an example outline containing three well-defined parts with multiple chapters (⚠️ draft).
Part I - Fundamentals
Chapter I - Mathematics
The mathematics chapter describes the following subchapters:
- Optimal Transport
- Linear Algebra
Chapter II - Image Processing Techniques
The image processing techniques chapter describes the following subchapters:
- Diffeomorphic Image Registration
- Feature Detectors
Chapter III - Biological Data Collection
The biological data collection chapter describes the following subchapters:
- Spatial Data
- Expression Profiles
Chapter IV - Machine Learning Models
The machine learning models chapter describes the following subchapters:
- Neural Optimal Transport
- Gaussian Processes
- Convolutional Neural Networks
Part II - Literature Review
Depending on what type of literature to include in the review, this part can contain multiple chapters.
Chapter V - Alignment Methods
The alignment methods chapter describes the following subchapters:
- GPSA
- PASTE(2)
- ST-Align
- ELD
- …
Part III - Implementation & Results
TODO
General Points Of Interest
- Try to keep an academic/scientific tone with the appropriate jargon
- Make sure to double check EVERYTHING for plagiarism
- Try to write consistent (i.e. same notation, same jargon, …)