Project for my Bachelor's Thesis in Computer Science.

---

Designing Convolutional Neural Networks for Classification of Metastatic Tissue in the Lymph Nodes

Evaluation of the Effect of Feature Map Distributions Across Convolutional Layers on Network Performance

This project was developed to answer to some extent the question: how does an Artificial Neural Network's (ANN) architecture design affect its performance? A more specific question that it gets closer to answering could be: how does the distribution of feature maps across convolutional layers affect the [metastatic tissue] image classification performance of a Convolutional Neural Network (CNN)?

For a more detailed explanation of the project and its outcomes, refer to abstarct or the full report added in the repository.

---

Jupyter Notebooks

The 3 jupyter notebooks used each have their own task:

• DefineArchitectures is the notebook used to define and store new architectures quickly in the form of untrained models.
• ModelTrainingAndTesting is the notebook used to train and test archtiectures and models as well as store the results of these procedures.
• GenerateGraphs is the notebook used to generate graphs with the data obtained from the previous notebook.

---

File hierarchy

The code in this project is very dependent on the file hierarchy used to store different data. Here you can see it briefly defined:

• ./datasets/Cancer Detection/labeled images/ : PCam Dataset location. Divided into 2 folders (named '0' and '1') according to their label.
• ./Graphs/ : Location where graphs generated in the Graph Generation notebook are stored.
• ./Models/ : Location where the untrained models/architectures defined in the Prepare models for testing notebook are stored. Each model is stored in one of the 5 sub-directories (/up, /down, /hill, /valley, /flat) according to their design.
• ./Results/ : Location where all the results are stored. The training histories are stored in a csv file inside the subdirectory "/history CSVs". Test results of trained models are stored in the subdirectory "/Test Results". The txt files containing the history+final result of each model are stored in one of the 5 subdirectories (/up, /down, /hill, /valley, /flat) according to the model design.
• ./Trained Models/ : Location where the fully trained models are stored.