Please enable JavaScript.
Coggle requires JavaScript to display documents.
Thesis - Coggle Diagram
Thesis
Chapter2: Nuclear detection and Segmentation
Detection
Methods
1- MDN
2- MDN for nuclear detection
3- Network architecture
Results
1- Performance with different network
2- Performance with different loss
3- Effect of number of Gaussians
Segmentation
Method
1 -Spatially Aware Neural Network
2- Proposal-Free Instance Segmentation
Results
1- Dataset
2- Networks Setup
3-Results and comparative analysis
Dual-head network for detection and segmentation
Triple-head network
Comparative Analysis
Introduction
Related works
Chapter4: Dealing with limited labels
Introduction
Related Works
Representaional learning for mOSCC
What is mOSCC
Dataset used in this study
methods
semi-supervised-GAN
Experiments
different budget of annotations
feature visualization
Different GANs
Self-supervised semi supervised
Experiments
different budget of nnotation
Multi-task setup
Domain Adaptation
GAN
self-supervision
Multi-task
Chapter3: Interactive Segmentation
Introduction
Related Works
Methods
Architecture
Guiding signals
Experimnets
Dataset
Implementation Details
Metrics
Network Selection
Validation Experiments
Discussions
Generalization study
Domain adaptation study
Segmentation Reliability Study
Sensitivity to Guiding Signals
Extreme Cases
User Correction
Chapter1: Introduction
Digital Pathology
Deep learning in Digital pathology
Challenges in DP
Aims and Objectives
Main contributions
Thesis Organization
Chapter 5: Conclusion and Future direction
Appendix
Monuseg Challenge
Chapter 5: Predicting patient outcome (Lung)