So I have finally submitted and successfully defended my PhD thesis (given below).
In it I have organised the results on how to achieve visual recognition of human rights violations using modern computer vision methods.
Hopefully it can now be seen as a complete body of work, accessible to as large an audience as possible, and also acting as an introduction to the notion of what people will refer to as Visual Recognition of Human Rights Violations.
The original thesis document can be obtained as a single PDF (221.7 MB) for best quality, as a compressed single PDF (20.3 MB), or as individual chapters (since the single files are fairly large):
- Contents (PDF, 34 KB)
- Chapter 1: Computer Vision in the Context of Human Rights (PDF, 698 KB)
- Chapter 2: Human Rights Technology (PDF, 1.9 MB)
- Chapter 3: Datasets (PDF, 10.9 MB)
- Chapter 4: Predicting Human Rights Violations from Images: (PDF, 1.56 MB)
- Chapter 5: Objects and Scenes: Combining Features for HRVR (PDF, 1.1 MB)
- Chapter 6: DisplaceNet (PDF, 1.6 MB)
- Chapter 7: GET-AID (PDF, 2.8 MB)
- Bibliography (PDF, 79 KB)
If you intend to use my code in your research or wish to refer to the results presented there, please use this BibTeX entry.
Wordcloud derived from PhD thesis Abstract