LABORATORY OF COMPUTER VISION FOR DIGITAL HUMANITIES

The "Laboratory of Computer Vision for Digital Humanities" course is designed to bridge the gap between cutting-edge computer vision and machine learning technologies and the interdisciplinary domain of Digital Humanities. In this era of vast digital collections and complex data, the course aims to equip Digital Humanities MSc students with the knowledge and practical skills necessary to harness the power of computer vision for their research and analysis.

This course provides students with a solid foundational understanding of machine learning, emphasising its applications within the Digital Humanities, while also equipping them with fundamental image processing skills essential for efficient visual data handling. The primary objective of this course is to empower students to apply their knowledge to practical applications in the field of Digital Humanities. Specifically, students will learn the creation of ground truth for supervised learning, the implementation of computer vision algorithms, the interpretation and evaluation of their outcomes, and the effective preprocessing of data.

At the end of this course, students will gain a comprehensive understanding of both the technical aspects of the most common computer vision methods and their practical applications within the Digital Humanities domain by coding within hands-on applications. Moreover, they will also acquire skills in utilising off-the-shelf tools for various Digital Humanities applications.

Basic Python programming, basic statistics, familiarity with Jupyter Notebooks

The program is provisional and is subject to change. Its contents may include a selection from the following potential topics:

Technical content:
- Introduction to Machine Learning
- Introduction to Neural networks
- Image classification using Convolutional Neural Networks
- Detection of faces, or objects in images using neural networks
- Digitization of written sources (OCR & Handwritten Text Recognition)
- Machine learning for automatic text analysis: Text classification and authorship analysis
- Art generation from text prompts, inpainting, outpainting (if time allows)
- Introducing Off-the-shelf Tools for DH Research

Possible Lab practices:
- Preparation of a dataset and binary classification of it
- Style classification of paintings
- Face or object detection in art images
- Classifying ancient coins
- Using CNNs to study/classify digitised historical newspapers
- DNN for detection of texts in historical maps
- Page segmentation for historical document images
- Handwritten recognition from archive documents

Appunti delle lezioni e tutorial sulla visione artificiale reperibili su internet.

Due libri rilevanti:
- Pixels & Paintings: Foundations of Computer-assisted Connoisseurship, David G. Stork, 2023
- Computational Formalism, Amanda Wasielewski, 2023

Assessment methods will include written / or oral exams, assignments, and a group project to encourage students to apply their learning to real-world DH challenges.

Theoretical lessons describe the various concepts and methods; active learning and open discussions.

English

written and oral