Laboratorio: computer vision for digital humanities
A.A. 2024/2025
Obiettivi formativi
This workshop aims to provide an introduction to deep learning-based computer vision aimed for humanities applications. In particular, it focuses on providing a high-level overview of machine learning based approaches to computer vision focusing on supervised learning. The workshop includes discussions on working with humanities data, and their hands-on application.
Risultati apprendimento attesi
By the end of this workshop, students will be able to:
1. Understand Key Concepts: Describe the goals of various computer vision tasks, such as image classification and segmentation.
2. Explain Principles of Computer Vision: Articulate the key principles of computer vision, including image processing, feature extraction, and machine learning techniques.
3. Utilize Computer Vision Tools: Apply computer vision tools (e.g., Transkribus) to process and analyze visual data in the context of historical and humanities research.
Students will have access to course materials and supplemental resources on the myAriel platform, as well as the opportunity to consult with the instructor during office hours or via email.
1. Understand Key Concepts: Describe the goals of various computer vision tasks, such as image classification and segmentation.
2. Explain Principles of Computer Vision: Articulate the key principles of computer vision, including image processing, feature extraction, and machine learning techniques.
3. Utilize Computer Vision Tools: Apply computer vision tools (e.g., Transkribus) to process and analyze visual data in the context of historical and humanities research.
Students will have access to course materials and supplemental resources on the myAriel platform, as well as the opportunity to consult with the instructor during office hours or via email.
Periodo: Secondo semestre
Modalità di valutazione: Giudizio di approvazione
Giudizio di valutazione: superato/non superato
Corso singolo
Questo insegnamento può essere seguito come corso singolo.
Programma e organizzazione didattica
Edizione unica
Responsabile
Periodo
Secondo semestre
Programma
Introduction to Computer Vision: This section will cover the foundational concepts of computer vision, including supervised learning, and key tasks like image recognition, segmentation, and detection. Students will explore the relevance of these techniques in humanities research, learning how they can be applied to analyze historical materials. The course will introduce various tools that enable hands-on applications, such as those for text recognition and object detection.
Automatic Text Recognition Using Computer Vision: Students will gain practical experience in preparing historical manuscript images for transcription. They will learn how to use automatic segmentation tools to break images into text lines, words, and characters. Through training models on historical documents, students will refine transcription accuracy with manual adjustments and enhance results.
Object Detection in Historical Datasets: Students will explore object detection techniques applied to historical image datasets. They will annotate images by drawing bounding boxes around objects, letters, and symbols, and apply relevant methods to automatically detect and classify these elements.
The course will explore how these techniques can be applied to various types of historical artifacts, such as Greek squeezes, ancient coin images, and historical handwritten archival texts.
Automatic Text Recognition Using Computer Vision: Students will gain practical experience in preparing historical manuscript images for transcription. They will learn how to use automatic segmentation tools to break images into text lines, words, and characters. Through training models on historical documents, students will refine transcription accuracy with manual adjustments and enhance results.
Object Detection in Historical Datasets: Students will explore object detection techniques applied to historical image datasets. They will annotate images by drawing bounding boxes around objects, letters, and symbols, and apply relevant methods to automatically detect and classify these elements.
The course will explore how these techniques can be applied to various types of historical artifacts, such as Greek squeezes, ancient coin images, and historical handwritten archival texts.
Prerequisiti
Familiarity with using computers and basic software applications. No prior knowledge of programming or computer science is required, but an interest in exploring digital tools and technologies in humanities research is recommended.
To apply for admission to the workshop, it is mandatory to follow the instructions on the webpage: https://scienzestoriche.cdl.unimi.it/it/insegnamenti/laboratori
To apply for admission to the workshop, it is mandatory to follow the instructions on the webpage: https://scienzestoriche.cdl.unimi.it/it/insegnamenti/laboratori
Metodi didattici
This course takes a hands-on, interactive approach, where students will apply computer vision techniques to analyze historical materials. Active participation, both individually and collaboratively, is required throughout the course.
Class attendance is mandatory.
Class attendance is mandatory.
Materiale di riferimento
- Slides, handouts, and scholarly articles provided by the instructor.
- Books on computer vision to support learning (these are not a prerequisite, just references):
o Foundations of Computer Vision by Antonio Torralba, Phillip Isola, and William T. Freeman (MIT Press, 2024): A comprehensive introduction to core computer vision concepts, suitable for those new to the field. (Online https://mitpress.ublish.com/ebook/foundations-of-computer-vision-1/12791/418)
o Deep Learning by Ian Goodfellow, Yoshua Bengio, and Aaron Courville (Adaptive Computation and Machine Learning series): A more advanced text focused on deep learning techniques, ideal for students seeking in-depth understanding of the algorithms behind computer vision. (online: https://www.deeplearningbook.org/)
o Biological and Computer Vision by Gabriel Kreiman (Cambridge University Press, 2021): Explores both biological and computational vision, providing useful insights into vision systems and their application in computer vision, accessible for students with a basic understanding of the subject. (online: https://klab.tch.harvard.edu/publications/Books/BiologicalAndComputerVision/TableOfContents.html)
o Artificial Intelligence, A Guide for Thinking Humans by Melanie Mitchell (2019): Provides a clear and accessible overview of AI concepts, challenges, and the history of the field. It is a helpful resource for students seeking to understand the broader context of AI and its implications beyond computer vision (https://melaniemitchell.me/aibook/)
o Pixels & Paintings: Foundations of Computer-assisted Connoisseurship by David G. Stork (2023): A specialized resource connecting computer vision to art and heritage analysis, can be useful for students interested in art and cultural heritage applications.
- Transkribus: Offers valuable online resources for handwritten text recognition. The platform provides guides and tools for working with historical manuscript images, suitable for students engaged in transcription tasks: https://www.transkribus.org/blog
- Books on computer vision to support learning (these are not a prerequisite, just references):
o Foundations of Computer Vision by Antonio Torralba, Phillip Isola, and William T. Freeman (MIT Press, 2024): A comprehensive introduction to core computer vision concepts, suitable for those new to the field. (Online https://mitpress.ublish.com/ebook/foundations-of-computer-vision-1/12791/418)
o Deep Learning by Ian Goodfellow, Yoshua Bengio, and Aaron Courville (Adaptive Computation and Machine Learning series): A more advanced text focused on deep learning techniques, ideal for students seeking in-depth understanding of the algorithms behind computer vision. (online: https://www.deeplearningbook.org/)
o Biological and Computer Vision by Gabriel Kreiman (Cambridge University Press, 2021): Explores both biological and computational vision, providing useful insights into vision systems and their application in computer vision, accessible for students with a basic understanding of the subject. (online: https://klab.tch.harvard.edu/publications/Books/BiologicalAndComputerVision/TableOfContents.html)
o Artificial Intelligence, A Guide for Thinking Humans by Melanie Mitchell (2019): Provides a clear and accessible overview of AI concepts, challenges, and the history of the field. It is a helpful resource for students seeking to understand the broader context of AI and its implications beyond computer vision (https://melaniemitchell.me/aibook/)
o Pixels & Paintings: Foundations of Computer-assisted Connoisseurship by David G. Stork (2023): A specialized resource connecting computer vision to art and heritage analysis, can be useful for students interested in art and cultural heritage applications.
- Transkribus: Offers valuable online resources for handwritten text recognition. The platform provides guides and tools for working with historical manuscript images, suitable for students engaged in transcription tasks: https://www.transkribus.org/blog
Modalità di verifica dell’apprendimento e criteri di valutazione
Assesment Method: Evaluation at the end of the course.
Type of Examination: In addition to attending lessons, students will complete a final assessment, which includes an oral exam and a presentation of a project to be agreed upon with the instructor. Detailed project guidelines will be provided during the course.
Evaluation Criteria: Ability to demonstrate and discuss key concepts; critical reflection on the completed work; clarity of communication; proficiency in using relevant terminology; efficient use of the computational tools introduced in the course; overall effectiveness of the presentation.
Type of evaluation method: approval of 3 CFUs.
Assessment result: approved/not approved.
The format of the assessment for students with disabilities should be arranged in advance with the lecturer.
Type of Examination: In addition to attending lessons, students will complete a final assessment, which includes an oral exam and a presentation of a project to be agreed upon with the instructor. Detailed project guidelines will be provided during the course.
Evaluation Criteria: Ability to demonstrate and discuss key concepts; critical reflection on the completed work; clarity of communication; proficiency in using relevant terminology; efficient use of the computational tools introduced in the course; overall effectiveness of the presentation.
Type of evaluation method: approval of 3 CFUs.
Assessment result: approved/not approved.
The format of the assessment for students with disabilities should be arranged in advance with the lecturer.
Docente/i