Teaching Units for a Computer Vision Course

KIRN, VASJA LEV (author), Peer, Peter (mentor), Emeršič, Žiga (co-mentor)

Rapidly advancing development of artificial intelligence (AI) technologies, including deep learning techniques in the field of computer vision, has encouraged the need for early education about artificial intelligence in schools. This thesis details the development of a computer vision (CV) curriculum, part of the AIM@VET (Artificial Intelligence Modules for Vocational Education and Training) project, targeting VET high-school students. The thesis is structured into three main teaching units (TUs): fundamentals of object detection, deep learning models for object detection, and fundamentals of image segmentation. Each TU consists of eight tasks and a final assignment, totaling 30 hours of classroom work. The course material, designed in Python notebooks, combines theoretical concepts with practical coding exercises. Unique versions for teachers and students facilitate effective learning and teaching, even for those unfamiliar with the topics. This approach to digital education in CV leverages interactive tools and open-source libraries like OpenCV, facilitating hands-on learning and immediate application of CV concepts. Also discussed are instructional design, content selection, and the initial evaluation of feedback, emphasizing the evolving need for digital education in the field of AI and CV.

• Language: English
• Keywords: Computer vision, Python notebook, object detection, segmentation
• Work type: Bachelor thesis/paper
• Typology: 2.11 - Undergraduate Thesis
• Organization: FRI - Faculty of Computer and Information Science
• Year: 2024
• PID: 20.500.12556/RUL-153473
• COBISS.SI-ID: 181849603
• Publication date in RUL: 09.01.2024

Secondary language

• Language: Slovenian
• Title: Učne enote za predmet računalniškega vida

Abstract

Hitro napredujoč razvoj tehnologij umetne inteligence (AI), ki vključuje tehnike globokega učenja na področju računalniškega vida, je spodbudil potrebo po zgodnjem izobraževanju o umetni inteligenci v šolah. V okviru projekta AIM@VET (Artificial Intelligence Modules for Vocational Education and Training), ki je namenjen dijakom strokovnih srednjih šol, se razvija učne enote za predmet računalniški vid (RV). V diplomski nalogi so bila razvita gradiva za tri glavne učne enote (UE): detekcija objektov, globoki modeli za detekcijo objektov in segmentacija slik. Vsaka UE vsebuje osem nalog in se zaključi s samostojnim projektom v dveh delih, skupaj je razvitega gradiva za 30 ur pouka. Dve različici gradiv, za učitelje in dijake, omogočata učinkovito poučevanje in učenje, tudi za tiste, ki niso seznanjeni s tematiko. Naloge, strukturirane v Python notebook, združujejo teoretične koncepte s praktičnimi vajami programiranja. Uporaba interaktivnih orodij in odprtokodnih knjižnic, kot je OpenCV, omogoča praktično učenje ter takojšen preiskus usvojenih konceptov. V diplomski nalogi so obravnavani načini poučevanja, izbori vsebin in evalvacija glede na prve povratne informacije, ki skupaj nakazujejo na naraščajočo potrebo po digitalnem izobraževanju na področju AI in RV.

• Keywords: računalniški vid, Python notebook, detekcija objektov, segmentacija