====== TTK37 Visual Analytics and Automation ======

===== Computer Vision and AI for Multi-Disciplinary Applications in Aquaculture and Industry =====

\\
**Study points:** 3.75\\

**Instructor:** Christian Schellewald

===== Motivation =====

Recent advancements and the integration of computer vision and artificial intelligence (AI) are quickly transforming the industry by enabling automated, data-driven decision-making. In aquaculture, these technologies enable continuous monitoring of fish health, behavior, and environmental conditions, helping farms become more sustainable and efficient.

This course is organized as a research seminar focusing on methods and recent advances in computer vision and artificial intelligence. Students gain an overview of presented research by critically reading, presenting, and discussing recent and influential research papers. The course emphasizes practical and applicable methods, enabling students to critically evaluate scientific work and understand how modern AI techniques may be applied to solve real-world problems.

While several examples may originate from aquaculture, the methods and insights are broadly transferable to robotics, industrial inspection, autonomous systems, and other application domains.

===== Learning Outcomes =====

By the end of this course, students will:
  * Understand recent advances in computer vision and artificial intelligence.
  * Critically read, evaluate, and discuss scientific literature.
  * Present and communicate scientific work in a clear and structured manner.
  * Assess the strengths, limitations, and applicability of modern artificial intelligence methods.
  * Relate current research to practical applications in aquaculture, robotics, industrial inspection, autonomous systems, and related fields.
  * Gain introductory experience with modern AI tools and open-source software where appropriate.
===== Seminar Structure =====

The course is organized as a research seminar centered around student presentations and discussions of recent scientific literature.

Students present selected research papers related to methods and recent advances in computer vision and artificial intelligence. Papers are chosen in consultation with the instructor and are typically aligned with the students' research interests or master projects.

Students are expected to actively participate in discussions and critically evaluate presented work.

Depending on the selected topics, short introductions, demonstrations, or practical sessions may be included to provide background on relevant concepts, methods, and software tools.

Topics may include:
  * Computer vision and visual perception
  * Machine learning and artificial intelligence
  * Scene understanding and visual reasoning
  * Detection, segmentation, and tracking
  * Foundation models and multimodal AI
  * AI for robotics and autonomous systems
  * Embedded and efficient AI
  * Reproducible research and experimental methodology
  * Applications in aquaculture, robotics, industrial inspection, and autonomous systems

Since the topics are based on current research and/or selected papers, the exact content varies from year to year.

===== Assessment =====

Assessment is based on an individual oral examination. Students are examined on their own presented paper together with selected papers discussed during the seminar.

===== Prerequisites =====

A background in computer vision, signal processing, machine learning, robotics, cybernetics, computer science, or related fields is beneficial. Familiarity with programming languages such as Python, C++, or MATLAB is advantageous but not required. Students should be willing to actively participate in scientific discussions and independently study research papers.