Course content

This course provides an introduction to artificial intelligence techniques used to analyse, interpret, and generate visual data. It covers core concepts in deep learning for images, video and 3D data, together with methods for building reliable and efficient AI pipelines for visual informatics.

Topics include:

• Deep learning architectures for visual tasks, such as CNNs and attention-based models
• Generative and reconstruction methods for visual content
• AI techniques for 3D, spectral and multimodal data
• Robustness, uncertainty, and basic explainability methods
• Ethical considerations and responsible use of AI in imaging

Students develop practical skills through small projects using modern AI frameworks.

Learning outcome

Knowledge

The candidate will be able to:

• Explain key AI methods used in visual informatics and how visual data are represented in modern models.
• Describe common approaches to deep learning, generative methods and multimodal analysis.
• Discuss limitations, error sources, and ethical considerations in AI-based visual systems.

Skills

The candidate will be able to:

• Implement and evaluate AI models for selected visual computing tasks.
• Prepare and process visual datasets in reproducible workflows.
• Use contemporary machine learning frameworks to train, analyse and compare models.

General competence

The candidate will:

• Critically assess the suitability and reliability of AI methods in different visual applications.
• Communicate findings clearly to both specialist and non-specialist audiences.
• Apply responsible practices when developing and deploying AI systems.

Learning methods and activities

Lectures, practical exercises and small projects. Some sessions may include group discussions or guest lectures.

Specific conditions