Skip to content

README.md

Latest commit

 

History

History
61 lines (45 loc) · 4.68 KB

README.md

File metadata and controls

61 lines (45 loc) · 4.68 KB

Thesis Topics for 2025

Welcome to the Master's Thesis Topics (2025) repository at Casa Paganini - InfoMus. Here, we share topics at the intersection of art, science, and technology, that can broadly be categorized as:

  • Explorative Research: Developing new computational models for interpreting non-verbal communication or exploring uncharted domains.
  • Applicative Research: Applying methods to solve real-world problems, such as interactive multimedia installations, adaptive soundscapes, and ML-driven tools for music therapy, rehabilitation, and education.

Take the Next Step

Pick a topic, refine it, or propose your own.

  • Chosen a topic you like? Contact Antonio Camurri (antoniocamurri@gmail.com) or Gualtiero Volpe (gualtiero.volpe@unige.it) to discuss your ideas and next steps.
  • Getting started: Select a topic that interests you and dive into it.
  • Got an idea we missed? Share it. We value fresh perspectives and are open to evaluating your proposals.
  • Need help? Reach out. We're here to guide you in putting together a thesis that aligns with your academic objectives.

If you propose a new topic, ensure it aligns with our lab's domain. Provide a clear summary, your intended approach, and potential applications. We will review and provide feedback to help shape it into a compelling thesis.

Proposed Topics

1. Quantifying Non-Verbal Communication Using Gaze Detection Algorithms

  • Macro-category: Computational Models, Non-verbal Communication
  • Summary: Gaze patterns provide insights into attention, engagement, and emotional states. Leveraging novel techniques and models (e.g., Moondream 2B) can help deepen our understanding of human interactions.
  • Applications: Social robotics, education platforms, therapy tools, and adaptive interactive systems.

2. Posture Quality Quantification Using Conformal Predictions

  • Macro-category: Rehabilitation
  • Summary: Posture impacts health and ergonomics, yet assessment methods often lack robustness. Conformal prediction offers uncertainty quantification, enabling reliable feedback.
  • Applications: Physiotherapy tools, workplace health monitoring, fitness training, and ergonomic software.

3. Automated audio/video semantic segmentation using Vision Language Models

  • Macro-category: Saliency
  • Summary: With growing digital content, efficient summarization techniques are helpful. Combining audio processing, signal analysis, and text understanding helps extract salient, relevant, and relevant audio-video segments.
  • Applications: Social media content, podcast highlights, educational material, and marketing campaigns.

4. Motion-Aware Animation Framework for Embodied Environments Using Three.js

  • Macro-category: Computational Models, Embodied Systems
  • Summary: Real-time motion-based animations enhance immersive experiences by linking movement features to dynamic visuals. These systems engage users through interactive feedback.
  • Applications: Museum exhibits, rehabilitation programs, therapeutic environments, and immersive art installations.

5. Class-Conditioned VAE for Trajectory Generation

  • Macro-category: Generative Models, Human Motion Synthesis
  • Summary: Human motion generation plays a vital role in virtual reality, robotics, and animation. This study focuses on developing a class-conditioned Variational Autoencoder (VAE) to synthesize expressive 3D motion sequences by leveraging novel conditioning techniques.
  • Applications: Virtual reality avatars, robotics navigation, animation tools, and interactive systems emphasizing expressive motion synthesis.

6. Video-Language Model for Expressive Qualities of Movement Annotations

  • Macro-category: Video Annotation
  • Summary: This thesis explores using Video-Language Models (Vid-LLMs) to enhance video annotation of expressive movement qualities and motion segmentation. It investigates state-of-the-art Vid-LLMs, their capabilities, and the feasibility of fine-tuning them on a small CP dataset.
  • Applications: Video editing, motion analysis, choreography annotation, and multimodal interaction systems.

Contribution Guidelines

Contributions and discussions are welcome. Ensure your proposals align with the lab's objectives to integrate technology and art. For queries or feedback, contact us at antoniocamurri@gmail.com or gualtiero.volpe@unige.it.

  • To propose new topics or extend existing ones, create a pull request.
  • Include a detailed description of your research idea, including its summary, approach, and potential applications.

Contact

For more information about Casa Paganini - InfoMus, visit our website.