SPD Matrices Reimagined: A Novel Approach to Computer Vision

Collective Intelligence, Optimization and Control

Type of work flexible, Master thesis, Bachelor thesis

Our lab has previously developed a novel SPD layer for causal learning, and now we aim to adapt this innovation to the field of computer vision.

The goal of this master thesis is to adapt and explore this layer for applications in computer vision, integrating it within the broader context of SPD networks. Specifically, the project will focus on:

• Implementation and Integration: Integrating the new layer into anexisting SPD network framework, adjusting parameters for optimal performance,and adapting the model for causal learning to computer vision.

• Evaluation and Testing: Applying the enhanced SPD network on selected computer vision tasks to assess the effectiveness and efficiency of the newlayer.

• Comparison with Existing Models: Analyzing the performance in comparison with existing methods.

• Scientific Contribution: Documenting the findings, presenting the results,and possibly contributing to a published paper in the field.