Our goal is to understand and extend how humans and nature analyze and synthesize visual and geometric data. We utilize a range of techniques from computer vision, graphics, machine learning, and artificial intelligence.
Core lab is a research group in the Department of Computer Science and Technology at the University of Cambridge.
We develop techniques for representing and processing 3D scenes in the forms of geometry, appearance, and motion. We are in particular interested in high-quality 3D asset creation and capture, neural object and scene representations, and multi-model representations with natural language.
We believe one of the best uses of machine learning and artificial intelligence is assisting people in creating and building new things. We design algorithms and systems for novel digital tools. The applications range from 3D asset capture & creation to interactive maps.
Going beyond inventing new digital tools, we aim at a comprehensive understanding of how and why our algorithms and systems work the way they do when processing visual data. This is not optional: in order to have unbiased and fair algorithms, and thorough and rigorous analysis of algorithms, we need to understand what we do. We work on many different flavors of explainable visual AI from interpretable-systems-by-construction to game-theoretical concepts.
Digital understanding of organic structures has numerous applications in agriculture from carbon emission estimation to plant health. We work on the core technologies of 3D sensing and semantic understanding for plant and field analysis.
