Jack Thorne
Thesis title: Superpixel-based Compression of Spatially Resolved Chemical Images
Franklin supervisor(s): Dr Mark Basham & Dr Felicia Green
University: University of Nottingham
University supervisor: Prof. Andrew French
Chemical imaging modalities such as Matrix-Assisted Laser Desorption/Ionisation (MALDI) or Scanning Transmission X-ray Microscopy (STXM) produce data-dense ‘datacubes.’ Every pixel contains discrete chemical information (e.g. mass spectra), each comprising thousands of data points. Consequently, traditional computer vision methods struggle; without semantic object structure, standard machine learning-based segmentation approaches fail. Superpixels overcome this by aggregating spectrally similar pixels using a combination of chemical and spatial information, reducing data volume several-fold whilst preserving image content.
Jack holds a Master of Science by Research from the University of Bristol, specialising in computer vision tool development for the Verkade group. Prior to this, whilst in the Baldock group, Jack’s work focused on real-time computer vision research to aid those who are colour-blind and genetic informatics for disease prediction. Despite working primarily on computer vision, Jack considers himself a classically trained biologist, holding a BSc (Hons) in Biology from the University of Gloucestershire, where he studied a mix of molecular biology, ecology, and biogeography.
No publications found.