A
Andrew H. Gee
Researcher at University of Cambridge
Publications - 213
Citations - 7557
Andrew H. Gee is an academic researcher from University of Cambridge. The author has contributed to research in topics: 3D ultrasound & Artificial neural network. The author has an hindex of 48, co-authored 204 publications receiving 7081 citations.
Papers
More filters
Journal ArticleDOI
Rapid calibration for 3-D freehand ultrasound
TL;DR: This paper describes a new calibration technique that takes only a few minutes to perform and produces results that compare favourably (in terms of both accuracy and precision) with previously published alternatives.
Journal ArticleDOI
Regularised marching tetrahedra: improved iso-surface extraction
TL;DR: A new algorithm, regularised marching tetrahedra (RMT), is presented, which combines marching tetahedra and vertex clustering to generate iso-surfaces which are topologically consistent with the data and contain a number of triangles appropriate to the sampling resolution with significantly improved aspect ratios.
Journal ArticleDOI
Determining the gaze of faces in images
Andrew H. Gee,Roberto Cipolla +1 more
TL;DR: A more flexible vision-based approach, which can estimate the direction of gaze from a single, monocular view of a face, which makes minimal assumptions about the structure of the face, requires lew image measurements, and produces an accurate estimate of the facial orientation.
Journal ArticleDOI
Sequential Monte Carlo Methods to Train Neural Network Models
TL;DR: A novel strategy for training neural networks using sequential Monte Carlo algorithms is discussed and a new hybrid gradient descent/sampling importance resampling algorithm (HySIR) is proposed, which outperforms extended Kalman filter training on several problems.
Journal ArticleDOI
High resolution cortical bone thickness measurement from clinical CT data
TL;DR: A novel technique is presented that is capable of producing unbiased thickness estimates down to 0.3 mm in the clinically relevant sub-millimetre range, where thresholding increasingly fails to detect the cortex at all, whereas the new technique continues to perform well.