Mammography is the gold standard for breast cancer detection. However, it has very
high false positive rates and is based on ionizing radiation. This has led to interest in
using multi-modal approaches. One modality is diagnostic ultrasound, which is based
on non-ionizing radiation and picks up many of the cancers that are generally missed
by mammography. However, the presence of speckle noise in ultrasound images has a
negative effect on image interpretation. Noise reduction, inconsistencies in capture
and segmentation of lesions still remain challenging open research problems in
The target of the proposed research is to enhance the state-of-art computer vision
algorithms used in ultrasound imaging and to investigate the role of computer
processed images in human diagnostic performance. [Continues.]
A Doctoral Thesis. Submitted in partial fulfilment of the requirements for the award of Doctor of Philosophy at Loughborough University.
Loughborough University, Department of Computer Science.