• Publications
  • Influence
UCID: an uncompressed color image database
A new dataset, UCID (pronounced "use it") - an Uncompressed Colour Image Dataset which tries to bridge the gap between standardised image databases and objective evaluation of image retrieval algorithms that operate in the compressed domain. Expand
Solving for Colour Constancy using a Constrained Dichromatic Reflection Model
This paper shows that a combination of physical and statistical knowledge leads to a surprisingly simple and powerful colour constancy algorithm, one that also works well for images of natural scenes and is significantly better than previous dichromatic algorithms. Expand
Lesion border detection in dermoscopy images
A systematic overview of the recent border detection methods in the literature paying particular attention to computational issues and evaluation aspects is presented. Expand
Illuminant and device invariant colour using histogram equalisation
This paper proposes a new colour invariant image representation based on an existing grey-scale image enhancement technique: histogram equalisation and applies the method to an image indexing application and shows that the method out performs all previous invariant representations. Expand
Anisotropic Mean Shift Based Fuzzy C-Means Segmentation of Dermoscopy Images
A new mean shift based fuzzy c-means algorithm that requires less computational time than previous techniques while providing good segmentation results is introduced. Expand
Convex and non-convex illuminant constraints for dichromatic colour constancy
This work shows how the intersection problem in the context of convex and non-convex illuminant constraints, based on the distribution of common light sources, can be solved and introduces a constraint on the colour of the illuminants. Expand
Colour indexing across devices and viewing conditions
Classification of melanocytic skin lesions from non-melanocytic lesions
A new tumor area extraction algorithm is developed to develop a pre-processor of an automated melanoma screening system and it is confirmed that this algorithm is capable of handling different dermoscopy images not only those of NoMSLs but also MSLs as well. Expand
Cost-sensitive decision tree ensembles for effective imbalanced classification
This paper introduces an effective ensemble of cost-sensitive decision trees for imbalanced classification, which is capable of outperforming other state-of-the-art algorithms and representing a useful and effective approach for dealing with imbalanced datasets. Expand
Colour and contrast enhancement for improved skin lesion segmentation
This paper considers the problems of poor contrast and lack of colour calibration which are often encountered when analysing dermoscopy images and presents an effective approach to improve the performance of lesion segmentation algorithms through a pre-processing step that enhances colour information and image contrast. Expand