Nonuniform Sampling for Bag-of-Features Classification in Melanoma Detection

Rui Hu, Ning Situ, Tarun Wadhawan, and George Zouridakis


Dermoscopic image classification, Bag-of-Features, Melanoma detection, Saliency


Image sampling is a critical component when bag-of-features methods are used for image classification. In this paper, we propose a nonuniform sampling strategy based on image patch saliency and pixel intensity. More patches are sampled from informative regions that contain more dermoscopically interesting features, while fewer patches are sampled from homogenous regions that provide only complementary information. The performance of the new method is evaluated on a dataset of 645 images representing pigmented skin lesions of known pathology. Experimental results show that the proposed method outperforms other sampling strategies in melanoma detection.

Important Links:

Go Back