Patch Selection for Melanoma Classification
This work addresses patch selection in medical image processing for melanoma diagnosis, but it is incremental as it compares existing criteria without introducing a new method.
The paper tackled the problem of automatically selecting relevant patches for melanoma classification by comparing entropy and spectral similarity criteria, finding that entropy-based selection leads to faster convergence and higher accuracy.
In medical image processing, the most important information is often located on small parts of the image. Patch-based approaches aim at using only the most relevant parts of the image. Finding ways to automatically select the patches is a challenge. In this paper, we investigate two criteria to choose patches: entropy and a spectral similarity criterion. We perform experiments at different levels of patch size. We train a Convolutional Neural Network on the subsets of patches and analyze the training time. We find that, in addition to requiring less preprocessing time, the classifiers trained on the datasets of patches selected based on entropy converge faster than on those selected based on the spectral similarity criterion and, furthermore, lead to higher accuracy. Moreover, patches of high entropy lead to faster convergence and better accuracy than patches of low entropy.