Handcrafted features vs deep-learned features: Hermite Polynomial Classification of Liver Images

Abstract

Liver cancer is one of the most common types of cancer according to World Health Statistics. Computer-aided diagnosis (CAD) systems are used in medical imaging for liver tumor identification and classification. Texture is a type of feature that can provide measurements of properties such as smoothness and regularity of the image. Handcraft techniques based on fractal geometry allow quantifying self-similarity properties present in images. However, new studies have shown that using information obtained from deep-learned feature maps can maximize the results of classical classifiers. This work presents an approach that investigates descriptors obtained by handcrafted and deep learning features, feature selection methods and the Hermite polynomial (HP) algorithm to classifier liver histological images. The results were evaluated using metrics such as accuracy (ACC) and the imbalance accuracy metric (IAM). The association with fractal features, Lasso regularization and the HP algorithm achieved 0.98 of IAM and 99.53% ACC, which was relevant when evaluated with other studies in the literature.