Texture Description for Classification of Fine Needle Aspirates

In breast cancer diagnosis, fine needle aspiration biopsy is an important diagnostic tool. It is used to estimate cancer malignancy grade that is further required for treatment determination. In this paper we describe a scheme based on pattern recognition and image processing techniques for automatic breast cancer malignancy grading from cytological slides of fine needle aspiration biopsies. To determine a malignancy classification of the slidewe propose to extract textural features of nucleiwith an application of local binary patterns. Based on texture determination, we present an improved classification system for cancer malignancy grading.