2017/01/15 by Aditya Shastri, Shastri, Aditya A., Deepti Tamrakar +3 · 1 citation
Biochemistry, Genetics and Molecular Biology · Computer Science · #AI in cancer detection #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Gene expression and cancer classification #Image Retrieval and Classification Techniques
paper · pdf · doi:10.48550/arxiv.1701.04010
openalex publication_date 2017/01/15 · openalex created_date 2022/10/03 · openalex updated_date 2026/07/28
Breast cancer is becoming pervasive with each passing day. Hence, its early\ndetection is a big step in saving the life of any patient. Mammography is a\ncommon tool in breast cancer diagnosis. The most important step here is\nclassification of mammogram patches as normal-abnormal and benign-malignant.\n Texture of a breast in a mammogram patch plays a significant role in these\nclassifications. We propose a variation of Histogram of Gradients (HOG) and\nGabor filter combination called Histogram of Oriented Texture (HOT) that\nexploits this fact. We also revisit the Pass Band - Discrete Cosine Transform\n(PB-DCT) descriptor that captures texture information well. All features of a\nmammogram patch may not be useful. Hence, we apply a feature selection\ntechnique called Discrimination Potentiality (DP). Our resulting descriptors,\nDP-HOT and DP-PB-DCT, are compared with the standard descriptors.\n Density of a mammogram patch is important for classification, and has not\nbeen studied exhaustively. The Image Retrieval in Medical Application (IRMA)\ndatabase from RWTH Aachen, Germany is a standard database that provides\nmammogram patches, and most researchers have tested their frameworks only on a\nsubset of patches from this database. We apply our two new descriptors on all\nimages of the IRMA database for density wise classification, and compare with\nthe standard descriptors. We achieve higher accuracy than all of the existing\nstandard descriptors (more than 92%).\n