Scalable Multi-Instance Multi-Shape Support Vector Machine for Whole Slide Breast Histopathology

Hoon Seo, Lodewijk Brand, Lucia Saldana Barco, Hua Wang

ICKG - 2022

Histopathological image analysis is critical in cancer diagnosis and treatment. Due to the huge size of histopathological images, the most works analyze the whole slide pathological image (WSI) as a bag and its patches are instances. However these approaches are limited to analyze the patches in a fixed shape, while the malignant lesions can form a various shapes. We propose the Multi-Instance Multi-Shape Support Vector Machine (MIMSSVM) to analyze the multiple images (instances) jointly where each instance consists of multiple patches in the various shapes. In our approach, we can identify the various morphologic abnormalities of nuclei shapes from the multiple images. In addition to the multi-instance multi-shape learning capability, we provide an efficient algorithm to optimize the proposed model which scales well to the large number of features. Our experimental results show the proposed MIMSSVM method outperforms the existing SVM and deep learning models in histopathological classification. The proposed model also identifies the tissue segments in an image exhibiting an indication of an abnormality which provides the utility in the early detection of malignant tumor. The implementation codes of proposed model is provided at!AiFpD21bgf2wgSXMUpr6K7cJmqx1?e=VPpqLg