Chapter 6: Classification Algorithms and Applications

Title: Hydatidiform mole analysis tool based on anomaly detection

Author(s): Patison Palee, Bernadette Sharp, Leonardo Antonio Noriega, Sebastien Pascal

Address: College of Arts Media and Technology, Chiang Mai University, Huaykeaw Rd., T. Suthep, A. Muang, 50200 Chiang Mai, Thailand and Faculty of Computing, Engineering and Technology, Staffordshire University, Beaconside, Stafford ST18 0AD, UK | Faculty of Computing, Engineering and Technology, Staffordshire University, Beaconside, Stafford ST18 0AD, UK | Faculty of Computing, Engineering and Technology, Staffordshire University, Beaconside, Stafford ST18 0AD, UK | Ecole Nationale Supérieure d’ingénieurs de Caen (ENSICAEN), Computer Science Section, 14000 Caen, France

Reference: Software, Knowledge, Information Management and Applications (SKIMA 2013) pp. 253 - 261

Abstract/Summary: Molar pregnancy, also known as hydatidiform mole, can be divided into partial and complete mole. These moles are caused by abnormal fertilisation. The incidence rate of molar pregnancy in some countries in Asia such as Japan, India and Thailand is higher than North America, Australia, New Zealand, and Europe. Morphological and histopathological examinations are commonly applied to identify hydatidiform moles. The histopathological examinations of molar pregnancy remain a challenge to expert pathologists. In this paper, we describe a tool developed to support the pathologists in their examination, referred herein as the hydatidiform mole analysis tool (HYMAT). The tool is based on anomaly detection approach aimed at analysing the key anomalous morphological features of villi such as red blood cells, stroma and trophoblast and villi size.

Order a copy of this article