Title: Quantification of Lung Disease from High-resolution CT Imagery using Textural Features and Machine Learning

Dr. Abhir Bhalerao; Dept. Computer Science, University of Warwick.

About the talk: This talk will introduce the background to the problem of measuring the extent of fibrotic lung disease in high-resolution CT (HRCT) images. The severity of the disease can be seen as textural changes in the appearance of the lung tissue. The talk will present the image analysis and machine learning framework that have been built to allow for automatically locating the lungs region, and grade and quantify the disease across the data. The study has applied techniques from integral geometry: Minkowski functionals as texture features. The talk will explain the texture feature extraction process and pattern classification methods that have been used. Validation of the work is being carried out on a prospective study with 50 patients over 2 years, imaged at 4 time points. The talk will show some of the validation results obtained thus far on subjects, and how they correlate with traditional functional measurements (spirometry) of the lung.

Bio-sketch: Dr. Abhir Bhalerao is an Associate Professor in Computer Science at the University of Warwick. He received his Ph.D. in 1992. He joined the faculty at Warwick in 1998 having been a research scientist with the NHS and Kings Medical School, London, and a Research Fellow at Harvard Medical School (1996-97). His research interests include modelling chronic lung diseases using CT scan data, multi-camera reconstruction problems, and forensic image analysis. He has published about 70 peer reviewed articles in image analysis, medical imaging, graphics and computer vision. He has served as the co-chair of the British Machine Vision Conference, 2007 and Medical Image Understanding & Analysis, 2010, and is a reviewer for several IEEE transactions journals. He was a co-founder of Warwick Warp Ltd., a university spin out company specializing in biometric technologies. Since 2010 he has been a director of Pattern Analytics Research Ltd., which makes software for access control and forensics markets. He is co-author of several patents on imaging technologies. Others collaborating with him are Prof. Charles Hutchison (WMS/UCHW); and Drs. Emma Helm (UCHW), Elke Thonnes (Dept of Statistics), Felix Woodhead (UCHW), and David Parr (UCHW).

Registration: The seminar is open to all and a light lunch will be served from 12:00pm, for catering purposes please register your interest with sallyann.edwards@warwick.ac.uk