Abstract: The large volume, enormous medical information, and stain variation are depicted as pathology image attributes; correspondingly, the products in the deep learning approaches, e.g., the heavy annotation workload, the diverse labeling strategy, the stain normalization unreliability, and the various artifacts, have hindered the diagnoses of pathology image. This talk presents some practical difficulties with deep learning techniques and our state-of-the-art resolutions to the issues.
Venue: The seminar will be held between 1-2pm on Wednesday 26th July at MR 12 , Centre for mathematical Sciences, CB3 0WA . Alternatively you can also join online via ZOOM : https://maths-cam-ac-uk.zoom.us/j/93331132587?pwd=MlpReFY3MVpyVThlSi85TmUzdTJxdz09.
Biography: Jing Ke, Ph.D. of the University of New South Wales, is an assistant professor at Shanghai Jiao Tong University, and an adjunct lecturer at the University of New South Wales. She has many years of commercial experience in Advanced Micro Devices, Inc. (AMD) and Commonwealth Scientific and Industrial Research Organization (CSIRO). Her research interest is in computational pathology and parallel computing of large-scale images. She has 20+ publications as the first/last author in medical journals and conferences, e.g., IEEE Trans on Medical Imaging Analysis, Medical Image Analysis, and MICCAI . She has 10 conference papers in computer architecture and high-performance computing, e.g., MICRO , DAC, and IPDPS , as well as a textbook in GPGPU programming. She has received a grant of the National Natural Science Foundation of China, a grant of the Natural Science Foundation of Shanghai, and two Shanghai Jiao Tong University Medical Engineering Cross Fund. She is a topic editor of Frontiers in Radiology and a reviewer of TPAMI , TMI, and MICCAI .
