På svenska

Presentation Information     2017-04-24 (14:15)   •  4307

Speaker Damian Matuszewski  (CBA)
Title Investigating deep segmentation for viruses in TEM images
Abstract Gustaf Kylberg, a former PhD-student of Ida-Maria Sintorn, created a unique dataset containing over 1300 Transmission Electron Microscopy images with 22 different types of viruses. He also addressed the segmentation and classification of this challenging image collection using “classical” image analysis methods. In the days of the Deep Learning glory I intent to continue his legacy and investigate the potential application of CNNs to this dataset. During the seminar I will present the dataset, its challenges and Deep Learning methods I plan to use to improve the automatic virus segmentation and classification.