(It is a topic for the seminar "Advanced Algorithms in Medical Image Processing" in my study matrix)

Introduction:

The number of mitotic figures visible in histology sections is an important indicator for cancer screening and assessment the counting is performed manually by pathologists, histologists (This process is cumbersome) Mitotic cell may look alike, and this will make assessments more difficult A low density of mitosis makes it more labor intensive for a pathologist. differences in staining and tissue artifacts complicate this task.

Furthermore, mitotic detection manually often cost 24-96h and this period will be painful for the patients

Because of its importance in determining the severity of the cancer, the development of automated mitosis detection has become necessary and imperative

Our game is to 1) train the players with labeld mitosis images 2) let the trained players label the unlabeld mitosis images and get the final label with a certain aggregation method.

During the seminar, we finished the first part.

Screenshots: