Benjamin El-Zein: Data Augmentation for Deep Learning based Collimator Detection

In digital radiography, only diagnostic relevant areas of the image should be presented to the radiologists. Since cropping algorithms yet have trouble with detecting corner points or edges robustly, a deep learning approach is investigated. To generate training data for this application, an image processing pipeline is presented that simulates real collimator characteristics.


24 Mar 2023


10:00 am - 10:15 pm

Local Time

  • Timezone: America/New_York
  • Date: 24 Mar 2023
  • Time: 5:00 am - 5:15 pm