New paper on “Evaluation of automated airway morphological quantification for assessing fibrosing lung disease” to be presented at Deep Generative Models workshop @ MICCAI 2022. We present a completely labelless automated airway measurement method that learns to regress to an airway’s diameter. It trains fast from synthetic data generated using style transfer. We evaluate our method against mortality data demonstrating that it better predicts mortality than state-of-the-art.
Many thanks to my co-authors without whom this work would not have been possible Mou-Cheng Xu Tony Wing Keung Cheung Marie Vermant Tinne Goos Laurens De Sadeleer Stijn Verleden Wim Wuyts Professor John Hurst Joseph Jacob.
ALSO this is completely compatible with the open-source airway analysis framework, AirQuant!