search

antonioguj / bronchinet

Airway segmentation from chest CTs using deep Convolutional Neural Networks
MIT License
62 stars 26 forks source link

Ask if the framework needs to prepare its own ground-truth for CT image preprocessing？ #1

Closed harzn closed 4 years ago

harzn commented 5 years ago

Thank you very much for the code you provided！I'm a rookie of deep learning.So i ask if the framework needs to prepare its own ground-truth for CT image preprocessing？Or the preprocessing module in the framework can output results with ground-truth？Can I use this framework directly without ground-truth CT images?

antonioguj commented 5 years ago

Hi Harzn, I'm sorry for the late reply. I did not receive notification of your inquiry. So this framework assumes you have available your own ground-truth, whether it's manually or with another software. The preprocessing modules only convert these data to the format used in the training algorithms, and basic processing such as cropping, rescaling... Moreover, this framework is only for supervised training, i.e. when you provide both the images and the ground-truth. Please do not hesitate to ask any other question.