INNOVATIVE CONVERGENCE TO THE WORLD

Al for radio-medicine

Home   >   Research   >   Al for radio-medicine

■ Students involved

  Jimin Lee (ljm861@gmail.com) 

  Hyungjoo Cho (phelahab@gmail.com)

 

 

■ Description

  Machine Learning (including Deep Learning) techniques have played an important role in many areas such as computer vision, natural language processing, etc. In addition, Deep Learning also shows great performance in medical image analysis problems.

 

  In our research group, we applied Machine Learning technique into prediction of MLC (Multi-Leaf Collimator) positional error and had a great result. For further analysis, we are going to apply Deep Learning technique into same problem and compare those two results. Also we are going to apply Deep Learning technique to various medical image processing such as metal artifact reduction, image resolution enhancement, etc.

 

  We recently participated in AAPM prostatex-2 challenge (SPIE-AAPM-NCI Prostate MR Gleason Grade Group Challenge) and won the 8th place among 143 teams.


https://www.aapm.org/GrandChallenge/PROSTATEx-2/default.asp 


  The goal of this challenge is to predict the GGG(Gleason Grade Group) of prostate cancer only with four sets of MRI scan data. We applied various machine learning and deep learning techniques (i.e. various Convolutional Neural Networks) and successfully predicted GGG.


  At last, we are also interested in Texture Analysis on Fluence Map and Radiomics feature for prognosis prediction as well.