Difference between revisions of "2017 Winter Project Week/Convolutional neural nets for lung cancer survival prediction"

From NAMIC Wiki
Jump to: navigation, search
(Created page with "==Project Description== {| class="wikitable" ! style="text-align: left; width:27%" | Objective ! style="text-align: left; width:27%" | Approach and Plan ! style="text-alig...")
(No difference)

Revision as of 14:11, 9 January 2017

Home < 2017 Winter Project Week < Convolutional neural nets for lung cancer survival prediction

Project Description

Objective Approach and Plan Progress and Next Steps
  • Improve the predictive performance and retrian our CNN's - benchmark "deep" features results with those from "hand-crafted" radiomic features.
  • Identify best data preprocessing techniques for medical images: normalization vs standardization
  • 2d single slice vs 2.5d stacked orientations (RGB)
  • Identify dest data augmentation methods for medical images: cropping, brigtness, contrast..etc