search

wentaozhu / DeepLung

WACV18 paper "DeepLung: Deep 3D Dual Path Nets for Automated Pulmonary Nodule Detection and Classification"
Apache License 2.0
316 stars 143 forks source link

About test results. #68

Closed anruoxi123 closed 5 years ago

anruoxi123 commented 5 years ago

Hello, I have reproduced the detection part of pulmonary nodules in your code, and after adding Dropout layer to the end of each residual block, the parameter of sensitivity has been improved strangely, but the FROC score has decreased and is only more than 60%, may I ask what is the reason for this? Thank you very much!

wentaozhu commented 5 years ago

Is the code has the dropout layer? How about adding dropout in the second last or third last layer?

anruoxi123 commented 5 years ago

The source code was there before, but it was commented out, and I just opened the comments section. LRCB0 R}OEC8Z~(}G Q@XWC

wentaozhu commented 5 years ago

You may try different ways of dropout.

anruoxi123 commented 5 years ago

Ok, thank you very much!