search

wentaozhu / DeepLung

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

Some questions about the size of pixel feature #117

Closed 3learning12340 closed 4 years ago

3learning12340 commented 4 years ago

Thank you for sharing the code. If I remember correctly, the maximum size of the nodules is greater than 16 in LIDC/IDRI dataset. However, for pixel feature, you use the cropped size of 16x16x16 and center as the detected nodule center in the experiments. Therefore, I think the pixel feature may not contain all information of nodules. what's reason that you crop 16x16x16 cube voxel ? just to save GPU meomery? I am looking forward to your reply.

3learning12340 commented 4 years ago

I've duplicated most of the code and I read your paper carefully. However, I still have some doubts about the paper and I want to ask you some questions about the paper while you are free. I used to do some study on the maglignant and benign nodule classification in two-dimensional CT images. I want to do some study on the detection and classification in the three-dimensional images and I am really thankful for your code sharing, so that I can have a better start. Can I make a friend with you by adding Wechat, QQ or facebook friends, My google email is perfect0insist@gmail.com.

wentaozhu commented 4 years ago

Sure. Please use GitHub to report issues. My wechat is zhuwentao_1991