3D Image Segmentation for Lung Cancer Using U-Net Architecture

EasyChair Preprint no. 8924, version history

VersionDatePagesVersion notes
1October 3, 20225
2October 24, 20225

Segmentation of pulmonary cancer is the big ambiguity of medical staff in their diagnostic, we will present our U-NET algorithm in this paper. The general idea is to create an optimal segmentation that allowed the medical staff  to distinct the different parts of the tumor using the U-Net architecture which represent the more elegant architecture, called a fully convolution network. The main idea is to complete a contracting network by successive layers; pooling operations are replaced by over sampling operators. So, the resolution of the output is increased by these layers. To obtain the best result of the performance from the different data merged we choose the technique of 3D-UNET architecture.

To elicit the optimal and best quality result of the segmented images, we choose this technique that allowed merging the different data source. The experimental results of the segmentation are approved by using U-NET for image segmentation.

Keyphrases: Classification., Conventional network, Segmentation, UNet architecture

BibTeX entry
BibTeX does not have the right entry for preprints. This is a hack for producing the correct reference:
  author = {Elloumi Nabila and Ben Chaabane Salim and Seddik Hassen},
  title = {3D Image Segmentation for Lung Cancer Using U-Net Architecture},
  howpublished = {EasyChair Preprint no. 8924},

  year = {EasyChair, 2022}}