CT TRAINING AND VALIDATION SERIES FOR 3D AUTOMATED SEGMENTATION OF INNER EAR USING U-NET ARCHITECTURE DEEP-LEARNING MODEL

Citation Author(s):
Jonathan
Lim
Julien
Ognard
Arnaud
Attye
Submitted by:
Jonathan Lim
Last updated:
Thu, 10/19/2023 - 09:34
DOI:
10.21227/y91d-4p39
Data Format:
License:
0
0 ratings - Please login to submit your rating.

Abstract 

This data set contains: 

- Training dataset: 271 CT-scans of inner ears used for optimization and training of the model. 

- Validation dataset: 70 CT-scans of inner ears used for external validation. 

- U-net architecture deep-learning model's weight after optimized training. 

- All manual segmentations performed for both datasets. 

- All post-processed automated segmentations performed by the model for bothd atasets. 

All CT-scans are related to matching automated and manual segmentations and have been named identically. 

Example : CT-scan “TDMvalext001“ in folder “Validation CT-scans" is related to “TDMvalext001” manual segmentation in folder “Validation manual segmentation”, and to “TDMvalext001_PRED” in folder “Validation automated segmentation”. 

Instructions: 

Training and validation datasets are both provided with CT-scans and their matching manual segmentation (same name) and automated segmentation (same name + _PRED) in .nii format.

Model's weight is provided in .txt format.