A new neural network for automatic head and neck cancer (HNC) segmentation from magnetic resonance imaging (MRI) is presented. The proposed neural network is based on U-net, which combines features from different resolutions to achieve end-to-end locating and segmentation of medical images. In this work, the dilated convolution is introduced into U-net, to obtain larger receptive field so that extract multi-scale features. Also, this network uses Dice loss to reduce the imbalance between classes. The proposed algorithm is trained and tested on real MRI data. The cross-validation results show that the new network outperformed the original Unet by 5% (Dice score) on head and neck tumour segmentation.
|Number of pages||4|
|Publication status||Published - 2 Sep 2019|
|Event||27th European Signal Processing Conference - A Coruna, Spain|
Duration: 2 Sep 2019 → 6 Sep 2019
|Conference||27th European Signal Processing Conference|
|Abbreviated title||EUSIPCO 2019|
|Period||2/09/19 → 6/09/19|
- MRI data
- head and neck cancer
- dilated convolution
- semantic segmentation
Zhao, B., Soraghan, J., Di Caterina, G., & Grose, D. (2019). Segmentation of head and neck tumours using modified U-net. Paper presented at 27th European Signal Processing Conference, A Coruna, Spain.