OPTIMIZED MULTI-ATLAS PROSTATE SEGMENTATION FROM 3D CT IMAGES

Yitian Zhou , Laurent Launay , Julien Bert (1) , Renaud de Crevoisier (2, 3) , Oscar Acosta (2, 3)

1 LaTIM - Laboratoire de Traitement de l'Information Medicale
2 LTSI - Laboratoire Traitement du Signal et de l'Image
3 CRLCC - CRLCC Eugène Marquis

The purpose of this study was to evaluate and optimize the performance of a multi-atlas based method for the segmentation of prostate in CT scans improving it up to the limit of the inter-observer variability. We assessed and optimized the atlas selection, the Non-Rigid Registration (NRR) and the label fusion steps by introducing new similarity measures based on image features and a multi-scale weighted majority voting. Cross validation results on 45 CT images suggested that the similarity measure based on the local feature histogram of oriented gradients outperformed classical intensity-based metrics for atlas selection. Besides, the NiftyReg optimized in a region of interest was found to be the optimal NRR algorithm. For the label fusion, the multi-scale weighted majority voting outperformed other approaches. All those improvements led to Dice scores of 0.84 +/- 0.03, which are comparable to the inter-observer variability for manual contouring.

Mots clés: Multi-atlas segmentation; CT; prostate; inter-observer variability; radiotherapy planning

Domaines: Ingénierie biomédicale

Yitian Zhou, Laurent Launay, Julien Bert, Renaud de Crevoisier, Oscar Acosta. OPTIMIZED MULTI-ATLAS PROSTATE SEGMENTATION FROM 3D CT IMAGES. 16th IEEE International Symposium on Biomedical Imaging (ISBI), Apr 2019, Venice, Italy. ⟨hal-02302409⟩