improving the efficiency of manual ground truth labeling
play

Improving the Efficiency of Manual Ground Truth Labeling Using - PowerPoint PPT Presentation

Improving the Efficiency of Manual Ground Truth Labeling Using Automated Anatomy Segmentation Hongzhi Wang PhD, Prasanth Prasanna MD, Jose Morey MD, Tanveer F. Syeda-Mahmood PhD Medical Sieve Group, IBM Almaden Research Center Anatomy


  1. Improving the Efficiency of Manual Ground Truth Labeling Using Automated Anatomy Segmentation Hongzhi Wang PhD, Prasanth Prasanna MD, Jose Morey MD, Tanveer F. Syeda-Mahmood PhD Medical Sieve Group, IBM Almaden Research Center

  2. Anatomy Segmentation: labeling anatomical structures of interest in medical images • Corner stone for quantitative image analysis

  3. Gold standard: manual segmentation Manual tracing in ITK-SNAP Manual tracing in Amira

  4. Gold standard: manual segmentation Manual tracing in ITK-SNAP Manual tracing in Amira • Very time consuming, hours or even days to annotate a single 3D volume

  5. Techniques for Assisting Manual Segmentation • Interactive manual segmentation enhanced with interpolation techniques, e.g. MITK and Amira • Only a subset of 2D slices are manually annotated • Annotation for the full 3D volume is generated through interpolation • Annotation time is a fraction of standard manual segmentation, depending on percentage of manually annotated slices • Semi-automatic segmentation, e.g. Pluta et al. 2009, Daisne & Blumhofer 2013 • Automatic segmentation produces initial annotation • Mistakes corrected by human experts • Most focusing on single anatomical structure • ~ 40% time reduction comparing to standard manual segmentation (Daisne & Blumhofer 2013)

  6. Aims • Investigate full potential in time reduction by semi-automatic segmentation • Employing state of the art automatic anatomy segmentation algorithm • Challenging multi-structure segmentation task • cardiac CT anatomy segmentation with 20 anatomical structures • Comparison with interpolation-based interactive manual segmentation

  7. Data Description • 33 cardiac CT studies o 28 cases used for training automatic segmentation o 5 testing cases for experimental validation • 20 structures studied o Bone: sternum, vertebrae o Artery/vein: pulmonary artery (left/right/trunk), aorta (root/ascending/arch/descending), Superior/inferior vena cava o Cardiac structure: Left/right ventricle/atrium, left ventricular myocardium o Valve : aortic valve, tricuspid valve, pulmonary valve, mitral valve

  8. Experimental Setup • Semi-Automatic Segmentation: o Automatic segmentation by multi-atlas label fusion o Manual correction by one clinician (PP) using Amira commercial software (FEI Corporate, Hillsboro, Oregon USA) • Manual segmentation o produced by the same clinician one week after semi-automatic segmentation is finished using Amira with the interpolation technique

  9. Overview for Automatic Anatomy Segmentation Training Training . . . Image 1 Image k Target Image Registration Registration and Warping and Warping Candidate Candidate Segmentation Segmentation Joint Label Fusion Initial Final Post Processing Segmentation Segmentation

  10. Overview for Automatic Anatomy Segmentation Training Training . . . Image 1 Image k Target Image Registration Registration and Warping and Warping Multi-Atlas Label Fusion Candidate Candidate Segmentation Segmentation Joint Label Fusion Initial Final Post Processing Segmentation Segmentation

  11. Multi-Atlas Label Fusion Atlases (Training anatomical volumes) Warped Atlases Given CT Scan registration Label Image Registration Fusion . . . . . . A t ti R lt

  12. Automatic segmentation: Post processing • Remove small isolated segments • Ensure boundaries between right ventricle and pulmonary artery and boundaries between substructures of aorta do not cross axial planes. image multi-atlas segmentation after post processing after manual correction The colored anatomical structures are: sternum; right ventricle; pulmonary artery trunk; myocardium; aortic root; ascending aorta; descending aorta; left atrium; right atrium; vertebrae.

  13. Results: Leave-One-Out Performance of Multi-Atlas Segmentation Inter-rater precision for non-valve structures Automatic Segmentation

  14. Results: Overall Time Reduction 37% time reduction statistically significant with p<0.0001

  15. Conclusions and Discussion • Multi-atlas anatomy segmentation is accurate enough to save time for manual segmentation, even with advanced interpolation tools • Without post processing, manual correction does not save time! • Manually correcting small, isolated segments is time consuming! • Having smooth segmentation is important for efficient manual correction!

  16. Backup Material

  17. Results: Ratio of Corrected Voxels

  18. Results: Corrections are Heterogeneous with respect to Anatomical Structures

Recommend


More recommend