OPAL: Optimized PatchMatch for Accurate Label fusion

Back to research page


  • Optimized PatchMatch for Accurate Label fusion (OPAL) and segmentation of 3D MRI structures
  • New multi-scale and multi-feature framework
  • Fast and accurate results on hippocampus segmentation (<1s)
  • Inter-class expert reliability is reached on healthy and pathological datasets
  • Extensions to Cerebellum segmentation and Alzheimer's disease classification
  • Integration into the online volBrain platform


  • Use of a linearly registered database with expert-based segmentation of structures
  • Patch-based approximate nearest neighbors (ANN) matching and transfer of manual segmentation information

Fast search of patch-based neighbors
  • PatchMatch algorithm [6] introduced for 2D patch matching between 2 images
  • New Optimized PatchMatch (OPM) algorithm for multiple 3D patch matching within a database
  • Search based on propagation step (PS) of good matches, with new constrained initialization (CI) and random search (CRS) steps

Label fusion from expert-based segmentations
  • Patch-wise label fusion from selected ANN, inspired from non-local means [7]

Multi-scale and multi-feature framework
  • Fast ANN search and label fusion on several scales and features
  • Late fusion of estimator maps

  • Leave-one-out validation procedure. Comparison on the Dice index [8]
  • OPAL is more accurate and much faster than state-of-the-art methods

On the ICBM database

  • Database of 80 healthy subjects [9]

On the EADC-ADNI database

  • Database of 100 subjects with AD pathologies [10]

Main (to cite)

  1. R. Giraud, V.-T. Ta, N. Papadakis, et al.: An Optimized PatchMatch for multi-scale and multi-feature label fusion
    NeuroImage, 2015   https://hal.archives-ouvertes.fr/hal-01198703/file/Giraud_2015_NR.pdf   BibTex

  1. V.-T. Ta, R. Giraud, D. L. Collins and P. Coupé: Optimized PatchMatch for Near Real Time and Accurate Label Fusion
    Int. Conf. on Medical Image Computing and Computer Assisted Interventions (MICCAI), 2014   https://hal.archives-ouvertes.fr/hal-01006329/file/Ta_MICCAI2014.pdf   BibTex
  2. R. Giraud, V.-T. Ta, et al.: Optimisation de l'algorithme PatchMatch pour la segmentation de structures anatomiques
    Groupement de Recherche en Traitement du Signal et des Images (GRETSI), 2015   https://hal.archives-ouvertes.fr/hal-01170197/file/Giraud_2015_GRETSI.pdf   BibTex  

  • New method for cerebellum lobule segmentation on 3D MRI
  • Extension of OPAL to multiple structure segmentation
  • Integrated into online platform volBrain

  1. J. Romero, P. Coupé, R. Giraud, V.-T. Ta, V. Fonov, et al.: CERES: A new cerebellum lobule segmentation method
    NeuroImage, 2016

  • Fast feature extraction on DTI (Diffusion Tensor Imaging) using OPAL
  • The extracted features lead to more accurate Alzheimer's disease detection
       (AD = Alzheimer's Disease, NC = Normal Control, MCI = Mild Cognitive Impairment)

  1. K. Hett, V.-T. Ta, R. Giraud, M. Mondino, et al.: Patch-based DTI grading: Application to Alzheimer's disease classication
    Int. Work. on Patch-Based Techniques in Medical Imaging (Patch-MI) at MICCAI, 2016

  1. C. Barnes, E. Shechtman, A. Finkelstein, et al.: PatchMatch: A randomized correspondence algorithm for structural image editing
    ACM Transactions of Graphics, 2009
  2. P. Coupé, J. V. Manjón, et al.: Patch-based segmentation using expert priors: application to hippocampus and ventricle segmentation
    NeuroImage, 2011
  3. A. P. Zijdenbos, B. M. Dawant, R. A. Margolin, et al.: Morphometric analysis of white matter lesions in MR images: method and validation
    IEEE Trans. on Medical Imaging, 1994
  4. J. C. Mazziotta, A. W. Toga, A. C. Evans, et al.: A probabilistic atlas of the human brain: theory and rationale for its development
    NeuroImage, 1995
  5. M. Boccardi, et al.: Delphi definition of the EADC-ADNI harmonized protocol for hippocampal segmentation on magnetic resonance
    Alzheimer's & Dementia, 2014
  6. D. L. Collins and J. C. Pruessner: Towards accurate, automatic segmentation of the hippocampus and amygdala from MRI by augmenting ANIMAL with a template library and label fusion.
    NeuroImage, 2010
  7. T. Tong, R. Wolz, P. Coupé, J. V. Hajnal, D. Rueckert, et al.: Segmentation of MR images via discriminative dictionary learning and sparse coding: Application to hippocampus labeling.
    NeuroImage, 2013
  8. S. Tangaro, N. Amoroso, et al.: Automated voxel-by-voxel tissue classification for hippocampal segmentation: Methods and validation
    Physica Medica, 2014
  9. F. Roche, J. Schaerer, S. Gouttard, et al.: Accuracy of BMAS Hippocampus Segmentation Using the Harmonized Hippocampal Protocol
    Alzheimer's & Dementia, 2014
  10. K. R. Gray, M. Austin, et al.: Integration of EADC-ADNI Harmonised hippocampus labels into the LEAP automated segmentation technique
    Alzheimer's & Dementia, 2014

Back to top