Faculty Name | Contact

Content

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Vivamus viverra, eros sed ullamcorper adipiscing, dolor ante aliquam justo, sed mollis turpis augue ut nisl. Nunc id orci metus, a rhoncus massa. In hac habitasse platea dictumst. Fusce sit amet arcu a dui mattis ultricies.

Curvilinear Structures Segmentation in Medical Images
  • Accelerating Convolutional Sparse Coding for Curvilinear Structures Segmentation by Refining
    SCIRD-TS Filter Banks

    Short Description:
    Deep learning has shown great potential for curvilinear structure (e.g. retinal blood vessels and neurites) segmentation as demonstrated by a recent auto-context regression architecture based on filter banks learned by convolutional sparse coding. However, learning such filter banks is very time-consuming, thus limiting the amount of filters employed and the adaptation to other data sets (i.e. slow re-training). We address this limitation by proposing a novel acceleration strategy to speed-up convolutional sparse coding filter learning for curvilinear structure segmentation. Our approach is based on a novel initialisation strategy (warm start), and therefore it is different from recent methods improving the optimisation itself. Our warm-start strategy is based on carefully designed hand-crafted filters (SCIRD-TS), modelling appearance properties of curvilinear structures which are then refined by convolutional sparse coding. Experiments on four diverse data sets, including retinal blood vessels and neurites, suggest that the proposed method reduces significantly the time taken to learn convolutional filter banks (i.e. up to -82%) compared to conventional initialisation strategies. Remarkably, this speed-up does not worsen performance; in fact, filters learned with the proposed strategy often achieve a much lower reconstruction error and match or exceed the segmentation performance of random and DCT-based initialisation, when used as input to a random forest classifier.
    MATLAB code   Downloads:
    Related publication:
    R. Annunziata and E. Trucco:
    "Accelerating Convolutional Sparse Coding for Curvilinear Structures Segmentation by Refining
    SCIRD-TS Filter Banks"
    ,
    IEEE Transactions on Medical Imaging, 2016.

  • SCIRD-TS (Scale and Curvature Invariant Ridge Detector for Thin Structures).
    Short Description:
    SCIRD (Annunziata et al., MICCAI, 2015) has shown good detection performance on curvilinear structures acquired at a good resolution (diameter > 2/3 pixels). However, when resolution is low, SCIRD's detection performance tends to degrade (the reason is discussed in Annunziata et al., IEEE TMI, 2016). We addressed this limitation by proposing a different mathematical derivation of SCIRD, which allows us to obtain better detection performance at a parity of computational complexity; this new ridge detector is called SCIRD-TS.
    MATLAB code   Downloads:
    Related publication:
    R. Annunziata and E. Trucco:
    "Accelerating Convolutional Sparse Coding for Curvilinear Structures Segmentation by Refining
    SCIRD-TS Filter Banks"
    ,
    IEEE Transactions on Medical Imaging, 2016.

  • SCIRD (Scale and Curvature Invariant Ridge Detector) for Tortuous and Fragmented Structures.
    Short Description:
    SCIRD is a novel (hand-crafted) ridge detector designed to improve the enhancement/segmentation of tortuous and fragmented curvilinear structures in the medical domain, such as corneal nerve fibres and neurites. In addition to the elongation and width parameters typically used to tune filter banks performing this task, SCIRD includes 2 other parameters:
    1) "k" controlling the curvature of each filter, allowing a better enhancement/segmentation of tortuous structures;
    2) "alpha" controlling the level of contrast enhancement required to deal with severe contrast variations for each specific dataset.
    MATLAB code   Downloads:
    Related publication:
    Annunziata R., Kheirkhah A., Hamrah P., Trucco E.:
    "Scale and Curvature Invariant Ridge Detector for Tortuous and Fragmented Structures",
    Proc. of Medical Image Computing and Computer Assisted Interventions (MICCAI), 2015, Munich, Germany.