.. -*- coding: utf-8 -*- .. _bob.ip.binseg.results.cod.vessel: ============================================== Retinal Vessel Segmentation for Retinography ============================================== .. list-table:: :header-rows: 2 * - - - :py:mod:`driu ` - :py:mod:`hed ` - :py:mod:`m2unet ` - :py:mod:`unet ` * - Dataset - 2nd. Annot. - 15M - 14.7M - 0.55M - 25.8M * - :py:mod:`drive ` - 0.788 (0.021) - `0.768 (0.031) `_ - `0.750 (0.036) `_ - `0.771 (0.027) `_ - `0.775 (0.029) `_ * - :py:mod:`stare ` - 0.759 (0.028) - `0.786 (0.100) `_ - `0.738 (0.193) `_ - `0.800 (0.080) `_ - `0.806 (0.072) `_ * - :py:mod:`chasedb1 ` - 0.768 (0.023) - `0.778 (0.031) `_ - `0.777 (0.028) `_ - `0.776 (0.031) `_ - `0.779 (0.028) `_ * - :py:mod:`hrf ` - - `0.742 (0.049) `_ - `0.719 (0.047) `_ - `0.735 (0.045) `_ - `0.746 (0.046) `_ * - :py:mod:`iostar-vessel ` - - `0.790 (0.023) `_ - `0.792 (0.020) `_ - `0.788 (0.021) `_ - `0.783 (0.019) `_ Notes ----- * The following table describes recommended batch sizes for 24Gb of RAM GPU card, for supervised training of COD-systems: .. code-block:: sh # change and by one of items bellow $ bob binseg experiment -vv --batch-size= --device="cuda:0" .. list-table:: * - **Models / Datasets** - :py:mod:`drive-covd ` - :py:mod:`stare-covd ` - :py:mod:`chasedb1-covd ` - :py:mod:`iostar-vessel-covd ` - :py:mod:`hrf-covd ` * - :py:mod:`driu ` / :py:mod:`driu-bn ` - 4 - 4 - 2 - 2 - 2 * - :py:mod:`m2unet ` - 8 - 4 - 4 - 4 - 4 .. include:: ../../links.rst