A Learning Based Formulation of Parametric Curve Fitting for Bioimage Analysis.
model.py contains the network design.
loss.py contains the loss function for training.
Run training.py for training.
Run test.py for testing.
We provide trained weights for the network for quick testing on BBBC038 in the checkpoints folder. We also provide the dataset. For using over a new dataset, just replace the imgs_train.npy (training images), contours_train.npy (training labels), imgs_test.npy (test images), and contours_test.npy (test labels) with relevant files.
Requirements -
numpy
Tensorflow 1.14