example.py

## generate a batch of training data from 'data' generated by fakearray

from showit import image
import matplotlib.pyplot as plot
import numpy as np
from fakearray import calcium_imaging

from neurocatcher import data_train

data,series,truth=calcium_imaging(shape=(100,100), n=12, t=10, withparams=True)

data=data*255/data.max()
data=np.transpose(data,(1,2,0))

data=[data]

batch_data,batch_truth=data_train(data,[truth],10,20,20-6)

for i,pic in enumerate(batch_data):
	image(np.mean(pic,axis=2))
	plot.show()
	image(batch_truth[i,:,:,0])
	plot.show()


##train a network and check how it performs

import neurocatcher as nc
import fakearray as fa

from showit import image
import numpy as np
import matplotlib.pyplot as plot

# generate some faux calcium imaging data
data, series, truth = fa.calcium_imaging(withparams=True)

# we will train a network that takes the mean image as input
data = [data.mean(axis=0)[..., np.newaxis]]
truth = [truth]

# each layer is defined by (filter footprint, 3 of features)
layers = [(3, 10), (3, 10)]

# the input will be a 15-by-15 patch with a single channel
inputShape = (15, 1)

# train network
acc, network = nc.network.train_conv_net(layers, inputShape, data, truth, batch_size=30, steps=1000)

# show how the loss changes during training
plot.plot(acc[:, 0])
plot.show()
##predict output using trained network

#generate new data to predict on
data,series,truth=calcium_imaging(shape=(240,240), n=75, t=10, noise=0.0, withparams=True)
data = [data.mean(axis=0)[..., np.newaxis]]
truth = [truth]

# get a prediction from the full data set
yhat, ytarget = nc.network.predict_conv_net(network, data, truth)

# visualize the results
plot.figure(figsize=(20, 15))

plot.subplot(1, 2, 1)
plot.imshow(ytarget[0], cmap='bone')

plot.subplot(1, 2, 2)
plot.imshow(yhat[0], cmap='bone')

plot.show()