Commit ac5ae422226145edf277c89f7e6fb6c2c73b08da
1 parent
27fd135f
added digital staining functions
Showing
1 changed file
with
50 additions
and
0 deletions
Show diff stats
| 1 | +# -*- coding: utf-8 -*- | |
| 2 | +""" | |
| 3 | +Created on Tue Jul 25 16:28:37 2017 | |
| 4 | + | |
| 5 | +@author: david | |
| 6 | +""" | |
| 7 | + | |
| 8 | +import spectral | |
| 9 | +import envi | |
| 10 | +import classify | |
| 11 | +import numpy | |
| 12 | +import scipy | |
| 13 | +import scipy.misc | |
| 14 | +import sklearn | |
| 15 | +import sklearn.naive_bayes | |
| 16 | +import sklearn.neural_network | |
| 17 | +import glob | |
| 18 | +import matplotlib.pyplot as plt | |
| 19 | +import random | |
| 20 | + | |
| 21 | +def generate_stain(envifile, stainfile, N=5000, batch_size=10000, validate=True): | |
| 22 | + E = envi.envi(envifile) | |
| 23 | + mask = classify.random_mask(E.mask, N) | |
| 24 | + | |
| 25 | + Ft = E.loadmask(mask).transpose() | |
| 26 | + | |
| 27 | + stain = numpy.rollaxis(scipy.misc.imread(stainfile), 2) | |
| 28 | + Tt = spectral.sift2(stain, mask).transpose() | |
| 29 | + | |
| 30 | + CLASS = sklearn.neural_network.MLPRegressor(solver='lbfgs', alpha=1e-5, hidden_layer_sizes=(), random_state=1, verbose=True) | |
| 31 | + CLASS.fit(Ft, Tt) | |
| 32 | + | |
| 33 | + if validate == False: | |
| 34 | + return CLASS | |
| 35 | + | |
| 36 | + plt.ion() | |
| 37 | + Fv = E.loadbatch(batch_size) #load the first batch | |
| 38 | + n = 0 | |
| 39 | + while not Fv == []: #loop until an empty batch is returned | |
| 40 | + if n == 0: | |
| 41 | + Tv = CLASS.predict(Fv.transpose()).transpose() | |
| 42 | + else: | |
| 43 | + Tv = numpy.append(Tv, CLASS.predict(Fv.transpose()).transpose(), 1) #append the predicted labels from this batch to those of previous batches | |
| 44 | + COLORS = spectral.unsift2(Tv, E.batchmask()) #convert the matrix of class labels to a 2D array | |
| 45 | + RGB = numpy.rollaxis(COLORS, 0, 3).astype(numpy.ubyte) | |
| 46 | + plt.imshow(RGB) #display it | |
| 47 | + plt.pause(0.05) | |
| 48 | + Fv = E.loadbatch(batch_size) #load the next batch | |
| 49 | + n = n + 1 | |
| 50 | + return CLASS, RGB | ... | ... |