-
Notifications
You must be signed in to change notification settings - Fork 2
/
GS.py
418 lines (360 loc) · 15.5 KB
/
GS.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
import numpy as np
from tqdm import trange
## check gpu functionality! ##
global gpu_func
gpu_func = True
try:
__import__("pycuda")
except ImportError:
gpu_func = False
else:
import pycuda.autoinit
import pycuda.gpuarray as gpuarray
import skcuda.fft as cu_fft
from pycuda.elementwise import ElementwiseKernel
from skcuda import misc
misc.init()
def to_gpu_c(somedata):
# all my complex data
return gpuarray.to_gpu(somedata.astype('complex64'))
def to_gpu_f(somedata):
# all my float data
return gpuarray.to_gpu(somedata.astype('float32'))
def to_gpu_b(somedata):
# all my bool data
return gpuarray.to_gpu(somedata)
## check gpu functionality! ##
class GS():
"""
Gerchberg-Saxton (GS) algorithm for non-linear wave refinement. GS algorithm
uses a loop of propagation and back-propagation operations of this wave
convolved with a contrast transfer function (CTF) and a spatial coherence
envelope that model the imaging system. The refinement works by combining
experimental intensities with the propagated waves. Thus, it is mandatory
to provide a (padded) focal series; An initial wave and CTF should also be
provided, but can be initialized from input (more below).
Example
-------
>>> from GS import GS
>>> gsax = GS(fs) # Init: sets inputs for GS # Init: sets inputs for GS
>>> wave = gsax() # Iterate
Parameters can be also updated by using specialized "set" methods:
>>> gsax.set_initial_wave(wave_other)
The code can run in a GPU, if available:
>>> gsax = GS(using_gpu=True) # Init will try to use GPU...
>>> wave = mftie() # if available, this step is the sped-up!
Authors
-------
Alberto Eljarrat and Christoph T. Koch.
Institute für Physik. Humboldt Universität zu Berlin.
No commercial use or modification of this code
is allowed without permission of the authors.
"""
def __init__(self, fs, wave=None, ctf=None, alpha=None, using_gpu=False,
*args, **kw):
"""
Parameters
----------
fs : Hyperspy Image
Dataset containing an experimental focal series and including some
pad region.
wave : Hyperspy Image
Image containing a wave with same signal dimensions as the input focal
series. The wave is expected in its real-space representation, a 2D
complex function. If not provided the algorithm is initialized with the
in-focus image and zero phase.
ctf : Hyperspy Image
Dataset containing the contrast transfer function with the same shape as
the input focal series. The CTF is expected in its reciprocal-space
representation, a 2D complex function. If not provided, the
"fs.get_contrast_transfer(*args, **kw)" result is used.
alpha : number
Convergence-semi angle, use this to set a spatial coherence envelope.
using_gpu : bool
Sets the GPU/CPU functionality
*args, **kw passed to the "set_contrast_transfer" method.
"""
if (using_gpu and not gpu_func):
# Not so fast ...
self.using_gpu = - using_gpu
raise Exception('GPU functionality not available !')
elif using_gpu:
# Set-up GPU Machinery
self.using_gpu = using_gpu
# EWK to avoid phase wrapping
self.cuphase = ElementwiseKernel(
"pycuda::complex<float> *a,pycuda::complex<float> *b,float *c",
"c[i] = c[i] + arg(a[i]/b[i]);")
else:
# User selects CPU
self.using_gpu = using_gpu
if wave is None:
# Get amplitude from middle image
amp = fs.data[self.Nz_half,:,:]
wave = fs._get_signal_signal()
wave.metadata = fs.metadata.deepcopy()
wave.data = np.sqrt(amp) * np.exp(1j*np.zeros_like(amp))
self.set_initial_wave(wave)
self.set_focal_series(fs)
self.set_contrast_transfer(ctf, *args, **kw)
self.set_convergence_envelope(alpha)
def __call__(self, Niters=5, unpad=False):
"""
Will run the reconstruction, in the GPU or CPU, depending on how the
parameters were set-up, and return the resulting wave with or without
padding. This result is also stored in the "new_wave" attribute.
Parameters
----------
Niters : int
Numer of GS-loop iterations. Set to 5 by defect.
unpad : bool
Pad / Unpad switch for the returned wave. Set to False by defect.
Returns
-------
new_wave: Hyperspy Image
Gerchberg-Saxton refined complex wave.
Notes
-----
The flux-preserving G-S algorithm contains an additional convolution
step if a convergence envelope is used. This is activated by the
presence of said envelope. In case a calculation without the
additional convolution should run after one (or several ones) with it,
the convergence envelope attribute must be destroyed. Use, for instance:
>>> del self.Esdata
"""
if self.using_gpu:
self.run_gpu(Niters)
else:
self.run_cpu(Niters)
return self._return_new_wave(unpad)
def _return_new_wave(self, unpad=False):
"""
This method returns the current wave, with or without padding depending
on a boolean parameter unpad.
"""
if self.using_gpu:
self.new_wave.data = self.wdata.get()
else:
self.new_wave.data = self.wdata
if unpad:
return self.new_vave.unset_padding()
else:
return self.new_wave
def _return_new_phase(self, unpad=False):
"""
This method returns the current phase, with or without padding depending
on a boolean parameter unpad.
"""
new_phase = self.new_wave.deepcopy()
if self.using_gpu:
new_phase.data = self.phase_data.get()
else:
new_phase.data = self.phase_data.copy()
if unpad:
return new_phase.unset_padding()
else:
return new_phase
def set_focal_series(self, fs, defoci=None):
"""
Setting the focal series also sets important parameters as the padding
and defocus values.
Parameters
----------
fs : Hyperspy Image
Mandatory! A focal series with all the appropriate parameters. Padding
is read from metadata.
defoci : array
Optional, use it to explicitly set defoci values. Must have a dimension
equal to the navigation axis of fs.
"""
# Real space parameters
Nz, Ny, Nx = fs.data.shape
if defoci is None:
defoci = fs.axes_manager.navigation_axes[0].axis
Nz = len(defoci)
self.zdim = defoci
self.k2 = fs.get_fourier_space()
self.Nz = Nz
# In the middle of the focal series is the reference plane
Nz_half = np.ceil(Nz/2).astype('int') - 1
# Pad mask
Npy, Npx = fs.metadata.Signal.pad_tuple
mask = np.zeros((Nz,Ny,Nx), dtype=np.bool)
mask[:,Npy[0]:(Ny-Npy[1]), Npx[0]:(Nx-Npx[1])] = True
# Set some parameters
self.new_wave = fs._get_signal_signal()
self.new_wave.metadata = fs.metadata.deepcopy()
self.shape = (Nz, Ny, Nx)
# Allocate things ...
if self.using_gpu:
# ... in GPU!
self.Iexp = to_gpu_f(fs.data)
self.mask = to_gpu_b(mask)
# - The plans for FFT
self.pft3dcc = cu_fft.Plan((Ny, Nx), np.complex64, np.complex64, Nz)
self.pft2dcc = cu_fft.Plan((Ny, Nx), np.complex64, np.complex64)
else:
# ... in CPU!
self.Iexp = fs.data.astype('complex64')
self.mask = mask
def set_initial_complex_wave(self, amplitude, phase):
"""
Use it to build a complex wave from amplitude and phase. Padding must be
set in advance for both signals and should coincide. Calls "set_wave".
Parameters
----------
amplitude : Hyperspy Image
The amplitude of complex wave.
phase : Hyperspy Image
The phase of complex wave.
"""
wave = amplitude.deepcopy()
wave.metadata = amplitude.metadata.deepcopy()
wave.data = np.sqrt(amplitude.data) * np.exp(1j*phase.data)
self.set_input_wave(wave, phase.data.copy())
def set_initial_wave(self, wave, phase_data=None):
"""
Use it to directly set the wave. It writes to an attribute "wdata" that
is overwritten everytime a G-S calculation runs. This allows to call G-S
refinements consecutively. Nevertheless, the attribute "wave" will
contain the original input wave.
Parameters
----------
wave : Hyperspy Image
A complex wave with all the appropriate parameters.
phase_data : numpy array
Use this to set the phase data separately to avoid wrapping issues.
"""
# TODO: trigget set_contrast_transfer if wave changes something important
self.wave = wave
# The phase data is stored separately to avoid wrapping issues
if phase_data is None:
phase_data = np.angle(wave.data)
if self.using_gpu:
self.wdata = to_gpu_c(wave.data)
self.phase_data = to_gpu_f(phase_data)
else:
self.wdata = wave.data.astype('complex64')
self.phase_data = phase_data.copy()
def set_contrast_transfer(self, ctf, *args, **kw):
"""
Use it to set the contrast transfer function. If a ctf image is not
provided, the get_contrast_transfer method is called with defoci set by
the focal series navigation axes. More info in the docs therein.
"""
if ctf is None:
ctf = self.wave.get_contrast_transfer(self.zdim, *args, **kw)
if self.using_gpu:
self.ctfd = to_gpu_c(ctf.data)
else:
self.ctfd = ctf.data.astype('complex64')
def set_convergence_envelope(self, alpha):
"""
Sets the spatial-coherence envelope function. The convergence semiangle,
alpha, is set by input preferentially or, if possible, by metadata.
Parameters
----------
alpha : float or None
To set the convergence semi-angle, in radians. If None, will try to
read from metadata.
Note
----
This version also has sets an ElementwiseKernel for combining the
experimental intensities and simulated waves in each GS iteration.
"""
k2 = self.k2
metadata = self.new_wave.metadata
if (alpha is None):
if metadata.has_item('ModImage.convergence_semiangle'):
alpha = metadata.ModImage.convergence_semiangle
else:
alpha = 0.
envelope = np.exp( -k2 * ((0.5*self.zdim*alpha)**2)[:,None,None] )
if self.using_gpu and (alpha != 0):
self.Esdata = to_gpu_f(envelope)
# EWK for wave combo
self.cuwave = ElementwiseKernel(
"float *a,pycuda::complex<float> *b,float *c,pycuda::complex<float> *d",
"const pycuda::complex<float> j(0.0,1.0); \
d[i] = (abs(b[i])+sqrt(a[i])-sqrt(c[i])) * exp(j*arg(b[i]));")
elif self.using_gpu and (alpha == 0):
# No spatial-coherence envelope, but we need
# EWK for wave combo because GPU
self.cuwave = ElementwiseKernel(
"float *a,pycuda::complex<float> *b,pycuda::complex<float> *c",
"const pycuda::complex<float> j(0.0,1.0); \
c[i] = sqrt(a[i]) * exp(j*arg(b[i]));")
else:
self.Esdata = envelope.astype('float32')
self.new_wave.metadata.set_item('ModImage.convergence_semiangle', alpha)
def run_gpu(self, Niters):
"""
Run G-S on GPU. The result is overwritten on the attribute "self.wdata"
containing a pycuda array.
"""
Nz, Ny, Nx = self.shape
# Allocate output data
wdata = gpuarray.empty((Ny, Nx), np.complex64)
sim = gpuarray.empty((Nz, Ny, Nx), np.complex64)
Isim = gpuarray.empty((Nz, Ny, Nx), np.complex64)
for io in trange(Niters):
# Propagate the initial wave to simulate defocused waves
# Psi(x,y,z) = convolve[Psi(x,y,0), CTF(x,y,z)]
cu_fft.fft(self.wdata, wdata, self.pft2dcc)
for kk in range(Nz):
sim[kk,:,:] = self.ctfd[kk,:,:] * wdata
cu_fft.ifft(sim, sim, self.pft3dcc, True)
if hasattr(self, 'Esdata'):
# Use the intensities, Isim = |Psi|**2
# Convolve with spatial-coherence envelope
# Isim = convolve[Isim, Es]
Isim = sim * sim.conj()
cu_fft.fft(Isim, Isim, self.pft3dcc)
cu_fft.ifft(Isim*self.Esdata, Isim, self.pft3dcc, True)
# Combine experimental and simulated amplitudes with simulated phase
# Psi' = [abs(Psi)+sqrt(Iexp)-sqrt(Isim)]*exp[i*arg(Psi)]
self.cuwave(self.Iexp, sim, Isim.real, Isim)
else:
# Combine experimental amplitudes with simulated phase
# Psi' = [sqrt(Iexp)]*exp[i*arg(Psi)]
self.cuwave(self.Iexp, sim, Isim)
sim = gpuarray.if_positive(self.mask, Isim, sim)
# then back-propagate to the exit plane and take average
# Psi(x,y,0) = < convolve[Psi, CTF*] >_z
cu_fft.fft(sim, sim, self.pft3dcc)
sim = sim * self.ctfd.conj()
cu_fft.ifft(sim, sim, self.pft3dcc, True)
wdata = misc.mean(sim.reshape(Nz,Nx*Ny),0).reshape(Ny,Nx)
# update phase and wave
self.cuphase(wdata, self.wdata, self.phase_data)
self.wdata = wdata.copy()
def run_cpu(self, Niters):
"""
Run G-S on CPU. The result is overwritten on the attribute "self.wdata"
containing a numpy array.
"""
Nz, Ny, Nx = self.shape
for io in trange(Niters):
# Propagate the initial wave to simulate defocused waves
# Psi(x,y,z) = convolve[Psi(x,y,0), CTF(x,y,z)]
Psi = np.fft.ifft2(np.fft.fft2(self.wdata)[None,:,:]*self.ctfd)
if hasattr(self, 'Esdata'):
# Use the intensities, Isim = |Psi|**2
# Convolve with spatial-coherence envelope
# Isim = convolve[Isim, Es]
Psim = np.fft.ifft2(np.fft.fft2(Psi * Psi.conj()) * self.Esdata)
# Combine experimental and simulated amplitudes with simulated phase
# Psi' = [abs(Psi)+sqrt(Iexp)-sqrt(Isim)]*exp[i*arg(Psi)]
Psim = (np.abs(Psi)+np.sqrt(self.Iexp)-np.sqrt(Psim.real)) * \
np.exp(1j*np.angle(Psi))
else:
# Combine experimental amplitudes with simulated phase
# Psi' = [sqrt(Iexp)]*exp[i*arg(Psi)]
Psim = np.sqrt(self.Iexp) * np.exp(1j*np.angle(Psi))
Psi[self.mask] = Psim[self.mask]
# then back-propagate to the exit plane and take average
# Psi(x,y,0) = < convolve[Psi, CTF*] >_z
wdata = np.mean(np.fft.ifft2(np.fft.fft2(Psi)*self.ctfd.conj()),0)
# Update phase and wave
self.phase_data = self.phase_data + np.angle(wdata/self.wdata)
self.wdata = wdata.copy()