# Copyright 2020 Diamond Light Source Ltd.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
.. module:: dezinger_gpu
:synopsis: A 3D median-based GPU dezinger plugin to apply to raw projection data
.. moduleauthor::Daniil Kazantsev <scientificsoftware@diamond.ac.uk>
"""
from savu.plugins.filters.base_filter import BaseFilter
from savu.plugins.driver.gpu_plugin import GpuPlugin
from savu.plugins.utils import register_plugin
import numpy as np
from larix.methods.misc_gpu import MEDIAN_DEZING_GPU
[docs]@register_plugin
class DezingerGpu(BaseFilter, GpuPlugin):
def __init__(self):
super(DezingerGpu, self).__init__("DezingerGpu")
self.GPU_index = None
self.res = False
self.start = 0
self.frame_limit = 8
[docs] def pre_process(self):
inData = self.get_in_datasets()[0]
self.proj_dim = inData.data.proj_dim
self._kernel = [1]*3
self._kernel[self.proj_dim] = self.kernel_size
pad_list = [(0, 0)]*3
pad_list[self.proj_dim] = (self.pad, self.pad)
dark = inData.data.dark()
flat = inData.data.flat()
if dark.size:
dark = np.pad(inData.data.dark(), pad_list, mode='edge')
dark = self._process_calibration_frames(dark)
inData.data.update_dark(dark[self.pad:-self.pad])
if flat.size:
flat = np.pad(inData.data.flat(), pad_list, mode='edge')
flat = self._process_calibration_frames(flat)
inData.data.update_flat(flat[self.pad:-self.pad])
def _process_calibration_frames(self, data):
nSlices = data.shape[self.proj_dim] - 2*self.pad
nSublists = int(np.ceil(nSlices/float(self.frame_limit)))
idx = np.array_split(np.arange(self.pad, nSlices+self.pad), nSublists)
idx = [np.arange(a[0]-self.pad, a[-1]+self.pad+1) for a in idx]
out_sl = np.tile([slice(None)]*3, [len(idx), 1])
out_sl[:, self.proj_dim] = idx
result = np.empty_like(data)
for sl in out_sl:
result[tuple(sl)] = self._dezing(data[tuple(sl)])
return result
def _dezing(self, data):
result = data[...]
indices = np.where(np.isnan(result))
result[indices] = 0.0
std_dev = np.std(result)
result = MEDIAN_DEZING_GPU(result.copy(order='C'), self.parameters['kernel_size'], std_dev*self.parameters['outlier_mu'])
return result
[docs] def process_frames(self, data):
input_temp = data[0]
indices = np.where(np.isnan(input_temp))
input_temp[indices] = 0.0
std_dev = np.std(input_temp)
result = MEDIAN_DEZING_GPU(input_temp.copy(order='C'), self.parameters['kernel_size'], std_dev*self.parameters['outlier_mu'])
return result
[docs] def get_max_frames(self):
""" Setting nFrames to multiple with an upper limit of 4 frames. """
return ['multiple', self.frame_limit]
[docs] def raw_data(self):
return True
[docs] def set_filter_padding(self, in_data, out_data):
# kernel size must be odd
ksize = self.parameters['kernel_size']
self.kernel_size = ksize+1 if ksize % 2 == 0 else ksize
in_data = in_data[0]
self.pad = (self.kernel_size - 1) // 2
self.data_size = in_data.get_shape()
in_data.padding = {'pad_multi_frames': self.pad}
out_data[0].padding = {'pad_multi_frames': self.pad}