remove-howto (#19139)
This commit is contained in:
Sebastian Golebiewski
GitHub
parent
9fd8a13fe6
commit
260273932a
@ -1,3 +0,0 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:f7c8ab4f15874d235968471bcf876c89c795d601e69891208107b8b72aa58eb1
|
||||
size 70014
|
||||
@ -1,3 +0,0 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:3d5ccf51fe1babb93d96d042494695a6a6e055d1f8ebf7eef5083d54d8987a23
|
||||
size 58789
|
||||
@ -1,40 +0,0 @@
|
||||
# Copyright (C) 2018-2023 Intel Corporation
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
|
||||
#! [complex:transformation]
|
||||
|
||||
from openvino.tools.mo.front.common.replacement import FrontReplacementSubgraph
|
||||
from openvino.tools.mo.graph.graph import Graph
|
||||
|
||||
|
||||
class Complex(FrontReplacementSubgraph):
|
||||
enabled = True
|
||||
|
||||
def pattern(self):
|
||||
return dict(
|
||||
nodes=[
|
||||
('strided_slice_real', dict(op='StridedSlice')),
|
||||
('strided_slice_imag', dict(op='StridedSlice')),
|
||||
('complex', dict(op='Complex')),
|
||||
],
|
||||
edges=[
|
||||
('strided_slice_real', 'complex', {'in': 0}),
|
||||
('strided_slice_imag', 'complex', {'in': 1}),
|
||||
])
|
||||
|
||||
@staticmethod
|
||||
def replace_sub_graph(graph: Graph, match: dict):
|
||||
strided_slice_real = match['strided_slice_real']
|
||||
strided_slice_imag = match['strided_slice_imag']
|
||||
complex_node = match['complex']
|
||||
|
||||
# make sure that both strided slice operations get the same data as input
|
||||
assert strided_slice_real.in_port(0).get_source() == strided_slice_imag.in_port(0).get_source()
|
||||
|
||||
# identify the output port of the operation producing datat for strided slice nodes
|
||||
input_node_output_port = strided_slice_real.in_port(0).get_source()
|
||||
input_node_output_port.disconnect()
|
||||
|
||||
# change the connection so now all consumers of "complex_node" get data from input node of strided slice nodes
|
||||
complex_node.out_port(0).get_connection().set_source(input_node_output_port)
|
||||
#! [complex:transformation]
|
||||
@ -1,27 +0,0 @@
|
||||
# Copyright (C) 2018-2023 Intel Corporation
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
|
||||
#! [complex_abs:transformation]
|
||||
import numpy as np
|
||||
|
||||
from openvino.tools.mo.ops.elementwise import Pow
|
||||
from openvino.tools.mo.ops.ReduceOps import ReduceSum
|
||||
from openvino.tools.mo.front.common.replacement import FrontReplacementOp
|
||||
from openvino.tools.mo.graph.graph import Graph, Node
|
||||
from openvino.tools.mo.ops.const import Const
|
||||
|
||||
|
||||
class ComplexAbs(FrontReplacementOp):
|
||||
op = "ComplexAbs"
|
||||
enabled = True
|
||||
|
||||
def replace_op(self, graph: Graph, node: Node):
|
||||
pow_2 = Const(graph, {'value': np.float32(2.0)}).create_node()
|
||||
reduce_axis = Const(graph, {'value': np.int32(-1)}).create_node()
|
||||
pow_0_5 = Const(graph, {'value': np.float32(0.5)}).create_node()
|
||||
|
||||
sq = Pow(graph, dict(name=node.in_node(0).name + '/sq', power=2.0)).create_node([node.in_node(0), pow_2])
|
||||
sum = ReduceSum(graph, dict(name=sq.name + '/sum')).create_node([sq, reduce_axis])
|
||||
sqrt = Pow(graph, dict(name=sum.name + '/sqrt', power=0.5)).create_node([sum, pow_0_5])
|
||||
return [sqrt.id]
|
||||
#! [complex_abs:transformation]
|
||||
@ -1,33 +0,0 @@
|
||||
# Copyright (C) 2018-2023 Intel Corporation
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
|
||||
# ! [fft_ext:extractor]
|
||||
from ...ops.FFT import FFT
|
||||
from openvino.tools.mo.front.extractor import FrontExtractorOp
|
||||
|
||||
|
||||
class FFT2DFrontExtractor(FrontExtractorOp):
|
||||
op = 'FFT2D'
|
||||
enabled = True
|
||||
|
||||
@classmethod
|
||||
def extract(cls, node):
|
||||
attrs = {
|
||||
'inverse': 0
|
||||
}
|
||||
FFT.update_node_stat(node, attrs)
|
||||
return cls.enabled
|
||||
|
||||
|
||||
class IFFT2DFrontExtractor(FrontExtractorOp):
|
||||
op = 'IFFT2D'
|
||||
enabled = True
|
||||
|
||||
@classmethod
|
||||
def extract(cls, node):
|
||||
attrs = {
|
||||
'inverse': 1
|
||||
}
|
||||
FFT.update_node_stat(node, attrs)
|
||||
return cls.enabled
|
||||
# ! [fft_ext:extractor]
|
||||
@ -1,27 +0,0 @@
|
||||
# Copyright (C) 2018-2023 Intel Corporation
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
|
||||
#! [fft:operation]
|
||||
from openvino.tools.mo.front.common.partial_infer.elemental import copy_shape_infer
|
||||
from openvino.tools.mo.graph.graph import Graph
|
||||
from openvino.tools.mo.ops.op import Op
|
||||
|
||||
|
||||
class FFT(Op):
|
||||
op = 'FFT'
|
||||
enabled = False
|
||||
|
||||
def __init__(self, graph: Graph, attrs: dict):
|
||||
super().__init__(graph, {
|
||||
'type': self.op,
|
||||
'op': self.op,
|
||||
'version': 'custom_opset',
|
||||
'inverse': None,
|
||||
'in_ports_count': 1,
|
||||
'out_ports_count': 1,
|
||||
'infer': copy_shape_infer
|
||||
}, attrs)
|
||||
|
||||
def backend_attrs(self):
|
||||
return ['inverse']
|
||||
#! [fft:operation]
|
||||
@ -1,106 +0,0 @@
|
||||
# Copyright (C) 2018-2023 Intel Corporation
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
|
||||
#! [mri_demo:demo]
|
||||
import numpy as np
|
||||
import cv2 as cv
|
||||
import argparse
|
||||
import time
|
||||
from openvino.inference_engine import IECore
|
||||
|
||||
|
||||
def kspace_to_image(kspace):
|
||||
assert(len(kspace.shape) == 3 and kspace.shape[-1] == 2)
|
||||
fft = cv.idft(kspace, flags=cv.DFT_SCALE)
|
||||
img = cv.magnitude(fft[:,:,0], fft[:,:,1])
|
||||
return cv.normalize(img, dst=None, alpha=255, beta=0, norm_type=cv.NORM_MINMAX, dtype=cv.CV_8U)
|
||||
|
||||
|
||||
if __name__ == '__main__':
|
||||
parser = argparse.ArgumentParser(description='MRI reconstrution demo for network from https://github.com/rmsouza01/Hybrid-CS-Model-MRI (https://arxiv.org/abs/1810.12473)')
|
||||
parser.add_argument('-i', '--input', dest='input', help='Path to input .npy file with MRI scan data.')
|
||||
parser.add_argument('-p', '--pattern', dest='pattern', help='Path to sampling mask in .npy format.')
|
||||
parser.add_argument('-m', '--model', dest='model', help='Path to .xml file of OpenVINO IR.')
|
||||
parser.add_argument('-l', '--cpu_extension', dest='cpu_extension', help='Path to extensions library with FFT implementation.')
|
||||
parser.add_argument('-d', '--device', dest='device', default='CPU',
|
||||
help='Optional. Specify the target device to infer on; CPU, '
|
||||
'GPU, GNA is acceptable. For non-CPU targets, '
|
||||
'HETERO plugin is used with CPU fallbacks to FFT implementation. '
|
||||
'Default value is CPU')
|
||||
args = parser.parse_args()
|
||||
|
||||
xml_path = args.model
|
||||
assert(xml_path.endswith('.xml'))
|
||||
bin_path = xml_path[:xml_path.rfind('.xml')] + '.bin'
|
||||
|
||||
ie = IECore()
|
||||
ie.add_extension(args.cpu_extension, "CPU")
|
||||
|
||||
net = ie.read_network(xml_path, bin_path)
|
||||
|
||||
device = 'CPU' if args.device == 'CPU' else ('HETERO:' + args.device + ',CPU')
|
||||
exec_net = ie.load_network(net, device)
|
||||
|
||||
# Hybrid-CS-Model-MRI/Data/stats_fs_unet_norm_20.npy
|
||||
stats = np.array([2.20295299e-01, 1.11048916e+03, 4.16997984e+00, 4.71741395e+00], dtype=np.float32)
|
||||
# Hybrid-CS-Model-MRI/Data/sampling_mask_20perc.npy
|
||||
var_sampling_mask = np.load(args.pattern) # TODO: can we generate it in runtime?
|
||||
print('Sampling ratio:', 1.0 - var_sampling_mask.sum() / var_sampling_mask.size)
|
||||
|
||||
data = np.load(args.input)
|
||||
num_slices, height, width = data.shape[0], data.shape[1], data.shape[2]
|
||||
pred = np.zeros((num_slices, height, width), dtype=np.uint8)
|
||||
data /= np.sqrt(height * width)
|
||||
|
||||
print('Compute...')
|
||||
start = time.time()
|
||||
for slice_id, kspace in enumerate(data):
|
||||
kspace = kspace.copy()
|
||||
|
||||
# Apply sampling
|
||||
kspace[var_sampling_mask] = 0
|
||||
kspace = (kspace - stats[0]) / stats[1]
|
||||
|
||||
# Forward through network
|
||||
input = np.expand_dims(kspace.transpose(2, 0, 1), axis=0)
|
||||
outputs = exec_net.infer(inputs={'input_1': input})
|
||||
output = next(iter(outputs.values()))
|
||||
output = output.reshape(height, width)
|
||||
|
||||
# Save predictions
|
||||
pred[slice_id] = cv.normalize(output, dst=None, alpha=255, beta=0, norm_type=cv.NORM_MINMAX, dtype=cv.CV_8U)
|
||||
|
||||
print('Elapsed time: %.1f seconds' % (time.time() - start))
|
||||
|
||||
WIN_NAME = 'MRI reconstruction with OpenVINO'
|
||||
|
||||
slice_id = 0
|
||||
def callback(pos):
|
||||
global slice_id
|
||||
slice_id = pos
|
||||
|
||||
kspace = data[slice_id]
|
||||
img = kspace_to_image(kspace)
|
||||
|
||||
kspace[var_sampling_mask] = 0
|
||||
masked = kspace_to_image(kspace)
|
||||
|
||||
rec = pred[slice_id]
|
||||
|
||||
# Add a header
|
||||
border_size = 20
|
||||
render = cv.hconcat((img, masked, rec))
|
||||
render = cv.copyMakeBorder(render, border_size, 0, 0, 0, cv.BORDER_CONSTANT, value=255)
|
||||
cv.putText(render, 'Original', (0, 15), cv.FONT_HERSHEY_SIMPLEX, 0.5, color=0)
|
||||
cv.putText(render, 'Sampled (PSNR %.1f)' % cv.PSNR(img, masked), (width, 15), cv.FONT_HERSHEY_SIMPLEX, 0.5, color=0)
|
||||
cv.putText(render, 'Reconstructed (PSNR %.1f)' % cv.PSNR(img, rec), (width*2, 15), cv.FONT_HERSHEY_SIMPLEX, 0.5, color=0)
|
||||
|
||||
cv.imshow(WIN_NAME, render)
|
||||
cv.waitKey(1)
|
||||
|
||||
cv.namedWindow(WIN_NAME, cv.WINDOW_NORMAL)
|
||||
print(num_slices)
|
||||
cv.createTrackbar('Slice', WIN_NAME, num_slices // 2, num_slices - 1, callback)
|
||||
callback(num_slices // 2) # Trigger initial visualization
|
||||
cv.waitKey()
|
||||
#! [mri_demo:demo]
|
||||
Loading…
Reference in New Issue
Block a user