cryoETparticles/CTF.py at main · picossio/cryoETparticles · GitHub

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
import matplotlib.pyplot as plt
import numpy as np
import scipy.constants as sc

class transfer:
    def __init__(self, keV, C_s, z, phi, n_pixels, pixel_size, d, sigma, sigma2, name, num_protein, x_center, y_center):

        self.C_s = C_s
        self.z = z
        self.phi = phi
        self.n_pixels = n_pixels
        self.pixel_size = pixel_size
        self.D = d
        self.sigma = sigma
        self.sigma2 = sigma2
        self.name = name
        self.x_center = x_center
        self.y_center = y_center
        self.num_protein = num_protein

        self.eV = keV * 1000
        self.lam = sc.h / np.sqrt(2 * sc.electron_mass * self.eV * sc.electron_volt) / np.sqrt(1 + self.eV * sc.electron_volt / (2 * sc.electron_mass * sc.speed_of_light ** 2)) * 10 ** 10   #wavelength in angstrom
        self.k = 1 / self.lam  # wavenumber in 1 / angstrom
        self.theta_0 = 0.16 * self.lam  # unitless number descibiing angular acceptance of aperature

        print('lambda = ' + str(self.lam))
        print('k = ' + str(self.k))
        print('theta_0= ' + str(self.theta_0))

    def Lens_effects(self):

        FT_phi = np.fft.fft2(self.phi)   # outputs N x N 2d fourier space with each dimension being a spatial frequency
                                         # spatial frequency is discretized in 1/pixel_size units
        xi = np.zeros(shape = (self.n_pixels, self.n_pixels))

        # for x in range(self.n_pixels):
        #     for y in range(self.n_pixels):

                # xi[x,y] = np.sqrt(((x - self.n_pixels / 2) / self.pixel_size) ** 2 + (((y - self.n_pixels / 2) / self.pixel_size)) ** 2)    # assigned absolute value of spatial frequency for each pixel in fourier space
        for x in range(int(self.n_pixels/2)):
            for y in range(int(self.n_pixels/2)):

                xi[x,y] = np.sqrt(((x) / self.pixel_size) ** 2 + (((y) / self.pixel_size)) ** 2)
                xi[self.n_pixels -1 - x, self.n_pixels - 1 - y] = np.sqrt(((x) / self.pixel_size) ** 2 + (((y) / self.pixel_size)) ** 2)
                xi[self.n_pixels -1 - x, y] = np.sqrt(((x) / self.pixel_size) ** 2 + (((y) / self.pixel_size)) ** 2)
                xi[x, self.n_pixels - 1 - y] = np.sqrt(((x) / self.pixel_size) ** 2 + (((y) / self.pixel_size)) ** 2)

        theta = xi / (self.k * self.D)

        chi = 1/2 * self.k * self.z * theta ** 2 + 1/8 * self.k * self.C_s * theta ** 4

        w_2 = 0.10  # constant descibing percentage of amplitude contrast in the image
        w_1 = np.sqrt(1 - w_2 ** 2)
        B = 1 #np.exp(- theta ** 2 / self.theta_0 ** 2)     # envelope function which descibes the decay of CTF amplitude

        self.CTF = B * (w_1 * np.sin(chi) - w_2 * np.cos(chi))
        self.phi_CTF = self.CTF * FT_phi

    def plot_imageCTF(self):

            fig, ax = plt.subplots()
            ax.imshow(self.CTF, cmap='gray')
            # plt.savefig(self.name + '_CTF.png')
            plt.show()

    def intensity(self):

        incident_wave = np.fft.ifft2(self.phi_CTF)
        self.wave_intensity = np.real(incident_wave)
        self.wave_intensity_range = np.clip(self.wave_intensity, -100,  0.01)

    def gaussian_noise_machine(self):

        rng = np.random.default_rng()
        g_noise = rng.normal(0,self.sigma, size = (self.n_pixels, self.n_pixels))
        g_noise = np.reshape(g_noise, (self.n_pixels, self.n_pixels))
        self.image_g_noise = self.wave_intensity + g_noise

    def gaussian_noise_machine2(self):

        rng = np.random.default_rng()
        g_noise = rng.normal(0,self.sigma2, size = (self.n_pixels, self.n_pixels))
        g_noise = np.reshape(g_noise, (self.n_pixels, self.n_pixels))
        self.image_g_noise2 = self.wave_intensity + g_noise

    def poisson_noise_machine(self):
        rng = np.random.default_rng()
        self.image_p_noise = np.zeros(shape=(self.n_pixels,self.n_pixels))
        minimum = np.min(self.wave_intensity)

        for x in range(self.n_pixels):
            for y in range(self.n_pixels):

                p_noise = rng.poisson(self.wave_intensity[x,y] - minimum)
                self.image_p_noise[x,y] = self.wave_intensity[x,y] + p_noise

    def output_data(self):

        file_path = '/Users/dkleebatt/VS_Code_projects/Micrograph_Generator/txt_data_files/' + self.name + '.txt'

        with open(file_path, mode='a') as f:

            f.write('#' + str(self.num_protein) + ' Proteins' + ' @ (x,y)' + '\n' + '\n')

            for i in range(int(self.num_protein)):
                f.write(str(self.x_center[i]) + ' ' + str(self.y_center[i]))
                f.write('\n')

            f.write('\n' + '# ' + 'x, ' + 'y, ' + 'incident wave, ' + 'CTF, ' + 'image, ' + 'image + gaussian, ' + 'image + gaussian2, ' + 'image + poisson' + '\n' + '\n')

        for x in range(int(self.n_pixels)):
                for y in range(int(self.n_pixels)):

                    with open(file_path, mode='a') as f:

                        f.write(str(x) + ' ' + str(y) + ' ' + str(self.phi[x,y]) + ' ' + str(self.CTF[x,y]) + ' ' + str(self.wave_intensity[x,y]) + ' ' + str(self.image_g_noise[x,y]) +  ' ' + str(self.image_g_noise2[x,y]) + ' ' + str(self.image_p_noise[x,y]))
                        f.write('\n')

    def plot_image_out(self):

        fig, ax = plt.subplots()
        ax.imshow(self.wave_intensity, cmap='gray')
        ax.set_title(self.name + ' CTF Convoluted Image')
        # plt.savefig(self.name + '_no_noise.png')
        plt.show()

    def plot_image_g_noise(self):

        fig,ax = plt.subplots()
        ax.imshow(self.image_g_noise, cmap='gray')
        ax.set_title(self.name + ' Gaussian Noise' + 'sigma = ' + str(self.sigma))
        plt.savefig(self.name + '_g_noise.png')
        plt.show()

    def plot_image_g_noise2(self):

        fig,ax = plt.subplots()
        ax.imshow(self.image_g_noise2, cmap='gray')
        ax.set_title(self.name + ' Gaussian Noise' + 'sigma = ' + str(self.sigma2))
        # plt.savefig(self.name + '_g_noise2.png')
        plt.show()

    def plot_image_p_noise(self):

        fig,ax = plt.subplots()
        ax.imshow(self.image_p_noise, cmap='gray')
        ax.set_title(self.name + ' Poisson Noise')
        # plt.savefig(self.name + '_p_noise.png')
        plt.show()

    def CTF_1d(self):

        self.frequency = np.zeros(self.n_pixels)

        for i in range (int(self.n_pixels)):

            self.frequency[i] = i / self.pixel_size

        theta = self.frequency / (self.k * self.D)
        chi = 1/2 * self.k * self.z * theta ** 2 + 1/8 * self.k * self.C_s * theta ** 4
        w_2 = 0.10
        w_1 = np.sqrt(1 - w_2 ** 2)
        B = np.exp(- theta ** 2 / self.theta_0 ** 2)

        self.CTF_1d = B * (w_1 * np.sin(chi) - w_2 * np.cos(chi))

        print(np.shape(self.plot_CTF_1d))

    def plot_CTF_1d(self):

        fig,ax = plt.subplots()
        ax.plot(self.frequency, self.CTF_1d)
        ax.set_title(self.name + ' 1D CTF' + ' z = ' + str(self.z ))
        ax.set_xlabel('Frequency')
        ax.set_ylabel('CTF')
        # plt.savefig(self.name + '_1d_CTF.png')
        plt.show()

    # definitions below this comment are not used in the current implementation

    def noise_machine(self):

        rng = np.random.default_rng()
        w_noise = rng.normal(0, self.sigma, size = (self.n_pixels, self.n_pixels))
        w_noise = np.reshape(w_noise, (self.n_pixels, self.n_pixels))
        FT_w_noise = np.fft.fft2(w_noise)

        rad_freq = np.zeros(shape = (self.n_pixels, self.n_pixels))
        f = np.zeros(shape = (self.n_pixels, self.n_pixels))

        for x in range(self.n_pixels):
            for y in range(self.n_pixels):

                rad_freq [x,y] = np.sqrt(((x - self.n_pixels / 2) / self.pixel_size) ** 2 + (((y - self.n_pixels / 2) / self.pixel_size)) ** 2)
                f[x,y] = 1 / np.sqrt((1 + rad_freq[x,y] ** 2))

        c_noise = f * FT_w_noise
        self.phi_CTF_noise = self.phi_CTF + c_noise

        print(np.shape(self.phi_CTF_noise))
        print(self.phi_CTF_noise)