imgal 0.2.0

A fast and open-source scientific image processing and algorithm library.
Documentation 
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

use ndarray::{
    Array3, ArrayBase, ArrayView1, ArrayViewMut3, AsArray, Axis, Ix1, Ix3, ViewRepr, Zip,
};
use rand::SeedableRng;
use rand::prelude::*;
use rand::rngs::StdRng;
use rand_distr::{Distribution, Poisson};
use rayon::prelude::*;

use crate::error::ImgalError;
use crate::traits::numeric::AsNumeric;

/// Apply Poisson noise on a 1-dimensional array.
///
/// # Description
///
/// Applies Poisson noise (_i.e._ shot noise) on a 1-dimensional array of data.
/// An element-wise lambda value (scaled by the `scale` parameter) is used to
/// simulate the Poisson noise with variable signal strength.
///
/// This function creates a new array and does not mutate the input array.
///
/// # Arguments
///
/// * `data`: The input 1-dimensional array.
/// * `scale`: The scale factor.
/// * `seed`: Pseudorandom number generator seed. Set the `seed` value to apply
///   homogenous noise to the input array. If `None`, then heterogenous noise
///   is applied to the input array.
///
/// # Returns
///
/// * `Vec<f64>`: A 1-dimensional array of the input data with Poisson noise
///   applied.
pub fn poisson_noise_1d<'a, T, A>(data: A, scale: f64, seed: Option<u64>) -> Vec<f64>
where
    A: AsArray<'a, T, Ix1>,
    T: 'a + AsNumeric,
{
    let view: ArrayBase<ViewRepr<&'a T>, Ix1> = data.into();
    let s = seed.unwrap_or(0);
    let mut rng = StdRng::seed_from_u64(s);
    let mut n_data = vec![0.0; view.len()];
    n_data.iter_mut().zip(view.iter()).for_each(|(n, &d)| {
        if d.to_f64() > 0.0 {
            let l: f64 = d.to_f64() * scale;
            let p = Poisson::new(l).unwrap();
            *n = p.sample(&mut rng);
        } else {
            *n = 0.0;
        }
    });

    n_data
}

/// Apply Poisson noise on a 1-dimensional array.
///
/// # Description
///
/// Applies Poisson noise (_i.e._ shot noise) on a 1-dimensional array of data.
/// An element-wise lambda value (scaled by the `scale` parameter) is used to
/// simulate the Poisson noise with variable signal strength.
///
/// This function mutates the input array and does not create a new array.
///
/// # Arguments
///
/// * `data`: The input 1-dimensional array view to mutate.
/// * `scale`: The scale factor.
/// * `seed`: Pseudorandom number generator seed. Set the `seed` value to apply
///   homogenous noise to the input array. If `None`, then heterogenous noise
///   is applied to the input array.
pub fn poisson_noise_1d_mut(data: &mut [f64], scale: f64, seed: Option<u64>) {
    let s = seed.unwrap_or(0);
    let mut rng = StdRng::seed_from_u64(s);
    data.iter_mut().for_each(|x| {
        if *x > 0.0 {
            let l = *x * scale;
            let p = Poisson::new(l).unwrap();
            *x = p.sample(&mut rng);
        } else {
            *x = 0.0;
        }
    });
}

/// Apply Poisson noise on a 3-dimensional array.
///
/// # Description
///
/// Applies Poisson noise (_i.e._ shot noise) on a 3-dimensional array of data.
/// An element-wise lambda value (scaled by the `scale` parameter) is used to
/// simulate Poisson noise with variable signal strength.
///
/// This function creates a new array and does not mutate the input array.
///
/// # Arguments
///
/// * `data`: The input 3-dimensional array.
/// * `scale`: The scale factor.
/// * `seed`: Pseudorandom number generator seed. Set the `seed` value to apply
///   homogenous noise to the input array. If `None`, then heterogenous noise
///   is applied to the input array.
/// * `axis`: The signal data axis. If `None`, then `axis = 2`.
///
/// # Returns
///
/// * `Ok(Array3<f64>)`: A 3-dimensional array of the input data with Poisson noise
///   applied.
/// * `Err(ImgalError)`: If `axis >= 3`.
pub fn poisson_noise_3d<'a, T, A>(
    data: A,
    scale: f64,
    seed: Option<u64>,
    axis: Option<usize>,
) -> Result<Array3<f64>, ImgalError>
where
    A: AsArray<'a, T, Ix3>,
    T: 'a + AsNumeric,
{
    // validate the axis value
    let a = axis.unwrap_or(2);
    if a >= 3 {
        return Err(ImgalError::InvalidAxis {
            axis_idx: a,
            dim_len: 3,
        });
    }

    // apply homogenous Poisson noise (with seed value set) or heterogenous
    // noise (no seed value set, each lane gets a new rng)
    let view: ArrayBase<ViewRepr<&'a T>, Ix3> = data.into();
    let shape = view.dim();
    let mut n_data = Array3::<f64>::zeros(shape);
    let src_lanes = view.lanes(Axis(a));
    let dst_lanes = n_data.lanes_mut(Axis(a));
    if let Some(s) = seed {
        Zip::from(src_lanes)
            .and(dst_lanes)
            .par_for_each(|s_ln, d_ln| {
                let mut rng = StdRng::seed_from_u64(s);
                Zip::from(s_ln).and(d_ln).for_each(|s, d| {
                    if (*s).to_f64() > 0.0 {
                        let l = (*s).to_f64() * scale;
                        let p = Poisson::new(l).unwrap();
                        *d = p.sample(&mut rng);
                    } else {
                        *d = 0.0;
                    }
                });
            });
    } else {
        Zip::from(src_lanes)
            .and(dst_lanes)
            .par_for_each(|s_ln, d_ln| {
                let mut rng = rand::rng();
                Zip::from(s_ln).and(d_ln).for_each(|s, d| {
                    if (*s).to_f64() > 0.0 {
                        let l = (*s).to_f64() * scale;
                        let p = Poisson::new(l).unwrap();
                        *d = p.sample(&mut rng);
                    } else {
                        *d = 0.0
                    }
                });
            });
    }

    Ok(n_data)
}

/// Apply Poisson noise on a 3-dimensional array.
///
/// # Description
///
/// Applies Poisson noise (_i.e._ shot noise) on a 3-dimensional array of data.
/// An element-wise lambda value (scaled by the `scale` parameter) is used to
/// simulate Poisson noise with variable signal strength.
///
/// This function mutates the input array and does not create a new array.
///
/// # Arguments
///
/// * `data`: The input 3-dimensional array to mutate.
/// * `scale`: The scale factor.
/// * `seed`: Pseudorandom number generator seed. Set the `seed` value to apply
///   homogenous noise to the input array. If `None`, then heterogenous noise
///   is applied to the input array.
/// * `axis`: The signal data axis. If `None`, then `axis = 2`.
pub fn poisson_noise_3d_mut(
    mut data: ArrayViewMut3<f64>,
    scale: f64,
    seed: Option<u64>,
    axis: Option<usize>,
) {
    // apply homogenous Poisson noise (with seed value set) or heterogenous
    // noise (no seed value set, each lane gets a new rng)
    let a = axis.unwrap_or(2);
    let lanes = data.lanes_mut(Axis(a));
    if let Some(s) = seed {
        lanes.into_iter().par_bridge().for_each(|mut ln| {
            if let Some(l) = ln.as_slice_mut() {
                poisson_noise_1d_mut(l, scale, Some(s));
            } else {
                let mut l = ln.to_vec();
                poisson_noise_1d_mut(&mut l, scale, Some(s));
                let l = ArrayView1::from(&l);
                ln.assign(&l);
            }
        });
    } else {
        lanes.into_iter().par_bridge().for_each(|mut ln| {
            let mut rng = rand::rng();
            let s = rng.next_u64();
            if let Some(l) = ln.as_slice_mut() {
                poisson_noise_1d_mut(l, scale, Some(s));
            } else {
                let mut l = ln.to_vec();
                poisson_noise_1d_mut(&mut l, scale, Some(s));
                let l = ArrayView1::from(&l);
                ln.assign(&l);
            }
        });
    }
}