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

/*!
# Gray Level Run Length Matrix (GLRLM).

This module contains generation of GLRLM and features that can be extracted from it.

## Features

 */

use tch::{Tensor, index::*, Kind};

/**
Generate GLRLM from an image.

# Arguments
- `image`: [N, 1, H, W] grayscale tensor of the image with values in [0, 1].
- `num_levels`: number of gray levels to use. From 2 to 254 with a mask and From 2 to 255 otherwise.
- `max_run_length`: maximum run length to use.
- `direction` : a tuple (dx, dy) of the direction of the run length.
- `normalize`: whether to normalize the GLRLM.
- `mask`: [N, 1, H, W] tensor of the mask with values in [0, 1]. 

# Returns
- [N, num_levels, max_run_length] GLRLM tensor.
 */
pub fn glrlm(
    image: &Tensor,
    num_levels: u8,
    max_run_length: i64,
    direction: (i64, i64),
    mask: Option<&Tensor>,
) -> Tensor {
    let (batch_size, _, height, width) = image.size4().unwrap();
    let (dx, dy) = direction;
    assert!(num_levels >= 2 && num_levels <= 254, "num_levels must be in the range [2, 254]");
    assert!(max_run_length >= 1, "max_run_length must be at least 1");
    assert!(dx.abs() < 2 && dy.abs() < 2, "dx and dy must be in the range [-1, 1]");
    assert!(dx != 0 || dy != 0, "dx and dy cannot both be 0");
    let min = f32::from(image.min());
    let max = f32::from(image.max());
    let kind = image.kind();
    assert!(min >= 0.0 && max <= 1.0 && kind == Kind::Float, "image must be float in the range [0, 1]");

    // map the image to the range [0, num_levels - 1]
    let mut image = (image * (num_levels as f64 - 1e-6)).floor();
    // If we use a mask we add 255 to the masked pixels so that they are not counted in the GLCM.
    if let Some(mask) = mask {
        assert!(mask.size() == image.size(), "mask must have the same size as image");
        let mask = mask.to_kind(Kind::Float);
        let mask = (mask - 1.0) * - 255.0;
        image += mask;
    }
    let image = image.clamp(0.0, 255.0).to_kind(tch::Kind::Uint8);
    
    // We create a mask where the pixels are 1 if they aren't the first pixel of a run.
    let mask = {
        let conv = Tensor::zeros(&[3, 3], (tch::Kind::Float, image.device()));
        drop(conv.i((1, 1)).fill_(1.0));
        drop(conv.i((1 - dy, 1 - dx)).fill_(-1.0));
        let conv = conv.view([1, 1, 3, 3]);
        let image = image.to_kind(Kind::Float).conv2d::<&Tensor>(&conv, None, &[1, 1], &[1, 1], &[1, 1], 1);
        
        image.eq(0)
    };


    // Computing the same shade run length 
    let run_length = Tensor::ones_like(&image);
    let mut dest_slice = run_length.i((.., .., dy.max(0)..(height + dy).min(height), dx.max(0)..(width + dx).min(width)));
    let neigh_slice = run_length.i((.., .., (-dy).max(0)..(height - dy).min(height), (-dx).max(0)..(width - dx).min(width)));
    let mask_slice = mask.i((.., .., dy.max(0)..(height + dy).min(height), dx.max(0)..(width + dx).min(width)));
    for _ in 0..max_run_length{
        dest_slice.copy_(&(&neigh_slice * &mask_slice + 1));
    }
    
    // Generate a mask that only is true for the furthers point of a gray run 
    let mask = {
        let slice = mask.i((
            ..,
            .., 
            dy.max(0)..((height + dy).min(height)), 
            dx.max(0)..((width + dx).min(width))
        ));
        let slice = slice.pad(&[(-dx).max(0), dx.max(0), (-dy).max(0), dy.max(0)], "constant", Some(0.0)).to_kind(Kind::Int);
        -slice + 1.0
    };

    let run_length = (run_length * mask).clamp(0, max_run_length).to_kind(Kind::Int);
    
    let glrlm = Tensor::zeros(&[batch_size, num_levels as i64, max_run_length], (Kind::Int64, image.device()));
    for level in 0..num_levels{
        let lmask = image.eq(level as i64);
        for len in 1..=max_run_length{
            let rmask = run_length.eq(len);
            let count = (&lmask * rmask).sum_dim_intlist(Some(&[1, 2, 3][..]), false, Kind::Int);
            glrlm.i((.., level as i64, len - 1)).copy_(&count);
        }
    }
    glrlm
}

#[cfg(test)]
mod test {
    use tch::{Tensor, Kind, Device};
    use crate::{glrlm::glrlm, utils::assert_eq_tensor};

    #[test]
    fn test_glrlm_0deg(){
        let image = Tensor::of_slice(&[
            2.0, 0.0, 0.0, 2.0, 0.0, 1.0, 1.0, 1.0, 3.0, 0.0, 0.0, 2.0, 0.0, 3.0, 3.0,
            2.0, 0.0, 0.0, 0.0, 2.0, 0.0, 3.0, 3.0, 3.0, 2.0, 1.0, 2.0, 1.0, 1.0, 2.0,
            3.0, 0.0, 2.0, 0.0, 0.0, 0.0, 0.0, 3.0, 1.0, 1.0, 2.0, 0.0, 0.0, 3.0, 3.0,
        ]).view([1, 1, 3, 15]).to_kind(Kind::Float);
        let expected = Tensor::of_slice(&[
            4, 3, 1, 1,
            1, 2, 1, 0,
            10, 0, 0, 1,
            3, 2, 1, 0,
        ]).view([1, 4, 4]);
        let image = image / 4.0;
        let glrlm = glrlm(&image, 4, 4, (1, 0), None);
        assert_eq_tensor(&glrlm, &expected);
    }

    #[test]
    fn test_glrlm_45deg(){
        let image = Tensor::of_slice(&[
            1, 0, 1, 0, 0, 1, 1, 1, 0, 1,
            1, 0, 1, 1, 1, 1, 1, 0, 1, 0,
            1, 0, 1, 1, 0, 1, 1, 0, 1, 1,
            1, 0, 0, 1, 1, 0, 1, 1, 1, 0,
            0, 1, 0, 1, 1, 1, 0, 1, 1, 1,
            1, 0, 0, 1, 1, 1, 1, 1, 1, 0,
            1, 0, 1, 0, 0, 0, 0, 1, 0, 1,
            0, 1, 0, 1, 1, 1, 0, 0, 0, 0,
            0, 1, 0, 1, 1, 0, 1, 1, 1, 1,
            0, 0, 0, 0, 1, 1, 0, 0, 1, 0]).to_kind(Kind::Float).view([1, 1, 10, 10]);
        let glrlm = glrlm(&(&image/2.0), 2, 4, (1, 1), None);
        let expected = Tensor::of_slice(&[
            15, 12, 1, 0,
            13, 9, 1, 5,
        ]).view([1, 2, 4]);

        assert_eq_tensor(&glrlm, &expected);
    }

    #[test]
    fn test_glrlm_90deg(){
        let image = Tensor::of_slice(&[
            0, 1, 1, 1, 1, 1, 1, 1, 0, 0,
            0, 0, 0, 0, 0, 1, 0, 1, 1, 1,
            1, 0, 1, 0, 0, 0, 0, 0, 0, 1,
            1, 1, 1, 0, 1, 0, 1, 1, 0, 0,
            1, 1, 1, 0, 0, 1, 1, 1, 0, 1,
            0, 1, 1, 1, 0, 1, 1, 1, 1, 1,
            0, 1, 0, 1, 1, 1, 1, 0, 1, 1,
            1, 0, 0, 0, 0, 0, 1, 1, 0, 0,
            1, 0, 0, 1, 1, 0, 0, 0, 1, 0,
            1, 1, 1, 1, 1, 1, 1, 1, 0, 0,
        ]).view([1, 1, 10, 10]);
        let expect = Tensor::of_slice(&[
            12, 9, 3, 1,
            15, 7, 5, 3
        ]).view([1, 2, 4]);
        let glrlm = glrlm(&(&image/2.0), 2, 4, (0, 1), None);
        assert_eq_tensor(&glrlm, &expect);
    }

    const N: i64 = 100;
    #[test]
    fn sanity_check(){
        let mut image = Tensor::zeros(&[N, 1, 10, 1000], (Kind::Float, Device::Cpu));
        drop(image.uniform_(0.0, 99.9999));
        image/=100.0;
        let glrlm = glrlm(&(&image/2.0), 100, 10, (0, 1), None);
        // Check that the sum of all the elements is equal to the number of pixels
        let i = Tensor::arange(10, (Kind::Int64, Device::Cpu)) + 1;
        let i = i.view([1, -1]).repeat(&[100, 1]);
        assert_eq!(i64::from((glrlm * i).sum(Kind::Int64)), N * 10_000, "The sum of all longest run lengths should be equal to the number of pixels");
    }
}