rscolorq 0.2.0

Spatial color quantization, a Rust port of `scolorq`.
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

use core::marker::PhantomData;

use crate::{quant::FilterSize, Matrix2d, QuantError, SpatialQuant};
use palette::{white_point::D65, Lab};

impl SpatialQuant for Lab<D65, f64> {
    fn calculate_filter_weights(dithering_level: f64, filter_size: FilterSize) -> Matrix2d<Self> {
        let mut sum = 0.0;
        match filter_size {
            FilterSize::One => Matrix2d::from_vec(vec![Self::one()], 1, 1),
            FilterSize::Three => {
                let mut filter_weights = Matrix2d::new(3, 3);
                for (j, fw_chunk) in filter_weights.rows_mut().enumerate() {
                    for (i, fw_entry) in fw_chunk.iter_mut().enumerate() {
                        let val = (-f64::from(i as i8 - 1).hypot(f64::from(j as i8 - 1))
                            / dithering_level.powi(2))
                        .exp();
                        sum += val;
                        *fw_entry = Self::new(val, val, val);
                    }
                }
                filter_weights.iter_mut().for_each(|fw| *fw /= sum);
                filter_weights
            }
            FilterSize::Five => {
                let mut filter_weights = Matrix2d::new(5, 5);
                for (j, fw_chunk) in filter_weights.rows_mut().enumerate() {
                    for (i, fw_entry) in fw_chunk.iter_mut().enumerate() {
                        let val = (-f64::from(i as i8 - 2).hypot(f64::from(j as i8 - 2))
                            / dithering_level.powi(2))
                        .exp();
                        sum += val;
                        *fw_entry = Self::new(val, val, val);
                    }
                }
                filter_weights.iter_mut().for_each(|fw| *fw /= sum);
                filter_weights
            }
        }
    }

    fn color_difference(&self, other: &Self) -> f64 {
        (*self - *other).norm_squared()
    }

    fn difference_threshold() -> f64 {
        // Using the 1976 delta E*_ab for just noticeable difference,
        // 2.3 seemed too high to produce good results
        1.0
    }

    fn direct_product(&self, other: &Self) -> Self {
        Self {
            l: self.l * other.l,
            a: self.a * other.a,
            b: self.b * other.b,
            white_point: PhantomData,
        }
    }

    fn dot_product(&self, other: &Self) -> f64 {
        self.l * other.l + self.a * other.a + self.b * other.b
    }

    fn one() -> Self {
        Self::new(1.0, 1.0, 1.0)
    }

    fn norm_squared(&self) -> f64 {
        self.l * self.l + self.a * self.a + self.b * self.b
    }

    fn random(rng: &mut impl rand::Rng) -> Self {
        Self {
            l: rng.gen_range(0.0..=100.0),
            a: rng.gen_range(-128.0..=127.0),
            b: rng.gen_range(-128.0..=127.0),
            white_point: PhantomData,
        }
    }

    fn refine_palette(
        s: &mut Matrix2d<Self>,
        coarse_variables: &crate::Matrix3d<f64>,
        a: &Matrix2d<Self>,
        palette: &mut Vec<Self>,
    ) -> Result<(), QuantError> {
        // Reflect s above the diagonal
        for v in 0..s.width() {
            for alpha in 0..v {
                *s.get_mut(v, alpha).ok_or("Could not reflect s")? =
                    *s.get(alpha, v).ok_or("Could not reflect s")?;
            }
        }

        let mut r: Vec<Self> = (0..palette.len()).map(|_| Self::default()).collect();
        for (v, r_item) in r.iter_mut().enumerate().take(palette.len()) {
            for (i_y, a_chunk) in a.rows().enumerate() {
                for (i_x, &a_entry) in a_chunk.iter().enumerate() {
                    *r_item += a_entry
                        * *coarse_variables
                            .get(i_x, i_y, v)
                            .ok_or("Could not access coarse variables to refine palette")?;
                }
            }
        }

        // Iterate through each of the component channels, clamp to valid range
        // Lightness
        let s_k = Matrix2d::from_vec(s.iter().map(|a| a.l).collect(), s.width(), s.height());
        let r_k = r.iter().map(|a| a.l).collect::<Vec<_>>();
        let palette_channel = (((s_k * 2.0).matrix_inverse()?) * -1.0) * r_k;
        for (v, pal) in palette.iter_mut().zip(&palette_channel) {
            v.l = pal.min(100.0).max(0.0);
        }
        // a star
        let s_k = Matrix2d::from_vec(s.iter().map(|a| a.a).collect(), s.width(), s.height());
        let r_k = r.iter().map(|a| a.a).collect::<Vec<_>>();
        let palette_channel = (((s_k * 2.0).matrix_inverse()?) * -1.0) * r_k;
        for (v, pal) in palette.iter_mut().zip(&palette_channel) {
            v.a = pal.min(127.0).max(-128.0);
        }
        // b star
        let s_k = Matrix2d::from_vec(s.iter().map(|a| a.b).collect(), s.width(), s.height());
        let r_k = r.iter().map(|a| a.b).collect::<Vec<_>>();
        let palette_channel = (((s_k * 2.0).matrix_inverse()?) * -1.0) * r_k;
        for (v, pal) in palette.iter_mut().zip(&palette_channel) {
            v.b = pal.min(127.0).max(-128.0);
        }

        Ok(())
    }
}