colorthief-dataset 0.1.0

Static xkcd color-hierarchy table with pre-computed LAB used by `colorthief` for human-vocabulary color naming.
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

//! x86 SSE4.1 CIE94 backend — 4 entries/iter via 128-bit
//! `_mm_loadu_ps` against the SoA `LABS_*` arrays.
//!
//! Mirrors the [`super::cie94_aarch64_neon`] shape exactly.

#![allow(unsafe_code, dead_code)]

use core::arch::x86_64::*;

use libm::sqrtf;

use super::{LABS_A, LABS_B, LABS_C, LABS_L};

/// CIE94 nearest-neighbor scan (SSE4.1).
///
/// # Safety
///
/// Caller must guarantee SSE4.1 is available at runtime.
#[target_feature(enable = "sse4.1")]
pub unsafe fn nearest_idx(query: [f32; 3]) -> usize {
  let l2 = _mm_set1_ps(query[0]);
  let a2 = _mm_set1_ps(query[1]);
  let b2 = _mm_set1_ps(query[2]);
  // C₂ = sqrt(a₂² + b₂²). Computed once, reused per chunk.
  let c2_sq = _mm_add_ps(_mm_mul_ps(a2, a2), _mm_mul_ps(b2, b2));
  let c2_v = _mm_sqrt_ps(c2_sq);

  let n = LABS_L.len();
  let chunks = n / 4;

  let mut best_d2 = f32::INFINITY;
  let mut best_idx: usize = 0;

  let l_ptr = LABS_L.as_ptr();
  let a_ptr = LABS_A.as_ptr();
  let b_ptr = LABS_B.as_ptr();
  let c_ptr = LABS_C.as_ptr();

  let zero = _mm_setzero_ps();
  let one = _mm_set1_ps(1.0);
  let kc = _mm_set1_ps(0.045);
  let kh = _mm_set1_ps(0.015);

  for chunk in 0..chunks {
    let i = chunk * 4;

    // SAFETY: chunks = n / 4 floor; last load is at offset
    // (chunks - 1) * 4, at most n - 4.
    let l1 = unsafe { _mm_loadu_ps(l_ptr.add(i)) };
    let a1 = unsafe { _mm_loadu_ps(a_ptr.add(i)) };
    let b1 = unsafe { _mm_loadu_ps(b_ptr.add(i)) };
    let c1 = unsafe { _mm_loadu_ps(c_ptr.add(i)) };

    let dl = _mm_sub_ps(l1, l2);
    let da = _mm_sub_ps(a1, a2);
    let db = _mm_sub_ps(b1, b2);

    let dc = _mm_sub_ps(c1, c2_v);
    let dab_sq = _mm_add_ps(_mm_mul_ps(da, da), _mm_mul_ps(db, db));
    let dc_sq = _mm_mul_ps(dc, dc);
    let dh_sq = _mm_max_ps(_mm_sub_ps(dab_sq, dc_sq), zero);

    // S_C = 1 + 0.045·C₁; S_H = 1 + 0.015·C₁; S_L = 1 (so dl_term = dl).
    let sc = _mm_add_ps(one, _mm_mul_ps(kc, c1));
    let sh = _mm_add_ps(one, _mm_mul_ps(kh, c1));

    let dl_sq = _mm_mul_ps(dl, dl);
    let dc_term = _mm_div_ps(dc, sc);
    let dc_term_sq = _mm_mul_ps(dc_term, dc_term);
    let sh_sq = _mm_mul_ps(sh, sh);
    let dh_term_sq = _mm_div_ps(dh_sq, sh_sq);

    let d2 = _mm_add_ps(_mm_add_ps(dl_sq, dc_term_sq), dh_term_sq);

    let mut buf = [0f32; 4];
    // SAFETY: 16-byte write into a 16-byte buffer.
    unsafe { _mm_storeu_ps(buf.as_mut_ptr(), d2) };
    for (lane, d) in buf.iter().enumerate() {
      if *d < best_d2 {
        best_d2 = *d;
        best_idx = i + lane;
      }
    }
  }

  // Tail.
  let c2_scalar = sqrtf(query[1] * query[1] + query[2] * query[2]);
  for i in (chunks * 4)..n {
    let l1 = LABS_L[i];
    let a1 = LABS_A[i];
    let b1 = LABS_B[i];
    let c1 = LABS_C[i];
    let dl = l1 - query[0];
    let da = a1 - query[1];
    let db = b1 - query[2];
    let dc = c1 - c2_scalar;
    let dh_sq = (da * da + db * db - dc * dc).max(0.0);
    let sc = 1.0 + 0.045 * c1;
    let sh = 1.0 + 0.015 * c1;
    let d2 = dl * dl + (dc / sc) * (dc / sc) + dh_sq / (sh * sh);
    if d2 < best_d2 {
      best_d2 = d2;
      best_idx = i;
    }
  }

  best_idx
}