-
Notifications
You must be signed in to change notification settings - Fork 476
/
lploc.js
119 lines (105 loc) · 3.51 KB
/
lploc.js
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
/* This library is released under the MIT license, see https://github.com/nenadmarkus/picojs */
lploc = {}
lploc.unpack_localizer = function(bytes)
{
//
const dview = new DataView(new ArrayBuffer(4));
let p = 0;
/*
read the number of stages, scale multiplier (applied after each stage),
number of trees per stage and depth of each tree
*/
dview.setUint8(0, bytes[p+0]), dview.setUint8(1, bytes[p+1]), dview.setUint8(2, bytes[p+2]), dview.setUint8(3, bytes[p+3]);
const nstages = dview.getInt32(0, true);
p = p + 4;
dview.setUint8(0, bytes[p+0]), dview.setUint8(1, bytes[p+1]), dview.setUint8(2, bytes[p+2]), dview.setUint8(3, bytes[p+3]);
const scalemul = dview.getFloat32(0, true);
p = p + 4;
dview.setUint8(0, bytes[p+0]), dview.setUint8(1, bytes[p+1]), dview.setUint8(2, bytes[p+2]), dview.setUint8(3, bytes[p+3]);
const ntreesperstage = dview.getInt32(0, true);
p = p + 4;
dview.setUint8(0, bytes[p+0]), dview.setUint8(1, bytes[p+1]), dview.setUint8(2, bytes[p+2]), dview.setUint8(3, bytes[p+3]);
const tdepth = dview.getInt32(0, true);
p = p + 4;
/*
unpack the trees
*/
const tcodes_ls = [];
const tpreds_ls = [];
for(let i=0; i<nstages; ++i)
{
// read the trees for this stage
for(let j=0; j<ntreesperstage; ++j)
{
// binary tests (we can read all of them at once)
Array.prototype.push.apply(tcodes_ls, bytes.slice(p, p+4*Math.pow(2, tdepth)-4));
p = p + 4*Math.pow(2, tdepth)-4;
// read the prediction in the leaf nodes of the tree
for(let k=0; k<Math.pow(2, tdepth); ++k)
for(let l=0; l<2; ++l)
{
dview.setUint8(0, bytes[p+0]), dview.setUint8(1, bytes[p+1]), dview.setUint8(2, bytes[p+2]), dview.setUint8(3, bytes[p+3]);
tpreds_ls.push(dview.getFloat32(0, true));
p = p + 4;
}
}
}
const tcodes = new Int8Array(tcodes_ls);
const tpreds = new Float32Array(tpreds_ls);
/*
construct the location estimaton function
*/
function loc_fun(r, c, s, pixels, nrows, ncols, ldim)
{
let root = 0;
const pow2tdepth = Math.pow(2, tdepth) >> 0; // '>>0' transforms this number to int
for(let i=0; i<nstages; ++i)
{
let dr=0.0, dc=0.0;
for(let j=0; j<ntreesperstage; ++j)
{
let idx = 0;
for(var k=0; k<tdepth; ++k)
{
const r1 = Math.min(nrows-1, Math.max(0, (256*r+tcodes[root + 4*idx + 0]*s)>>8));
const c1 = Math.min(ncols-1, Math.max(0, (256*c+tcodes[root + 4*idx + 1]*s)>>8));
const r2 = Math.min(nrows-1, Math.max(0, (256*r+tcodes[root + 4*idx + 2]*s)>>8));
const c2 = Math.min(ncols-1, Math.max(0, (256*c+tcodes[root + 4*idx + 3]*s)>>8));
idx = 2*idx + 1 + (pixels[r1*ldim+c1] > pixels[r2*ldim+c2])
}
const lutidx = 2*(ntreesperstage*pow2tdepth*i + pow2tdepth*j + idx - (pow2tdepth - 1))
dr += tpreds[lutidx + 0];
dc += tpreds[lutidx + 1];
root += 4*pow2tdepth - 4;
}
r = r + dr*s;
c = c + dc*s;
s = s*scalemul;
}
return [r, c];
}
/*
this function applies random perturbations to the default rectangle (r, c, s)
*/
function loc_fun_with_perturbs(r, c, s, nperturbs, image)
{
const rows=[], cols=[];
for(let i=0; i<nperturbs; ++i)
{
const _s = s*(0.925 + 0.15*Math.random());
let _r = r + s*0.15*(0.5 - Math.random());
let _c = c + s*0.15*(0.5 - Math.random());
[_r, _c] = loc_fun(_r, _c, _s, image.pixels, image.nrows, image.ncols, image.ldim)
rows.push(_r)
cols.push(_c)
}
// return the median along each axis
rows.sort()
cols.sort()
return [rows[Math.round(nperturbs/2)], cols[Math.round(nperturbs/2)]];
}
/*
we're done
*/
return loc_fun_with_perturbs;
}