This repository has been archived by the owner on Mar 20, 2024. It is now read-only.
forked from parallella/pal
-
Notifications
You must be signed in to change notification settings - Fork 0
/
p_median3x3.c
118 lines (107 loc) · 2.61 KB
/
p_median3x3.c
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
#include <pal.h>
#define SORT(a,b) \
do { \
if (a < b) { \
float d = a; \
a = b; \
b = d; \
} \
} while(0)
#define SORT_HI(a,b) \
do { \
b = (a < b) ? a : b; \
} while(0)
#define SORT_LO(a,b) \
do { \
a = (a < b) ? b : a; \
} while(0)
static __inline __attribute__((__always_inline__))
float my_median(
float v0, float v1, float v2,
float v3, float v4, float v5,
float v6, float v7, float v8);
/*
* A median 3x3 filter.
*
* @param x Pointer to input image, a 2D array of size 'rows' x 'cols'
* @param r Pointer to output image, a 2D array pf size 'rows-2' x 'cols-2'
* @param rows Number of rows in input image
* @param cols Number of columns in input image
* @return None
*/
void p_median3x3_f32(const float *x, float *r, int rows, int cols)
{
float v0,v1,v2,v3,v4,v5,v6,v7,v8;
const float *px = x;
float *pr = r;
int i, j;
for (i = 0; i < rows - 2; i++) {
v3 = *(px);
v4 = *(px + cols);
v5 = *(px + 2*cols);
SORT(v4,v5);
SORT(v3,v4);
SORT(v4,v5);
v6 = *(px + 1);
v7 = *(px + cols + 1);
v8 = *(px + 2*cols + 1);
SORT(v7,v8);
SORT(v6,v7);
SORT(v7,v8);
for (j = 0; j < cols - 2; j++) {
v0 = v3;
v1 = v4;
v2 = v5;
v3 = v6;
v4 = v7;
v5 = v8;
v6 = *(px + 2);
v7 = *(px + cols + 2);
v8 = *(px + 2*cols + 2);
SORT(v7,v8);
SORT(v6,v7);
SORT(v7,v8);
*(pr++) = my_median(v0,v1,v2,v3,v4,v5,v6,v7,v8);
px++;
}
px += 2;
}
}
/*
* A specialized median 3x3 filter.
*
* The routine is inlined for performance and requires that the v0-v2,
* v3-v5, and v6-v8 triplets be ordered from largest to smallest. This
* improves performance while sweeping through columns of values.
* Macros handle the swapping of values and do not swap values which
* won't be used again.
*
* @param v0 The largest value in the triplet v0-v2
* @param v1 The median value in the triplet v0-v2
* @param v2 The smallest value in the triplet v0-v2
* @param v3 The largest value in the triplet v3-v5
* @param v4 The median value in the triplet v3-v5
* @param v5 The smallest value in the triplet v3-v5
* @param v6 The largest value in the triplet v6-v8
* @param v7 The median value in the triplet v6-v8
* @param v8 The smallest value in the triplet v6-v8
* @return The median value of the set v0-v8
*/
static __inline __attribute((__always_inline__))
float my_median(
float v0, float v1, float v2,
float v3, float v4, float v5,
float v6, float v7, float v8)
{
SORT_HI(v0,v3);
SORT_LO(v5,v8);
SORT(v4,v7);
SORT_HI(v3,v6);
SORT_HI(v1,v4);
SORT_LO(v2,v5);
SORT_LO(v4,v7);
SORT(v4,v2);
SORT_HI(v6,v4);
SORT_LO(v4,v2);
return v4;
}