1 | /* -*- mode: c++; c-basic-offset: 4; indent-tabs-mode: nil -*- */ |
---|
2 | /* |
---|
3 | * Copyright (C) 2004-2013 HUBzero Foundation, LLC |
---|
4 | * |
---|
5 | */ |
---|
6 | |
---|
7 | #include <cstdlib> |
---|
8 | |
---|
9 | #include <vrmath/Vector3f.h> |
---|
10 | |
---|
11 | #include "ReaderCommon.h" |
---|
12 | #include "GradientFilter.h" |
---|
13 | |
---|
14 | using namespace nv; |
---|
15 | using namespace vrmath; |
---|
16 | |
---|
17 | float * |
---|
18 | nv::merge(float *scalar, float *gradient, int size) |
---|
19 | { |
---|
20 | float *data = (float *)malloc(sizeof(float) * 4 * size); |
---|
21 | |
---|
22 | Vector3f *g = (Vector3f *)gradient; |
---|
23 | |
---|
24 | int ngen = 0, sindex = 0; |
---|
25 | for (sindex = 0; sindex < size; ++sindex) { |
---|
26 | data[ngen++] = scalar[sindex]; |
---|
27 | data[ngen++] = 1.0 - g[sindex].x; |
---|
28 | data[ngen++] = 1.0 - g[sindex].y; |
---|
29 | data[ngen++] = 1.0 - g[sindex].z; |
---|
30 | } |
---|
31 | return data; |
---|
32 | } |
---|
33 | |
---|
34 | /** |
---|
35 | * \brief Normalize data to [0,1] based on vmin,vmax range |
---|
36 | * |
---|
37 | * Data outside of given range is clamped, and NaNs are set to |
---|
38 | * -1 in the output |
---|
39 | * |
---|
40 | * \param data Float array of unnormalized data, will be normalized on return |
---|
41 | * \param count Number of elts in array |
---|
42 | * \param stride Stride between values in data array |
---|
43 | * \param vmin Minimum value in data array |
---|
44 | * \param vmax Maximum value in data array |
---|
45 | */ |
---|
46 | void |
---|
47 | nv::normalizeScalar(float *data, int count, int stride, double vmin, double vmax) |
---|
48 | { |
---|
49 | double dv = vmax - vmin; |
---|
50 | dv = (dv == 0.0) ? 1.0 : dv; |
---|
51 | for (int pt = 0, i = 0; pt < count; ++pt, i += stride) { |
---|
52 | double v = data[i]; |
---|
53 | if (isnan(v)) { |
---|
54 | data[i] = -1.0f; |
---|
55 | } else if (v < vmin) { |
---|
56 | data[i] = 0.0f; |
---|
57 | } else if (v > vmax) { |
---|
58 | data[i] = 1.0f; |
---|
59 | } else { |
---|
60 | data[i] = (float)((v - vmin)/ dv); |
---|
61 | } |
---|
62 | } |
---|
63 | } |
---|
64 | |
---|
65 | /** |
---|
66 | * \brief Normalize data to [0,1] based on vmin,vmax range |
---|
67 | * |
---|
68 | * Data outside of given range is clamped, and NaNs are set to |
---|
69 | * -1 in the output |
---|
70 | * |
---|
71 | * \param data Float array of unnormalized data, will be normalized on return |
---|
72 | * \param count Number of elts in array |
---|
73 | * \param stride Stride between values in data array |
---|
74 | * \param vmin Minimum value in data array |
---|
75 | * \param vmax Maximum value in data array |
---|
76 | */ |
---|
77 | void |
---|
78 | nv::normalizeVector(float *data, int count, double vmin, double vmax) |
---|
79 | { |
---|
80 | for (int p = 0; p < count; p++) { |
---|
81 | int i = p * 4; |
---|
82 | data[i ] = data[i]/vmax; |
---|
83 | data[i+1] = data[i+1]/(2.0 * vmax) + 0.5; |
---|
84 | data[i+2] = data[i+2]/(2.0 * vmax) + 0.5; |
---|
85 | data[i+3] = data[i+3]/(2.0 * vmax) + 0.5; |
---|
86 | } |
---|
87 | } |
---|
88 | |
---|
89 | /** |
---|
90 | * \brief Compute Sobel filtered gradients for a 3D volume |
---|
91 | * |
---|
92 | * This technique is fairly expensive in terms of memory and |
---|
93 | * running time due to the filter extent. |
---|
94 | * |
---|
95 | * \param data Data array with X the fastest running, stride of |
---|
96 | * 1 float |
---|
97 | * \param nx number of values in X direction |
---|
98 | * \param ny number of values in Y direction |
---|
99 | * \param nz number of values in Z direction |
---|
100 | * \param dx Size of voxels in X direction |
---|
101 | * \param dy Size of voxels in Y direction |
---|
102 | * \param dz Size of voxels in Z direction |
---|
103 | * \param min Minimum value in data |
---|
104 | * \param max Maximum value in data |
---|
105 | * \return Returns a float array with stride of 4 floats |
---|
106 | * containing the normalized scalar and the x,y,z components of |
---|
107 | * the (normalized) gradient vector |
---|
108 | */ |
---|
109 | float * |
---|
110 | nv::computeGradient(float *data, |
---|
111 | int nx, int ny, int nz, |
---|
112 | float dx, float dy, float dz, |
---|
113 | float min, float max) |
---|
114 | { |
---|
115 | int npts = nx * ny * nz; |
---|
116 | float *gradients = (float *)malloc(npts * 3 * sizeof(float)); |
---|
117 | float *tempGradients = (float *)malloc(npts * 3 * sizeof(float)); |
---|
118 | int sizes[3] = { nx, ny, nz }; |
---|
119 | float spacing[3] = { dx, dy, dz }; |
---|
120 | computeGradients(tempGradients, data, sizes, spacing, DATRAW_FLOAT); |
---|
121 | filterGradients(tempGradients, sizes); |
---|
122 | quantizeGradients(tempGradients, gradients, sizes, DATRAW_FLOAT); |
---|
123 | free(tempGradients); |
---|
124 | normalizeScalar(data, npts, 1, min, max); |
---|
125 | float *dataOut = merge(data, gradients, npts); |
---|
126 | free(gradients); |
---|
127 | return dataOut; |
---|
128 | } |
---|
129 | |
---|
130 | /** |
---|
131 | * \brief Compute gradients for a 3D volume with cubic cells |
---|
132 | * |
---|
133 | * The gradients are estimated using the central difference |
---|
134 | * method. This function assumes the data are normalized |
---|
135 | * to [0,1] with missing data/NaNs represented by a negative |
---|
136 | * value. |
---|
137 | * |
---|
138 | * \param data Data array with X the fastest running. There |
---|
139 | * should be 4 floats allocated for each node, with the |
---|
140 | * first float containing the scalar value. The subsequent |
---|
141 | * 3 floats will be filled with the x,y,z components of the |
---|
142 | * gradient vector |
---|
143 | * \param nx The number of nodes in the X direction |
---|
144 | * \param ny The number of nodes in the Y direction |
---|
145 | * \param nz The number of nodes in the Z direction |
---|
146 | * \param dx The spacing (cell length) in the X direction |
---|
147 | * \param dy The spacing (cell length) in the Y direction |
---|
148 | * \param dz The spacing (cell length) in the Z direction |
---|
149 | */ |
---|
150 | void |
---|
151 | nv::computeSimpleGradient(float *data, |
---|
152 | int nx, int ny, int nz, |
---|
153 | float dx, float dy, float dz) |
---|
154 | { |
---|
155 | bool clampToEdge = true; |
---|
156 | double borderVal = 0.0; |
---|
157 | |
---|
158 | #define BORDER ((clampToEdge ? data[ngen] : borderVal)) |
---|
159 | |
---|
160 | // Compute the gradient of this data. BE CAREFUL: center |
---|
161 | // calculation on each node to avoid skew in either direction. |
---|
162 | int ngen = 0; |
---|
163 | for (int iz = 0; iz < nz; iz++) { |
---|
164 | for (int iy = 0; iy < ny; iy++) { |
---|
165 | for (int ix = 0; ix < nx; ix++) { |
---|
166 | // gradient in x-direction |
---|
167 | double valm1 = (ix == 0) ? BORDER : data[ngen - 4]; |
---|
168 | double valp1 = (ix == nx-1) ? BORDER : data[ngen + 4]; |
---|
169 | if (valm1 < 0.0 || valp1 < 0.0) { |
---|
170 | data[ngen+1] = 0.0; |
---|
171 | } else { |
---|
172 | data[ngen+1] = -(valp1-valm1)/(2. * dx); |
---|
173 | } |
---|
174 | |
---|
175 | // gradient in y-direction |
---|
176 | valm1 = (iy == 0) ? BORDER : data[ngen - 4*nx]; |
---|
177 | valp1 = (iy == ny-1) ? BORDER : data[ngen + 4*nx]; |
---|
178 | if (valm1 < 0.0 || valp1 < 0.0) { |
---|
179 | data[ngen+2] = 0.0; |
---|
180 | } else { |
---|
181 | data[ngen+2] = -(valp1-valm1)/(2. * dy); |
---|
182 | } |
---|
183 | |
---|
184 | // gradient in z-direction |
---|
185 | valm1 = (iz == 0) ? BORDER : data[ngen - 4*nx*ny]; |
---|
186 | valp1 = (iz == nz-1) ? BORDER : data[ngen + 4*nx*ny]; |
---|
187 | if (valm1 < 0.0 || valp1 < 0.0) { |
---|
188 | data[ngen+3] = 0.0; |
---|
189 | } else { |
---|
190 | data[ngen+3] = -(valp1-valm1)/(2. * dz); |
---|
191 | } |
---|
192 | // Normalize and scale/bias to [0,1] range |
---|
193 | // The volume shader will expand to [-1,1] |
---|
194 | double len = sqrt(data[ngen+1]*data[ngen+1] + |
---|
195 | data[ngen+2]*data[ngen+2] + |
---|
196 | data[ngen+3]*data[ngen+3]); |
---|
197 | if (len < 1.0e-6) { |
---|
198 | data[ngen+1] = 0.0; |
---|
199 | data[ngen+2] = 0.0; |
---|
200 | data[ngen+3] = 0.0; |
---|
201 | } else { |
---|
202 | data[ngen+1] = (data[ngen+1]/len + 1.0) * 0.5; |
---|
203 | data[ngen+2] = (data[ngen+2]/len + 1.0) * 0.5; |
---|
204 | data[ngen+3] = (data[ngen+3]/len + 1.0) * 0.5; |
---|
205 | } |
---|
206 | ngen += 4; |
---|
207 | } |
---|
208 | } |
---|
209 | } |
---|
210 | |
---|
211 | #undef BORDER |
---|
212 | } |
---|