/usr/include/gamera/pixel.hpp is in python-gamera-dev 3.3.3-2+deb7u1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 | /*
*
* Copyright (C) 2001-2005 Ichiro Fujinaga, Michael Droettboom, and Karl MacMillan
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version 2
* of the License, or (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
*/
#ifndef kwm11162001_pixel_hpp
#define kwm11162001_pixel_hpp
/**
* This header contains the definition for all of the standard pixels in
* Gamera. These include:
*
* RGB - color pixels
* Complex - complex number pixels are convenient for fourier image
* (frequency domain) processing algorithms. These values are
* similar to float values, but there are two values for each pixel.
* Float - floating point pixels that are convenient for many image processing
* algorithms
* GreyScale - grey scale pixels that hold values from 0 - 255 (8bit)
* OneBit - one bit pixels for black and white images. These pixels actually
* can hold more than 2 values, which is used for labeling the pixels
* (using connected-components for example). This seems like a lot
* of space to waste on one bit images, but if run-length encoding
* is used the space should be minimul.
*
* In addition to the pixels themselves, there is information about the pixels
* (white/black values, etc).
*/
#include "gamera_limits.hpp"
#include "vigra/rgbvalue.hxx"
#include "vigra/colorconversions.hxx"
#include <complex>
using namespace vigra;
namespace Gamera {
/************************************************************************
* PIXEL TYPES
************************************************************************/
/**
* Floating-point pixel.
*
* The Gamera::FloatPixel type represents a single pixel in a
* floating-point image. For floating-point images 0 is considerd
* black and max is considered white.
*/
typedef double FloatPixel;
/**
* GreyScalePixel
*
* The Gamera::GreyScalePixel type is for 8bit greyscale images. For GreyScale
* images 0 is considerd black and 255 is considered white.
*/
typedef unsigned char GreyScalePixel;
/**
* Grey16Pixel
*
* The Gamera::Grey16Pixel type is for 16bit greyscale images.
*/
typedef unsigned int Grey16Pixel;
/**
* OneBitPixel
*
* The Gamera::OneBitPixel type is for OneBitImages. For OneBit
* images > 0 is considerd black and 0 is considered white. Also, see the note
* at the beginning of this file about why OneBitPixels are so large.
*/
typedef unsigned short OneBitPixel;
/**
* ComplexPixel
*
* The Gamera::ComplexPixel type represents a pixel with two values:
* real and imaginary. These values are accessed by real() and imag()
* functions. Most functions follow normal std::complex behavior.
* Other behavior will generally mimic the floating-point pixel type often
* by the same operation applied only to the real part of the pixel.
*/
typedef std::complex<double> ComplexPixel;
/**
* RGB Pixels
*
* The Gamera::RGB pixel type is derived from the Vigra class RGBValue. The
* only reason that this is a derived class instead of directly using the
* Vigra type is to provide conversion operators to and from the standard
* Gamera types (instead of using Vigra style promotion traits) and to provide
* overloaded red, green, and blue functions instead of the set* functions
* in the Vigra class.
*/
template<class T>
class Rgb : public RGBValue<T> {
protected:
using RGBValue<T>::data_;
public:
using RGBValue<T>::luminance;
/**
* Construct a RGB pixel from a GreyScalePixel. RGB are all
* set to the passed in GreyScalePixel.
*/
explicit Rgb(GreyScalePixel grey) : RGBValue<T>(grey) { }
/**
* Construct a RGB pixel from a Grey16Pixel. RGB are all
* set to the passed in Grey16Pixel.
*/
explicit Rgb(Grey16Pixel grey) : RGBValue<T>(grey) { }
/**
* Construct a RGB pixel from a Float. RGB are all
* set to the passed in Float (which is truncated first).
*/
explicit Rgb(FloatPixel f) : RGBValue<T>((T)f) { }
/**
* Construct a RGB pixel from a Complex. RGB are all
* set to the real part passed in Complex (which is truncated
* first).
*/
explicit Rgb(ComplexPixel j) : RGBValue<T>((T)j.real()) { }
/**
* Construct a RGB Pixel from a OneBitPixel. Appropriate conversion
* is done.
*/
explicit Rgb(OneBitPixel s) {
// TODO: fix for new ONEBIT
if (s > 0) {
RGBValue<T>(1);
} else {
RGBValue<T>(0);
}
}
/**
* Default constructor - RGB are all set to 0.
*/
Rgb() : RGBValue<T>() { }
/**
* Copy constructor.
*/
template <class U>
Rgb(RGBValue<U> const & r) : RGBValue<T>(r) { }
Rgb(const Rgb& other) : RGBValue<T>(other) { }
/**
* Construct a RGB pixel from the passed in red, green, and blue
* values.
*/
Rgb(T red, T green, T blue) : RGBValue<T>(red, green, blue) { }
/**
* Construct a RGB pixel from the values contained in the iterator
* range passed in.
*/
template<class I>
Rgb(I i, const I end) : RGBValue<T>(i, end) { }
/* This is totally arbitrary, and doesn't make sense in terms
of "colors", but it will make using RGB as a key in a std::map
work.
*/
bool operator<(const Rgb<T>& other) const {
const typename vigra::NumericTraits<T>::Promote s =
(typename vigra::NumericTraits<T>::Promote)vigra::NumericTraits<T>::max;
const typename vigra::NumericTraits<T>::Promote s2 = s * s;
return (red() * s2 + green() * s + blue() <
other.red() * s2 + green() * s + blue());
}
/// Set the red component to the passed in value.
void red(T v) {
this->setRed(v);
}
/// Set the green component to the passed in value.
void green(T v) {
this->setGreen(v);
}
/// Set the blue component to the passed in value.
void blue(T v) {
this->setBlue(v);
}
/// Retrieve the red component - the returned value is an lvalue.
T const & red() const {
return data_[0];
}
/// Retrieve the green component - the returned value is an lvalue.
T const & green() const {
return data_[1];
}
/// Retrieve the blue component - the returned value is an lvalue.
T const & blue() const {
return data_[2];
}
/// Return the hue of this pixel.
FloatPixel const hue() {
FloatPixel maxc = (FloatPixel)std::max(data_[0], std::max(data_[1], data_[2]));
FloatPixel minc = (FloatPixel)std::min(data_[0], std::min(data_[1], data_[2]));
if (minc == maxc)
return 0;
FloatPixel den = (maxc - minc);
FloatPixel rc = (maxc - data_[0]) / den;
FloatPixel gc = (maxc - data_[1]) / den;
FloatPixel bc = (maxc - data_[2]) / den;
FloatPixel h;
if (data_[0] == maxc)
h = bc - gc;
else if (data_[1] == maxc)
h = 2.0 + rc - bc;
else
h = 4.0 + gc - rc;
h /= 6.0;
h -= floor(h);
return h;
}
/// Return the saturation of this pixel
FloatPixel const saturation() {
FloatPixel maxc = (FloatPixel)std::max(data_[0], std::max(data_[1], data_[2]));
FloatPixel minc = (FloatPixel)std::min(data_[0], std::min(data_[1], data_[2]));
if (minc == maxc)
return 0;
return (maxc - minc) / maxc;
}
/// Return the value of this pixel (max of RGB)
FloatPixel const value() {
return (FloatPixel)((float)(std::max(data_[0], std::max(data_[1], data_[2])))/255.0);
}
// conversion to CIE color space XYZ
FloatPixel const cie_x() {
RGB2XYZFunctor<FloatPixel> rgb2xyz_func;
RGB2XYZFunctor<FloatPixel>::result_type xyz;
xyz = rgb2xyz_func( RGBValue<FloatPixel>(data_[0], data_[1], data_[2]) );
return xyz[0];
}
FloatPixel const cie_y() {
RGB2XYZFunctor<FloatPixel> rgb2xyz_func;
RGB2XYZFunctor<FloatPixel>::result_type xyz;
xyz = rgb2xyz_func( RGBValue<FloatPixel>(data_[0], data_[1], data_[2]) );
return xyz[1];
}
FloatPixel const cie_z() {
RGB2XYZFunctor<FloatPixel> rgb2xyz_func;
RGB2XYZFunctor<FloatPixel>::result_type xyz;
xyz = rgb2xyz_func( RGBValue<FloatPixel>(data_[0], data_[1], data_[2]) );
return xyz[2];
}
// conversion to CIE color space Lab
FloatPixel const cie_Lab_L() {
RGB2LabFunctor<FloatPixel> rgb2lab_func;
RGB2LabFunctor<FloatPixel>::result_type lab;
lab = rgb2lab_func( RGBValue<FloatPixel>(data_[0], data_[1], data_[2]) );
return lab[0];
}
FloatPixel const cie_Lab_a() {
RGB2LabFunctor<FloatPixel> rgb2lab_func;
RGB2LabFunctor<FloatPixel>::result_type lab;
lab = rgb2lab_func( RGBValue<FloatPixel>(data_[0], data_[1], data_[2]) );
return lab[1];
}
FloatPixel const cie_Lab_b() {
RGB2LabFunctor<FloatPixel> rgb2lab_func;
RGB2LabFunctor<FloatPixel>::result_type lab;
lab = rgb2lab_func( RGBValue<FloatPixel>(data_[0], data_[1], data_[2]) );
return lab[2];
}
GreyScalePixel const cyan() {
return std::numeric_limits<T>::max() - data_[0];
}
GreyScalePixel const magenta() {
return std::numeric_limits<T>::max() - data_[1];
}
GreyScalePixel const yellow() {
return std::numeric_limits<T>::max() - data_[2];
}
// /// Conversion operator to a FloatPixel
// operator FloatPixel() {
// return FloatPixel(luminance());
// }
// /// Conversion operator to a ComplexPixel
// operator ComplexPixel() {
// ComplexPixel temp;
// temp.real = luminance();
// temp.imag = 0;
// return ComplexPixel(temp);
// }
// /// Conversion operator to a GreyScalePixel
// operator GreyScalePixel() {
// return GreyScalePixel(luminance());
// }
// /// Conversion operator to a Grey16Pixel
// operator Grey16Pixel() {
// return Grey16Pixel(luminance());
// }
// /// Conversion operator to a OneBitPixel
// operator OneBitPixel() {
// if (luminance())
// return 1;
// else
// return 0;
// }
};
/// This is the standard form of the RGB pixels
typedef Rgb<GreyScalePixel> RGBPixel;
/*
* This is a test for black/white regardless of the pixel type. For some
* pixel types this test is complicated and this also allows us to use 0
* for white in OneBit images and max for white in others without sacrificing
* generality in the algorithms.
*
* This default implementation is here mainly for CCProxies (see
* connected_components.hpp). Most of the real implementations are
* further down.
*/
template<class T>
inline bool is_black(T value) {
return value;
}
/*
* This is here for the same reason as is_black above.
*/
template<class T>
inline bool is_white(T value) {
return !value;
}
/*
* pixel_traits allows us to find out certain properties of pixels in a generic
* way. Again, this is primarily to allow the easy switching between min is white
* and min is black representations for different pixel types.
*/
template<class T>
struct pixel_traits {
static T white() {
return std::numeric_limits<T>::max();
}
static T black() {
return 0;
}
static T default_value() {
return white();
}
};
/*
* Helper functions to get black/white from a given T that has a value_type
* member that is a pixel - i.e.
*
* DenseImage<OneBitPixel> ob;
* black(ob);
*
* The pixel_traits syntax is just too horrible to make users go through to
* get white/black. From within a template function it looks like:
*
* Gamera::pixel_traits<typename T::value_type>::white();
*
*/
template<class T>
typename T::value_type black(T& container) {
return pixel_traits<typename T::value_type>::black();
}
template<class T>
typename T::value_type white(T& container) {
return pixel_traits<typename T::value_type>::white();
}
/*
Everything beyond this point is implementation
*/
// Specializations for black/white
template<>
inline bool is_black<FloatPixel>(FloatPixel value) {
return value <= 0;
}
template<>
inline bool is_black<GreyScalePixel>(GreyScalePixel value) {
return value == 0;
}
template<>
inline bool is_black<Grey16Pixel>(Grey16Pixel value) {
return value == 0;
}
template<>
inline bool is_black<RGBPixel>(RGBPixel value) {
return (value.green() == 0 && value.red() == 0 && value.blue() == 0);
}
template<>
inline bool is_black<OneBitPixel>(OneBitPixel value) {
return value != 0;
}
template<>
inline bool is_white<FloatPixel>(FloatPixel value) {
return (value == std::numeric_limits<GreyScalePixel>::max());
}
template<>
inline bool is_white<GreyScalePixel>(GreyScalePixel value) {
return (value == std::numeric_limits<GreyScalePixel>::max());
}
template<>
inline bool is_white<Grey16Pixel>(Grey16Pixel value) {
return (value == std::numeric_limits<Grey16Pixel>::max());
}
template<>
inline bool is_white<RGBPixel>(RGBPixel value) {
return (value.red() == std::numeric_limits<GreyScalePixel>::max()
&& value.green() == std::numeric_limits<GreyScalePixel>::max()
&& value.blue() == std::numeric_limits<GreyScalePixel>::max());
}
template<>
inline bool is_white<OneBitPixel>(OneBitPixel value) {
return value == 0;
}
/*
Specialization for pixel_traits
*/
template<>
inline OneBitPixel pixel_traits<OneBitPixel>::black() {
return 1;
}
template<>
inline OneBitPixel pixel_traits<OneBitPixel>::white() {
return 0;
}
template<>
inline Grey16Pixel pixel_traits<Grey16Pixel>::white() {
return 65535; // 2^16 - 1
}
template<>
inline RGBPixel pixel_traits<RGBPixel>::black() {
return RGBPixel(0, 0, 0);
}
template<>
inline RGBPixel pixel_traits<RGBPixel>::white() {
return RGBPixel(std::numeric_limits<GreyScalePixel>::max(),
std::numeric_limits<GreyScalePixel>::max(),
std::numeric_limits<GreyScalePixel>::max());
}
template<>
inline FloatPixel pixel_traits<FloatPixel>::default_value() {
return 0.0;
}
template<>
inline ComplexPixel pixel_traits<ComplexPixel>::white() {
return ComplexPixel(std::numeric_limits<double>::max(), 0.0);
}
template<>
inline ComplexPixel pixel_traits<ComplexPixel>::black() {
return ComplexPixel(0.0, 0.0);
}
template<>
inline ComplexPixel pixel_traits<ComplexPixel>::default_value() {
return pixel_traits<ComplexPixel>::black();
}
/*
* Inversion of pixel values
*
* Generically invert pixel values.
*/
inline FloatPixel invert(FloatPixel value) {
// Hard to know what makes sense here... MGD
return -value;
}
inline ComplexPixel invert(ComplexPixel value) {
return -value;
}
inline GreyScalePixel invert(GreyScalePixel value) {
return std::numeric_limits<GreyScalePixel>::max() - value;
}
inline Grey16Pixel invert(Grey16Pixel value) {
return std::numeric_limits<Grey16Pixel>::max() - value;
}
inline RGBPixel invert(RGBPixel value) {
return RGBPixel(std::numeric_limits<RGBPixel::value_type>::max() -
value.red(),
std::numeric_limits<RGBPixel::value_type>::max() -
value.green(),
std::numeric_limits<RGBPixel::value_type>::max() -
value.blue());
}
inline OneBitPixel invert(OneBitPixel value) {
if (is_white(value))
return pixel_traits<OneBitPixel>::black();
else
return pixel_traits<OneBitPixel>::white();
}
/*
* Blend pixels together.
*/
inline FloatPixel blend(FloatPixel original, FloatPixel add, double alpha) {
return alpha * original + (1.0 - alpha) * add;
}
inline ComplexPixel blend(ComplexPixel original, ComplexPixel add, double alpha) {
return alpha * original + (1.0 - alpha) * add;
}
inline GreyScalePixel blend(GreyScalePixel original, GreyScalePixel add, double alpha) {
return (GreyScalePixel)(alpha * double(original) + (1.0 - alpha) * double(add));
}
inline Grey16Pixel blend(Grey16Pixel original, GreyScalePixel add, double alpha) {
return (Grey16Pixel)(alpha * original + (1.0 - alpha) * add);
}
inline RGBPixel blend(RGBPixel original, RGBPixel add, double alpha) {
double inv_alpha = 1.0 - alpha;
return RGBPixel(GreyScalePixel(original.red() * alpha + add.red() * inv_alpha),
GreyScalePixel(original.green() * alpha + add.green() * inv_alpha),
GreyScalePixel(original.blue() * alpha + add.blue() * inv_alpha));
}
inline OneBitPixel blend(OneBitPixel original, RGBPixel add, double alpha) {
if (alpha > 0.5)
return original;
return add.luminance();
}
};
#endif
|