1 /*=========================================================================
4 Module: $RCSfile: gdcmPixelReadConvert.cxx,v $
6 Date: $Date: 2007/05/23 14:18:11 $
7 Version: $Revision: 1.115 $
9 Copyright (c) CREATIS (Centre de Recherche et d'Applications en Traitement de
10 l'Image). All rights reserved. See Doc/License.txt or
11 http://www.creatis.insa-lyon.fr/Public/Gdcm/License.html for details.
13 This software is distributed WITHOUT ANY WARRANTY; without even
14 the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR
15 PURPOSE. See the above copyright notices for more information.
17 =========================================================================*/
19 #include "gdcmPixelReadConvert.h"
20 #include "gdcmDebug.h"
22 #include "gdcmGlobal.h"
24 #include "gdcmDocEntry.h"
25 #include "gdcmRLEFramesInfo.h"
26 #include "gdcmJPEGFragmentsInfo.h"
29 #include <stdio.h> //for sscanf
31 #if defined(__BORLANDC__)
32 #include <mem.h> // for memset
35 namespace GDCM_NAME_SPACE
38 //bool ReadMPEGFile (std::ifstream *fp, char *inputdata, size_t lenght);
39 bool gdcm_read_JPEG2000_file (void* raw,
40 char *inputdata, size_t inputlength);
41 //-----------------------------------------------------------------------------
42 #define str2num(str, typeNum) *((typeNum *)(str))
44 //-----------------------------------------------------------------------------
45 // Constructor / Destructor
47 PixelReadConvert::PixelReadConvert()
63 /// Canonical Destructor
64 PixelReadConvert::~PixelReadConvert()
69 //-----------------------------------------------------------------------------
72 * \brief Predicate to know whether the image[s] (once Raw) is RGB.
73 * \note See comments of \ref ConvertHandleColor
75 bool PixelReadConvert::IsRawRGB()
78 || PlanarConfiguration == 2
86 * \brief Gets various usefull informations from the file header
87 * @param file gdcm::File pointer
88 * @param fileHelper gdcm::FileHelper pointer
90 void PixelReadConvert::GrabInformationsFromFile( File *file,
91 FileHelper *fileHelper )
93 // Number of Bits Allocated for storing a Pixel is defaulted to 16
94 // when absent from the file.
95 BitsAllocated = file->GetBitsAllocated();
96 if ( BitsAllocated == 0 )
101 else if (BitsAllocated > 8 && BitsAllocated < 16 && BitsAllocated != 12)
105 // Number of "Bits Stored", defaulted to number of "Bits Allocated"
106 // when absent from the file.
107 BitsStored = file->GetBitsStored();
108 if ( BitsStored == 0 )
110 BitsStored = BitsAllocated;
113 // High Bit Position, defaulted to "Bits Allocated" - 1
114 HighBitPosition = file->GetHighBitPosition();
115 if ( HighBitPosition == 0 )
117 HighBitPosition = BitsAllocated - 1;
120 XSize = file->GetXSize();
121 YSize = file->GetYSize();
122 ZSize = file->GetZSize();
123 TSize = file->GetTSize();
124 SamplesPerPixel = file->GetSamplesPerPixel();
125 //PixelSize = file->GetPixelSize(); Useless
126 PixelSign = file->IsSignedPixelData();
127 SwapCode = file->GetSwapCode();
129 IsPrivateGETransferSyntax = IsMPEG
130 = IsJPEG2000 = IsJPEGLS = IsJPEGLossy
131 = IsJPEGLossless = IsRLELossless
134 if (! file->IsDicomV3() ) // Should be ACR-NEMA file
140 std::string ts = file->GetTransferSyntax();
143 while (true) // shorter to write than 'if elseif elseif elseif' ...
145 // mind the order : check the most usual first.
146 if( IsRaw = Global::GetTS()->GetSpecialTransferSyntax(ts) == TS::ExplicitVRLittleEndian) break;
147 if( IsRaw = Global::GetTS()->GetSpecialTransferSyntax(ts) == TS::ImplicitVRLittleEndian ) break;
148 if( IsRaw = Global::GetTS()->GetSpecialTransferSyntax(ts) == TS::ExplicitVRBigEndian) break;
149 if( IsRaw = Global::GetTS()->GetSpecialTransferSyntax(ts) == TS::ImplicitVRBigEndianPrivateGE) break;
150 // DeflatedExplicitVRLittleEndian syntax means the whole Dataset (Header + Pixels) is compressed !
151 // Not dealt with ! (Parser hangs)
152 //if( IsRaw = Global::GetTS()->GetSpecialTransferSyntax(ts) == TS::DeflatedExplicitVRLittleEndian) break;
155 // cache whether this is a strange GE transfer syntax (which uses
156 // a little endian transfer syntax for the header and a big endian
157 // transfer syntax for the pixel data).
158 IsPrivateGETransferSyntax =
159 ( Global::GetTS()->GetSpecialTransferSyntax(ts) == TS::ImplicitVRBigEndianPrivateGE );
161 IsMPEG = IsJPEG2000 = IsJPEGLS = IsJPEGLossy = IsJPEGLossless = IsRLELossless = false;
166 // mind the order : check the most usual first.
167 if( IsJPEGLossy = Global::GetTS()->IsJPEGLossy(ts) ) break;
168 if( IsJPEGLossless = Global::GetTS()->IsJPEGLossless(ts) ) break;
169 if( IsRLELossless = Global::GetTS()->IsRLELossless(ts) ) break;
170 if( IsJPEG2000 = Global::GetTS()->IsJPEG2000(ts) ) break;
171 if( IsMPEG = Global::GetTS()->IsMPEG(ts) ) break;
172 if( IsJPEGLS = Global::GetTS()->IsJPEGLS(ts) ) break;
173 // DeflatedExplicitVRLittleEndian is considered as 'Unexpected' (we don't know yet how to process !)
174 gdcmWarningMacro("Unexpected Transfer Syntax :[" << ts << "]");
180 PixelOffset = file->GetPixelOffset();
181 PixelDataLength = file->GetPixelAreaLength();
182 RLEInfo = file->GetRLEInfo();
183 JPEGInfo = file->GetJPEGInfo();
185 IsMonochrome = file->IsMonochrome();
186 IsMonochrome1 = file->IsMonochrome1();
187 IsPaletteColor = file->IsPaletteColor();
188 IsYBRFull = file->IsYBRFull();
190 PlanarConfiguration = file->GetPlanarConfiguration();
192 /////////////////////////////////////////////////////////////////
194 HasLUT = file->HasLUT();
197 // Just in case some access to a File element requires disk access.
198 LutRedDescriptor = file->GetEntryString( 0x0028, 0x1101 );
199 LutGreenDescriptor = file->GetEntryString( 0x0028, 0x1102 );
200 LutBlueDescriptor = file->GetEntryString( 0x0028, 0x1103 );
202 // FIXME : The following comment is probabely meaningless, since LUT are *always*
203 // loaded at parsing time, whatever their length is.
205 // Depending on the value of Document::MAX_SIZE_LOAD_ELEMENT_VALUE
206 // [ refer to invocation of Document::SetMaxSizeLoadEntry() in
207 // Document::Document() ], the loading of the value (content) of a
208 // [Bin|Val]Entry occurence migth have been hindered (read simply NOT
209 // loaded). Hence, we first try to obtain the LUTs data from the file
210 // and when this fails we read the LUTs data directly from disk.
211 // \TODO Reading a [Bin|Val]Entry directly from disk is a kludge.
212 // We should NOT bypass the [Bin|Val]Entry class. Instead
213 // an access to an UNLOADED content of a [Bin|Val]Entry occurence
214 // (e.g. DataEntry::GetBinArea()) should force disk access from
215 // within the [Bin|Val]Entry class itself. The only problem
216 // is that the [Bin|Val]Entry is unaware of the FILE* is was
217 // parsed from. Fix that. FIXME.
220 file->LoadEntryBinArea(0x0028, 0x1201);
221 LutRedData = (uint8_t*)file->GetEntryBinArea( 0x0028, 0x1201 );
224 gdcmWarningMacro("Unable to read Red Palette Color Lookup Table data");
228 file->LoadEntryBinArea(0x0028, 0x1202);
229 LutGreenData = (uint8_t*)file->GetEntryBinArea(0x0028, 0x1202 );
232 gdcmWarningMacro("Unable to read Green Palette Color Lookup Table data");
236 file->LoadEntryBinArea(0x0028, 0x1203);
237 LutBlueData = (uint8_t*)file->GetEntryBinArea( 0x0028, 0x1203 );
240 gdcmWarningMacro("Unable to read Blue Palette Color Lookup Table data");
245 ComputeRawAndRGBSizes();
248 /// \brief Reads from disk and decompresses Pixels
249 bool PixelReadConvert::ReadAndDecompressPixelData( std::ifstream *fp )
251 // ComputeRawAndRGBSizes is already made by
252 // ::GrabInformationsFromfile. So, the structure sizes are
256 //////////////////////////////////////////////////
257 //// First stage: get our hands on the Pixel Data.
260 gdcmWarningMacro( "Unavailable file pointer." );
264 fp->seekg( PixelOffset, std::ios::beg );
265 if ( fp->fail() || fp->eof() )
267 gdcmWarningMacro( "Unable to find PixelOffset in file." );
273 //////////////////////////////////////////////////
275 CallStartMethod(); // for progress bar
276 unsigned int count = 0;
277 unsigned int frameSize;
278 unsigned int bitsAllocated = BitsAllocated;
279 //if(bitsAllocated == 12)
280 if(bitsAllocated > 8 && bitsAllocated < 16)
282 frameSize = XSize*YSize*SamplesPerPixel*bitsAllocated/8;
284 //// Second stage: read from disk and decompress.
286 if ( BitsAllocated == 12 ) // We suppose 'BitsAllocated' = 12 only exist for uncompressed files
288 ReadAndDecompress12BitsTo16Bits( fp);
292 // This problem can be found when some obvious informations are found
293 // after the field containing the image data. In this case, these
294 // bad data are added to the size of the image (in the PixelDataLength
295 // variable). But RawSize is the right size of the image !
296 if ( PixelDataLength != RawSize )
298 gdcmWarningMacro( "Mismatch between PixelReadConvert : "
299 << PixelDataLength << " and RawSize : " << RawSize );
302 //todo : is it the right patch?
303 char *raw = (char*)Raw;
304 uint32_t remainingLength;
306 unsigned int lengthToRead;
308 if ( PixelDataLength > RawSize )
309 lengthToRead = RawSize;
311 lengthToRead = PixelDataLength;
313 // perform a frame by frame reading
314 remainingLength = lengthToRead;
315 unsigned int nbFrames = lengthToRead / frameSize;
316 for (i=0;i<nbFrames; i++)
318 Progress = (float)(count+1)/(float)nbFrames;
319 fp->read( raw, frameSize);
321 remainingLength -= frameSize;
324 if (remainingLength !=0 )
325 fp->read( raw, remainingLength);
327 if ( fp->fail() || fp->eof())
329 gdcmWarningMacro( "Reading of Raw pixel data failed." );
333 else if ( IsRLELossless )
335 if ( ! RLEInfo->DecompressRLEFile
336 ( fp, Raw, XSize, YSize, ZSize, TSize, BitsAllocated ) )
338 gdcmWarningMacro( "RLE decompressor failed." );
344 //gdcmWarningMacro( "Sorry, MPEG not yet taken into account" );
346 // fp has already been seek to start of mpeg
347 //ReadMPEGFile(fp, (char*)Raw, PixelDataLength);
352 // Default case concerns JPEG family
353 if ( ! ReadAndDecompressJPEGFile( fp ) )
355 gdcmWarningMacro( "JPEG decompressor ( ReadAndDecompressJPEGFile()"
356 << " method ) failed." );
361 ////////////////////////////////////////////
362 //// Third stage: twigle the bytes and bits.
363 ConvertReorderEndianity();
364 ConvertReArrangeBits();
365 ConvertFixGreyLevels();
366 if (UserFunction) // user is allowed to Mirror, TopDown, Rotate,...the image
367 UserFunction( Raw, FileInternal);
368 ConvertHandleColor();
373 /// Deletes Pixels Area
374 void PixelReadConvert::Squeeze()
390 * \brief Build the RGB image from the Raw image and the LUTs.
392 bool PixelReadConvert::BuildRGBImage()
396 // The job is already done.
402 // The job can't be done
409 // The job can't be done
413 gdcmDebugMacro( "--> BuildRGBImage" );
419 if ( BitsAllocated <= 8 )
421 uint8_t *localRGB = RGB;
422 for (size_t i = 0; i < RawSize; ++i )
425 *localRGB++ = LutRGBA[j];
426 *localRGB++ = LutRGBA[j+1];
427 *localRGB++ = LutRGBA[j+2];
431 else // deal with 16 bits pixels and 16 bits Palette color
433 uint16_t *localRGB = (uint16_t *)RGB;
434 for (size_t i = 0; i < RawSize/2; ++i )
436 j = ((uint16_t *)Raw)[i] * 4;
437 *localRGB++ = ((uint16_t *)LutRGBA)[j];
438 *localRGB++ = ((uint16_t *)LutRGBA)[j+1];
439 *localRGB++ = ((uint16_t *)LutRGBA)[j+2];
446 //-----------------------------------------------------------------------------
449 //-----------------------------------------------------------------------------
452 * \brief Read from file a 12 bits per pixel image and decompress it
453 * into a 16 bits per pixel image.
455 void PixelReadConvert::ReadAndDecompress12BitsTo16Bits( std::ifstream *fp )
456 throw ( FormatError )
458 /// \todo Fix the 3D, 4D pb
459 int nbPixels = XSize * YSize * TSize;
460 uint16_t *localDecompres = (uint16_t*)Raw;
462 for( int p = 0; p < nbPixels; p += 2 )
466 fp->read( (char*)&b0, 1);
467 if ( fp->fail() || fp->eof() )
469 throw FormatError( "PixelReadConvert::ReadAndDecompress12BitsTo16Bits()",
470 "Unfound first block" );
473 fp->read( (char*)&b1, 1 );
474 if ( fp->fail() || fp->eof())
476 throw FormatError( "PixelReadConvert::ReadAndDecompress12BitsTo16Bits()",
477 "Unfound second block" );
480 fp->read( (char*)&b2, 1 );
481 if ( fp->fail() || fp->eof())
483 throw FormatError( "PixelReadConvert::ReadAndDecompress12BitsTo16Bits()",
484 "Unfound second block" );
487 // Two steps are necessary to please VC++
489 // 2 pixels 12bit = [0xABCDEF]
490 // 2 pixels 16bit = [0x0ABD] + [0x0FCE]
492 *localDecompres++ = ((b0 >> 4) << 8) + ((b0 & 0x0f) << 4) + (b1 & 0x0f);
494 *localDecompres++ = ((b2 & 0x0f) << 8) + ((b1 >> 4) << 4) + (b2 >> 4);
496 /// \todo JPR Troubles expected on Big-Endian processors ?
501 * \brief Reads from disk the Pixel Data of JPEG Dicom encapsulated
502 * file and decompress it.
503 * @param fp File Pointer
506 bool PixelReadConvert::ReadAndDecompressJPEGFile( std::ifstream *fp )
510 // make sure this is the right JPEG compression
511 assert( !IsJPEGLossless || !IsJPEGLossy || !IsJPEGLS );
512 // FIXME this is really ugly but it seems I have to load the complete
513 // jpeg2000 stream to use jasper:
514 // I don't think we'll ever be able to deal with multiple fragments properly
516 unsigned long inputlength = 0;
517 JPEGFragment *jpegfrag = JPEGInfo->GetFirstFragment();
520 inputlength += jpegfrag->GetLength();
521 jpegfrag = JPEGInfo->GetNextFragment();
523 gdcmAssertMacro( inputlength != 0);
524 uint8_t *inputdata = new uint8_t[inputlength];
525 char *pinputdata = (char*)inputdata;
526 jpegfrag = JPEGInfo->GetFirstFragment();
529 fp->seekg( jpegfrag->GetOffset(), std::ios::beg);
530 fp->read(pinputdata, jpegfrag->GetLength());
531 pinputdata += jpegfrag->GetLength();
532 jpegfrag = JPEGInfo->GetNextFragment();
534 // Warning the inputdata buffer is delete in the function
535 if ( ! gdcm_read_JPEG2000_file( Raw,
536 (char*)inputdata, inputlength ) )
540 // wow what happen, must be an error
541 gdcmWarningMacro( "gdcm_read_JPEG2000_file() failed ");
546 // make sure this is the right JPEG compression
547 assert( !IsJPEGLossless || !IsJPEGLossy || !IsJPEG2000 );
548 // WARNING : JPEG-LS is NOT the 'classical' Jpeg Lossless :
549 // [JPEG-LS is the basis for new lossless/near-lossless compression
550 // standard for continuous-tone images intended for JPEG2000. The standard
551 // is based on the LOCO-I algorithm (LOw COmplexity LOssless COmpression
552 // for Images) developed at Hewlett-Packard Laboratories]
554 // see http://datacompression.info/JPEGLS.shtml
557 std::cerr << "count:" << JPEGInfo->GetFragmentCount() << std::endl;
558 unsigned long inputlength = 0;
559 JPEGFragment *jpegfrag = JPEGInfo->GetFirstFragment();
562 inputlength += jpegfrag->GetLength();
563 jpegfrag = JPEGInfo->GetNextFragment();
565 gdcmAssertMacro( inputlength != 0);
566 uint8_t *inputdata = new uint8_t[inputlength];
567 char *pinputdata = (char*)inputdata;
568 jpegfrag = JPEGInfo->GetFirstFragment();
571 fp->seekg( jpegfrag->GetOffset(), std::ios::beg);
572 fp->read(pinputdata, jpegfrag->GetLength());
573 pinputdata += jpegfrag->GetLength();
574 jpegfrag = JPEGInfo->GetNextFragment();
577 //fp->read((char*)Raw, PixelDataLength);
579 std::ofstream out("/tmp/jpegls.jpg");
580 out.write((char*)inputdata, inputlength);
585 gdcmWarningMacro( "Sorry, JPEG-LS not yet taken into account" );
586 fp->seekg( JPEGInfo->GetFirstFragment()->GetOffset(), std::ios::beg);
587 // if ( ! gdcm_read_JPEGLS_file( fp,Raw ) )
592 // make sure this is the right JPEG compression
593 assert( !IsJPEGLS || !IsJPEG2000 );
594 // Precompute the offset localRaw will be shifted with
595 int length = XSize * YSize * ZSize * SamplesPerPixel;
596 int numberBytes = BitsAllocated / 8;
598 JPEGInfo->DecompressFromFile(fp, Raw, BitsStored, numberBytes, length );
604 * \brief Build Red/Green/Blue/Alpha LUT from File when :
605 * - (0028,0004) : Photometric Interpretation == [PALETTE COLOR ]
607 * - (0028,1101),(0028,1102),(0028,1102)
608 * xxx Palette Color Lookup Table Descriptor are found
610 * - (0028,1201),(0028,1202),(0028,1202)
611 * xxx Palette Color Lookup Table Data - are found
612 * \warning does NOT deal with :
613 * - 0028 1100 Gray Lookup Table Descriptor (Retired)
614 * - 0028 1221 Segmented Red Palette Color Lookup Table Data
615 * - 0028 1222 Segmented Green Palette Color Lookup Table Data
616 * - 0028 1223 Segmented Blue Palette Color Lookup Table Data
617 * no known Dicom reader deals with them :-(
618 * @return a RGBA Lookup Table
620 void PixelReadConvert::BuildLUTRGBA()
623 // Note to code reviewers :
624 // The problem is *much more* complicated, since a lot of manufacturers
625 // Don't follow the norm :
626 // have a look at David Clunie's remark at the end of this .cxx file.
633 // http://www.barre.nom.fr/medical/dicom2/limitations.html#Color%20Lookup%20Tables
635 if ( ! IsPaletteColor )
640 if ( LutRedDescriptor == GDCM_UNFOUND
641 || LutGreenDescriptor == GDCM_UNFOUND
642 || LutBlueDescriptor == GDCM_UNFOUND )
644 gdcmWarningMacro( "(At least) a LUT Descriptor is missing" );
648 ////////////////////////////////////////////
649 // Extract the info from the LUT descriptors
650 int lengthR; // Red LUT length in Bytes
651 int debR; // Subscript of the first Lut Value
652 int nbitsR; // Lut item size (in Bits)
653 int nbRead; // nb of items in LUT descriptor (must be = 3)
655 nbRead = sscanf( LutRedDescriptor.c_str(),
657 &lengthR, &debR, &nbitsR );
660 gdcmWarningMacro( "Wrong Red LUT descriptor" );
662 int lengthG; // Green LUT length in Bytes
663 int debG; // Subscript of the first Lut Value
664 int nbitsG; // Lut item size (in Bits)
666 nbRead = sscanf( LutGreenDescriptor.c_str(),
668 &lengthG, &debG, &nbitsG );
671 gdcmWarningMacro( "Wrong Green LUT descriptor" );
674 int lengthB; // Blue LUT length in Bytes
675 int debB; // Subscript of the first Lut Value
676 int nbitsB; // Lut item size (in Bits)
677 nbRead = sscanf( LutRedDescriptor.c_str(),
679 &lengthB, &debB, &nbitsB );
682 gdcmWarningMacro( "Wrong Blue LUT descriptor" );
685 gdcmDebugMacro(" lengthR " << lengthR << " debR "
686 << debR << " nbitsR " << nbitsR);
687 gdcmDebugMacro(" lengthG " << lengthG << " debG "
688 << debG << " nbitsG " << nbitsG);
689 gdcmDebugMacro(" lengthB " << lengthB << " debB "
690 << debB << " nbitsB " << nbitsB);
692 if ( !lengthR ) // if = 2^16, this shall be 0 see : CP-143
694 if ( !lengthG ) // if = 2^16, this shall be 0
696 if ( !lengthB ) // if = 2^16, this shall be 0
699 ////////////////////////////////////////////////////////
701 if ( ( ! LutRedData ) || ( ! LutGreenData ) || ( ! LutBlueData ) )
703 gdcmWarningMacro( "(At least) a LUT is missing" );
707 // -------------------------------------------------------------
709 if ( BitsAllocated <= 8 )
711 // forge the 4 * 8 Bits Red/Green/Blue/Alpha LUT
712 LutRGBA = new uint8_t[ 1024 ]; // 256 * 4 (R, G, B, Alpha)
717 memset( LutRGBA, 0, 1024 );
720 if ( ( nbitsR == 16 ) && ( BitsAllocated == 8 ) )
722 // when LUT item size is different than pixel size
723 mult = 2; // high byte must be = low byte
727 // See PS 3.3-2003 C.11.1.1.2 p 619
731 // if we get a black image, let's just remove the '+1'
732 // from 'i*mult+1' and check again
733 // if it works, we shall have to check the 3 Palettes
734 // to see which byte is ==0 (first one, or second one)
736 // We give up the checking to avoid some (useless ?) overhead
737 // (optimistic asumption)
741 //take "Subscript of the first Lut Value" (debR,debG,debB) into account!
743 //FIXME : +1 : to get 'low value' byte
744 // Trouble expected on Big Endian Processors ?
745 // 16 BIts Per Pixel Palette Color to be swapped?
747 a = LutRGBA + 0 + debR;
748 for( i=0; i < lengthR; ++i )
750 *a = LutRedData[i*mult+1];
754 a = LutRGBA + 1 + debG;
755 for( i=0; i < lengthG; ++i)
757 *a = LutGreenData[i*mult+1];
761 a = LutRGBA + 2 + debB;
762 for(i=0; i < lengthB; ++i)
764 *a = LutBlueData[i*mult+1];
769 for(i=0; i < 256; ++i)
771 *a = 1; // Alpha component
777 // Probabely the same stuff is to be done for 16 Bits Pixels
778 // with 65536 entries LUT ?!?
779 // Still looking for accurate info on the web :-(
781 gdcmWarningMacro( "Sorry Palette Color Lookup Tables not yet dealt with"
782 << " for 16 Bits Per Pixel images" );
784 // forge the 4 * 16 Bits Red/Green/Blue/Alpha LUT
786 LutRGBA = (uint8_t *)new uint16_t[ 65536*4 ]; // 2^16 * 4 (R, G, B, Alpha)
789 memset( LutRGBA, 0, 65536*4*2 ); // 16 bits = 2 bytes ;-)
791 LutItemNumber = 65536;
797 //take "Subscript of the first Lut Value" (debR,debG,debB) into account!
799 a16 = (uint16_t*)LutRGBA + 0 + debR;
800 for( i=0; i < lengthR; ++i )
802 *a16 = ((uint16_t*)LutRedData)[i];
806 a16 = (uint16_t*)LutRGBA + 1 + debG;
807 for( i=0; i < lengthG; ++i)
809 *a16 = ((uint16_t*)LutGreenData)[i];
813 a16 = (uint16_t*)LutRGBA + 2 + debB;
814 for(i=0; i < lengthB; ++i)
816 *a16 = ((uint16_t*)LutBlueData)[i];
820 a16 = (uint16_t*)LutRGBA + 3 ;
821 for(i=0; i < 65536; ++i)
823 *a16 = 1; // Alpha component
826 /* Just to 'see' the LUT, at debug time
827 // Don't remove this commented out code.
829 a16=(uint16_t*)LutRGBA;
830 for (int j=0;j<65536;j++)
832 std::cout << *a16 << " " << *(a16+1) << " "
833 << *(a16+2) << " " << *(a16+3) << std::endl;
841 * \brief Swap the bytes, according to \ref SwapCode.
843 void PixelReadConvert::ConvertSwapZone()
847 // If this file is 'ImplicitVR BigEndian PrivateGE Transfer Syntax',
848 // then the header is in little endian format and the pixel data is in
849 // big endian format. When reading the header, GDCM has already established
850 // a byte swapping code suitable for this machine to read the
851 // header. In TS::ImplicitVRBigEndianPrivateGE, this code will need
852 // to be switched in order to read the pixel data. This must be
853 // done REGARDLESS of the processor endianess!
855 // Example: Assume we are on a little endian machine. When
856 // GDCM reads the header, the header will match the machine
857 // endianess and the swap code will be established as a no-op.
858 // When GDCM reaches the pixel data, it will need to switch the
859 // swap code to do big endian to little endian conversion.
861 // Now, assume we are on a big endian machine. When GDCM reads the
862 // header, the header will be recognized as a different endianess
863 // than the machine endianess, and a swap code will be established
864 // to convert from little endian to big endian. When GDCM readers
865 // the pixel data, the pixel data endianess will now match the
866 // machine endianess. But we currently have a swap code that
867 // converts from little endian to big endian. In this case, we
868 // need to switch the swap code to a no-op.
870 // Therefore, in either case, if the file is in
871 // 'ImplicitVR BigEndian PrivateGE Transfer Syntax', then GDCM needs to switch
872 // the byte swapping code when entering the pixel data.
874 int tempSwapCode = SwapCode;
875 if ( IsPrivateGETransferSyntax )
877 gdcmWarningMacro(" IsPrivateGETransferSyntax found; turn the SwapCode");
878 // PrivateGETransferSyntax only exists for 'true' Dicom images
879 // we assume there is no 'exotic' 32 bits endianess!
880 if (SwapCode == 1234)
884 else if (SwapCode == 4321)
890 if ( BitsAllocated == 16 )
892 uint16_t *im16 = (uint16_t*)Raw;
893 switch( tempSwapCode )
900 for( i = 0; i < RawSize / 2; i++ )
902 im16[i]= (im16[i] >> 8) | (im16[i] << 8 );
906 gdcmWarningMacro("SwapCode value (16 bits) not allowed."
910 else if ( BitsAllocated == 32 )
915 uint32_t *im32 = (uint32_t*)Raw;
916 switch ( tempSwapCode )
921 for( i = 0; i < RawSize / 4; i++ )
923 low = im32[i] & 0x0000ffff; // 4321
924 high = im32[i] >> 16;
925 high = ( high >> 8 ) | ( high << 8 );
926 low = ( low >> 8 ) | ( low << 8 );
928 im32[i] = ( s32 << 16 ) | high;
932 for( i = 0; i < RawSize / 4; i++ )
934 low = im32[i] & 0x0000ffff; // 2143
935 high = im32[i] >> 16;
936 high = ( high >> 8 ) | ( high << 8 );
937 low = ( low >> 8 ) | ( low << 8 );
939 im32[i] = ( s32 << 16 ) | low;
943 for( i = 0; i < RawSize / 4; i++ )
945 low = im32[i] & 0x0000ffff; // 3412
946 high = im32[i] >> 16;
948 im32[i] = ( s32 << 16 ) | high;
952 gdcmWarningMacro("SwapCode value (32 bits) not allowed." << tempSwapCode );
958 * \brief Deal with endianness i.e. re-arange bytes inside the integer
960 void PixelReadConvert::ConvertReorderEndianity()
962 if ( BitsAllocated != 8 )
967 // Special kludge in order to deal with xmedcon broken images:
968 if ( BitsAllocated == 16
969 && BitsStored < BitsAllocated
972 int l = (int)( RawSize / ( BitsAllocated / 8 ) );
973 uint16_t *deb = (uint16_t *)Raw;
974 for(int i = 0; i<l; i++)
976 if ( *deb == 0xffff )
986 * \brief Deal with Grey levels i.e. re-arange them
987 * to have low values = dark, high values = bright
989 void PixelReadConvert::ConvertFixGreyLevels()
994 uint32_t i; // to please M$VC6
999 if ( BitsAllocated == 8 )
1001 uint8_t *deb = (uint8_t *)Raw;
1002 for (i=0; i<RawSize; i++)
1010 if ( BitsAllocated == 16 )
1013 for (j=0; j<BitsStored-1; j++)
1015 mask = (mask << 1) +1; // will be fff when BitsStored=12
1018 uint16_t *deb = (uint16_t *)Raw;
1019 for (i=0; i<RawSize/2; i++)
1029 if ( BitsAllocated == 8 )
1031 uint8_t smask8 = 255;
1032 uint8_t *deb = (uint8_t *)Raw;
1033 for (i=0; i<RawSize; i++)
1035 *deb = smask8 - *deb;
1040 if ( BitsAllocated == 16 )
1042 uint16_t smask16 = 65535;
1043 uint16_t *deb = (uint16_t *)Raw;
1044 for (i=0; i<RawSize/2; i++)
1046 *deb = smask16 - *deb;
1055 * \brief Re-arrange the bits within the bytes.
1056 * @return Boolean always true
1058 bool PixelReadConvert::ConvertReArrangeBits() throw ( FormatError )
1061 if ( BitsStored != BitsAllocated )
1063 int l = (int)( RawSize / ( BitsAllocated / 8 ) );
1064 if ( BitsAllocated == 16 )
1066 // pmask : to mask the 'unused bits' (may contain overlays)
1067 uint16_t pmask = 0xffff;
1068 pmask = pmask >> ( BitsAllocated - BitsStored );
1070 uint16_t *deb = (uint16_t*)Raw;
1072 if ( !PixelSign ) // Pixels are unsigned
1074 for(int i = 0; i<l; i++)
1076 *deb = (*deb >> (BitsStored - HighBitPosition - 1)) & pmask;
1080 else // Pixels are signed
1082 // smask : to check the 'sign' when BitsStored != BitsAllocated
1083 uint16_t smask = 0x0001;
1084 smask = smask << ( 16 - (BitsAllocated - BitsStored + 1) );
1085 // nmask : to propagate sign bit on negative values
1086 int16_t nmask = (int16_t)0x8000;
1087 nmask = nmask >> ( BitsAllocated - BitsStored - 1 );
1089 for(int i = 0; i<l; i++)
1091 *deb = *deb >> (BitsStored - HighBitPosition - 1);
1094 *deb = *deb | nmask;
1098 *deb = *deb & pmask;
1104 else if ( BitsAllocated == 32 )
1106 // pmask : to mask the 'unused bits' (may contain overlays)
1107 uint32_t pmask = 0xffffffff;
1108 pmask = pmask >> ( BitsAllocated - BitsStored );
1110 uint32_t *deb = (uint32_t*)Raw;
1114 for(int i = 0; i<l; i++)
1116 *deb = (*deb >> (BitsStored - HighBitPosition - 1)) & pmask;
1122 // smask : to check the 'sign' when BitsStored != BitsAllocated
1123 uint32_t smask = 0x00000001;
1124 smask = smask >> ( 32 - (BitsAllocated - BitsStored +1 ));
1125 // nmask : to propagate sign bit on negative values
1126 int32_t nmask = 0x80000000;
1127 nmask = nmask >> ( BitsAllocated - BitsStored -1 );
1129 for(int i = 0; i<l; i++)
1131 *deb = *deb >> (BitsStored - HighBitPosition - 1);
1133 *deb = *deb | nmask;
1135 *deb = *deb & pmask;
1142 gdcmWarningMacro("Weird image (BitsAllocated !=8, 12, 16, 32)");
1143 throw FormatError( "Weird image !?" );
1150 * \brief Convert (Red plane, Green plane, Blue plane) to RGB pixels
1151 * \warning Works on all the frames at a time
1153 void PixelReadConvert::ConvertRGBPlanesToRGBPixels()
1155 gdcmWarningMacro("--> ConvertRGBPlanesToRGBPixels");
1157 uint8_t *localRaw = Raw;
1158 uint8_t *copyRaw = new uint8_t[ RawSize ];
1159 memmove( copyRaw, localRaw, RawSize );
1161 int l = XSize * YSize * ZSize;
1163 uint8_t *a = copyRaw;
1164 uint8_t *b = copyRaw + l;
1165 uint8_t *c = copyRaw + l + l;
1167 for (int j = 0; j < l; j++)
1169 *(localRaw++) = *(a++);
1170 *(localRaw++) = *(b++);
1171 *(localRaw++) = *(c++);
1177 * \brief Convert (cY plane, cB plane, cR plane) to RGB pixels
1178 * \warning Works on all the frames at a time
1180 void PixelReadConvert::ConvertYcBcRPlanesToRGBPixels()
1182 // Remarks for YBR newbees :
1183 // YBR_FULL works very much like RGB, i.e. three samples per pixel,
1184 // just the color space is YCbCr instead of RGB. This is particularly useful
1185 // for doppler ultrasound where most of the image is grayscale
1186 // (i.e. only populates the Y components) and Cb and Cr are mostly zero,
1187 // except for the few patches of color on the image.
1188 // On such images, RLE achieves a compression ratio that is much better
1189 // than the compression ratio on an equivalent RGB image.
1191 gdcmWarningMacro("--> ConvertYcBcRPlanesToRGBPixels");
1193 uint8_t *localRaw = Raw;
1194 uint8_t *copyRaw = new uint8_t[ RawSize ];
1195 memmove( copyRaw, localRaw, RawSize );
1197 // to see the tricks about YBR_FULL, YBR_FULL_422,
1198 // YBR_PARTIAL_422, YBR_ICT, YBR_RCT have a look at :
1199 // ftp://medical.nema.org/medical/dicom/final/sup61_ft.pdf
1200 // and be *very* affraid
1203 /// \todo : find an example to see how 3rd dim and 4th dim work together
1204 int l = XSize * YSize * TSize;
1205 int nbFrames = ZSize;
1207 uint8_t *a = copyRaw + 0;
1208 uint8_t *b = copyRaw + l;
1209 uint8_t *c = copyRaw + l+ l;
1212 /// We replaced easy to understand but time consuming floating point
1213 /// computations by the 'well known' integer computation counterpart
1215 /// http://lestourtereaux.free.fr/papers/data/yuvrgb.pdf
1216 /// for code optimisation.
1218 for ( int i = 0; i < nbFrames; i++ )
1220 for ( int j = 0; j < l; j++ )
1222 R = 38142 *(*a-16) + 52298 *(*c -128);
1223 G = 38142 *(*a-16) - 26640 *(*c -128) - 12845 *(*b -128);
1224 B = 38142 *(*a-16) + 66093 *(*b -128);
1233 if (R > 255) R = 255;
1234 if (G > 255) G = 255;
1235 if (B > 255) B = 255;
1237 *(localRaw++) = (uint8_t)R;
1238 *(localRaw++) = (uint8_t)G;
1239 *(localRaw++) = (uint8_t)B;
1248 /// \brief Deals with the color decoding i.e. handle:
1249 /// - R, G, B planes (as opposed to RGB pixels)
1250 /// - YBR (various) encodings.
1251 /// - LUT[s] (or "PALETTE COLOR").
1253 void PixelReadConvert::ConvertHandleColor()
1255 //////////////////////////////////
1256 // Deal with the color decoding i.e. handle:
1257 // - R, G, B planes (as opposed to RGB pixels)
1258 // - YBR (various) encodings.
1259 // - LUT[s] (or "PALETTE COLOR").
1261 // The classification in the color decoding schema is based on the blending
1262 // of two Dicom tags values:
1263 // * "Photometric Interpretation" for which we have the cases:
1264 // - [Photo A] MONOCHROME[1|2] pictures,
1265 // - [Photo B] RGB or YBR_FULL_422 (which acts as RGB),
1266 // - [Photo C] YBR_* (with the above exception of YBR_FULL_422)
1267 // - [Photo D] "PALETTE COLOR" which indicates the presence of LUT[s].
1268 // * "Planar Configuration" for which we have the cases:
1269 // - [Planar 0] 0 then Pixels are already RGB
1270 // - [Planar 1] 1 then we have 3 planes : R, G, B,
1271 // - [Planar 2] 2 then we have 1 gray Plane and 3 LUTs
1273 // Now in theory, one could expect some coherence when blending the above
1274 // cases. For example we should not encounter files belonging at the
1275 // time to case [Planar 0] and case [Photo D].
1276 // Alas, this was only theory ! Because in practice some odd (read ill
1277 // formated Dicom) files (e.g. gdcmData/US-PAL-8-10x-echo.dcm) we encounter:
1278 // - "Planar Configuration" = 0,
1279 // - "Photometric Interpretation" = "PALETTE COLOR".
1280 // Hence gdcm will use the folowing "heuristic" in order to be tolerant
1281 // towards Dicom-non-conformant files:
1282 // << whatever the "Planar Configuration" value might be, a
1283 // "Photometric Interpretation" set to "PALETTE COLOR" forces
1284 // a LUT intervention >>
1286 // Now we are left with the following handling of the cases:
1287 // - [Planar 0] OR [Photo A] no color decoding (since respectively
1288 // Pixels are already RGB and monochrome pictures have no color :),
1289 // - [Planar 1] AND [Photo B] handled with ConvertRGBPlanesToRGBPixels()
1290 // - [Planar 1] AND [Photo C] handled with ConvertYcBcRPlanesToRGBPixels()
1291 // - [Planar 2] OR [Photo D] requires LUT intervention.
1293 gdcmDebugMacro("--> ConvertHandleColor "
1294 << "Planar Configuration " << PlanarConfiguration );
1298 // [Planar 2] OR [Photo D]: LUT intervention done outside
1299 gdcmDebugMacro("--> RawRGB : LUT intervention done outside");
1303 if ( PlanarConfiguration == 1 )
1307 // [Planar 1] AND [Photo C] (remember YBR_FULL_422 acts as RGB)
1308 gdcmDebugMacro("--> YBRFull");
1309 ConvertYcBcRPlanesToRGBPixels();
1313 // [Planar 1] AND [Photo C]
1314 gdcmDebugMacro("--> YBRFull");
1315 ConvertRGBPlanesToRGBPixels();
1320 // When planarConf is 0, and RLELossless (forbidden by Dicom norm)
1321 // pixels need to be RGB-fyied anyway
1325 gdcmDebugMacro("--> RLE Lossless");
1326 ConvertRGBPlanesToRGBPixels();
1329 // In *normal *case, when planarConf is 0, pixels are already in RGB
1332 /// Computes the Pixels Size
1333 void PixelReadConvert::ComputeRawAndRGBSizes()
1335 int bitsAllocated = BitsAllocated;
1336 // Number of "Bits Allocated" is fixed to 16 when it's 12, since
1337 // in this case we will expand the image to 16 bits (see
1338 // \ref ReadAndDecompress12BitsTo16Bits() )
1339 if ( BitsAllocated == 12 )
1344 RawSize = XSize * YSize * ZSize * TSize
1345 * ( bitsAllocated / 8 )
1349 RGBSize = 3 * RawSize; // works for 8 and 16 bits per Pixel
1357 /// Allocates room for RGB Pixels
1358 void PixelReadConvert::AllocateRGB()
1362 RGB = new uint8_t[RGBSize];
1365 /// Allocates room for RAW Pixels
1366 void PixelReadConvert::AllocateRaw()
1370 Raw = new uint8_t[RawSize];
1373 //-----------------------------------------------------------------------------
1376 * \brief Print self.
1377 * @param indent Indentation string to be prepended during printing.
1378 * @param os Stream to print to.
1380 void PixelReadConvert::Print( std::ostream &os, std::string const &indent )
1383 << "--- Pixel information -------------------------"
1386 << "Pixel Data: offset " << PixelOffset
1387 << " x(" << std::hex << PixelOffset << std::dec
1388 << ") length " << PixelDataLength
1389 << " x(" << std::hex << PixelDataLength << std::dec
1390 << ")" << std::endl;
1392 if ( IsRLELossless )
1396 RLEInfo->Print( os, indent );
1400 gdcmWarningMacro("Set as RLE file but NO RLEinfo present.");
1404 if ( IsJPEG2000 || IsJPEGLossless || IsJPEGLossy || IsJPEGLS )
1408 JPEGInfo->Print( os, indent );
1412 gdcmWarningMacro("Set as JPEG file but NO JPEGinfo present.");
1418 * \brief CallStartMethod
1420 void PixelReadConvert::CallStartMethod()
1424 CommandManager::ExecuteCommand(FH,CMD_STARTPROGRESS);
1428 * \brief CallProgressMethod
1430 void PixelReadConvert::CallProgressMethod()
1432 CommandManager::ExecuteCommand(FH,CMD_PROGRESS);
1436 * \brief CallEndMethod
1438 void PixelReadConvert::CallEndMethod()
1441 CommandManager::ExecuteCommand(FH,CMD_ENDPROGRESS);
1445 //-----------------------------------------------------------------------------
1446 } // end namespace gdcm
1448 // Note to developpers :
1449 // Here is a very detailled post from David Clunie, on the troubles caused
1450 // 'non standard' LUT and LUT description
1451 // We shall have to take it into accound in our code.
1456 Subject: Problem with VOI LUTs in Agfa and Fuji CR and GE DX images, was Re: VOI LUT issues
1457 Date: Sun, 06 Feb 2005 17:13:40 GMT
1458 From: David Clunie <dclunie@dclunie.com>
1459 Reply-To: dclunie@dclunie.com
1460 Newsgroups: comp.protocols.dicom
1461 References: <1107553502.040221.189550@o13g2000cwo.googlegroups.com>
1463 > THE LUT that comes with [my] image claims to be 16-bit, but none of the
1464 > values goes higher than 4095. That being said, though, none of my
1465 > original pixel values goes higher than that, either. I have read
1466 > elsewhere on this group that when that happens you are supposed to
1467 > adjust the LUT. Can someone be more specific? There was a thread from
1468 > 2002 where Marco and David were mentioning doing precisely that.
1475 You have encountered the well known "we know what the standard says but
1476 we are going to ignore it and do what we have been doing for almost
1477 a decade regardless" CR vendor bug. Agfa started this, but they are not
1478 the only vendor doing this now; GE and Fuji may have joined the club.
1480 Sadly, one needs to look at the LUT Data, figure out what the maximum
1481 value actually encoded is, and find the next highest power of 2 (e.g.
1482 212 in this case), to figure out what the range of the data is
1483 supposed to be. I have assumed that if the maximum value in the LUT
1484 data is less than a power of 2 minus 1 (e.g. 0xebc) then the intent
1485 of the vendor was not to use the maximum available grayscale range
1486 of the display (e.g. the maximum is 0xfff in this case). An alternative
1487 would be to scale to the actual maximum rather than a power of two.
1489 Very irritating, and in theory not totally reliable if one really
1490 intended the full 16 bits and only used, say 15, but that is extremely
1491 unlikely since everything would be too dark, and this heuristic
1494 There has never been anything in the standard that describes having
1495 to go through these convolutions. Since the only value in the
1496 standard that describes the bit depth of the LUT values is LUT
1497 Descriptor value 3 and that is (usually) always required to be
1498 either 8 or 16, it mystifies me how the creators' of these images
1499 imagine that the receiver is going to divine the range that is intended. Further, the standard is quite explicit that this 3rd
1500 value defines the range of LUT values, but as far as I am aware, all
1501 the vendors are ignoring the standard and indeed sending a third value
1504 This problem is not confined to CR, and is also seen with DX products.
1506 Typically I have seen:
1508 - Agfa CR, which usually (always ?) sends LUTs, values up to 0x0fff
1509 - Fuji CR, which occasionally send LUTs, values up to 0x03ff
1510 - GE DX, for presentation, which always have LUTs, up to 0x3fff
1512 Swissray, Siemens, Philips, Canon and Kodak never seem to send VOI LUTs
1513 at this point (which is a whole other problem). Note that the presence
1514 or absence of a VOI LUT as opposed to window values may be configurable
1515 on the modality in some cases, and I have just looked at what I happen
1516 to have received from a myriad of sites over whose configuration I have
1517 no control. This may be why the majority of Fuji images have no VOI LUTs,
1518 but a few do (or it may be the Siemens system that these Fuji images went
1519 through that perhaps added it). I do have some test Hologic DX images that
1520 are not from a clinical site that do actually get this right (a value
1521 of 12 for the third value and a max of 0xfff).
1523 Since almost every vendor that I have encountered that encodes LUTs
1524 makes this mistake, perhaps it is time to amend the standard to warn
1525 implementor's of receivers and/or sanction this bad behavior. We have
1526 talked about this in the past in WG 6 but so far everyone has been
1527 reluctant to write into the standard such a comment. Maybe it is time
1528 to try again, since if one is not aware of this problem, one cannot
1529 effectively implement display using VOI LUTs, and there is a vast
1530 installed base to contend with.
1532 I did not check presentation states, in which VOI LUTs could also be
1533 encountered, for the prevalence of this mistake, nor did I look at the
1534 encoding of Modality LUT's, which are unusual. Nor did I check digital
1535 mammography images. I would be interested to hear from anyone who has.
1539 PS. The following older thread in this newsgroup discusses this:
1541 "http://groups-beta.google.com/group/comp.protocols.dicom/browse_frm/t hread/6a033444802a35fc/0f0a9a1e35c1468e?q=voi+lut&_done=%2Fgroup%2Fcom p.protocols.dicom%2Fsearch%3Fgroup%3Dcomp.protocols.dicom%26q%3Dvoi+lu t%26qt_g%3D1%26searchnow%3DSearch+this+group%26&_doneTitle=Back+to+Sea rch&&d#0f0a9a1e35c1468e"
1543 PPS. From a historical perspective, the following may be of interest.
1545 In the original standard in 1993, all that was said about this was a
1546 reference to the corresponding such where Modality LUTs are described
1549 "The third value specifies the number of bits for each entry in the
1550 LUT Data. It shall take the value 8 or 16. The LUT Data shall be stored
1551 in a format equivalent to 8 or 16 bits allocated and high bit equal
1554 Since the high bit hint was not apparently explicit enough, a very
1555 early CP, CP 15 (submitted by Agfa as it happens), replaced this with:
1557 "The third value conveys the range of LUT entry values. It shall take
1558 the value 8 or 16, corresponding with the LUT entry value range of
1561 Note: The third value is not required for describing the
1562 LUT data and is only included for informational usage
1563 and for maintaining compatibility with ACRNEMA 2.0.
1565 The LUT Data contains the LUT entry values."
1567 That is how it read in the 1996, 1998 and 1999 editions.
1569 By the 2000 edition, Supplement 33 that introduced presentation states
1570 extensively reworked this entire section and tried to explain this in
1573 "The output range is from 0 to 2^n-1 where n is the third value of LUT
1574 Descriptor. This range is always unsigned."
1576 and also added a note to spell out what the output range meant in the
1579 "9. The output of the Window Center/Width or VOI LUT transformation
1580 is either implicitly scaled to the full range of the display device
1581 if there is no succeeding transformation defined, or implicitly scaled
1582 to the full input range of the succeeding transformation step (such as
1583 the Presentation LUT), if present. See C.11.6.1."
1585 It still reads this way in the 2004 edition.
1587 Note that LUTs in other applications than the general VOI LUT allow for
1588 values other than 8 or 16 in the third value of LUT descriptor to permit
1589 ranges other than 0 to 255 or 65535.
1591 In addition, the DX Image Module specializes the VOI LUT
1592 attributes as follows, in PS 3.3 section C.8.11.3.1.5 (added in Sup 32):
1594 "The third value specifies the number of bits for each entry in the LUT
1595 Data (analogous to ìbits storedî). It shall be between 10-16. The LUT
1596 Data shall be stored in a format equivalent to 16 ìbits allocatedî and
1597 ìhigh bitî equal to ìbits storedî - 1. The third value conveys the range
1598 of LUT entry values. These unsigned LUT entry values shall range between
1599 0 and 2^n-1, where n is the third value of the LUT Descriptor."
1601 So in the case of the GE DX for presentation images, the third value of
1602 LUT descriptor is allowed to be and probably should be 14 rather than 16.