~ [ source navigation ] ~ [ diff markup ] ~ [ identifier search ] ~ [ freetext search ] ~ [ file search ] ~

Linux Cross Reference
Tina6/tina-libs/ChangeLog

Version: ~
Architecture: ~ [ i386 ] ~ [ alpha ] ~ [ m68k ] ~ [ mips ] ~ [ ppc ] ~ [ sparc ] ~ [ sparc64 ] ~

  1 Project     : tina-libs
  2 Description : main tina library code
  3 ------------------------------------
  4 
  5 ChangeLog
  6 ---------
  7 
  8 This is the changelog file for the main tina library system (tina-libs).  Make an entry
  9 in this file for each change to the code stored in the central repository on server6, 
 10 copying the format of previous entries.  Entries should:
 11 
 12  - be made in reverse chronological order i.e. most recent at the top;
 13 
 14  - include details of files modified and added, details of the nature and purpose of 
 15    the change, the platform on which you are working, and any other details which
 16    other developers should be aware of;
 17 
 18  - tagged with an ID (incremented from the last).
 19 
 20 The build number in the configure.ac file in the tina-libs directory should also be 
 21 changed to the ID of the entry.
 22 
 23 
 24 -------------------------------------------------------------------------------------
 25 ID:8 pab 13/2/2012 (linux 2.6.22)
 26 
 27         . Added the docs and scripts subdirectories to the distribution tarball.
 28 
 29 
 30 -------------------------------------------------------------------------------------
 31 ID:7 pab 9/12/2011 (linux 2.6.22)
 32 
 33         . Changes to support Manual Landmark tool v1.6.
 34 
 35         . Changed the allocation of the TE_arr, TR_arr and FA_arr vectors in 
 36           dicom_hdr_TE_extract().  When a sequence is loaded from a non-zero 
 37           offset (first slice) these vectors are allocated starting at the same
 38           offset, and entered onto the props list.  When the props list is freed
 39           they are passed straight to rfree without any attempt to set the offset 
 40           back to zero.  This was causing a segementation fault when a sequence
 41           loaded from a non-zero offset was freed.  In order to avoid passing the 
 42           offset into the props list freeing functions, I simply chanegd the 
 43           allocation so that these vectors always start from zero offset.
 44 
 45           Also fixed a small memory leak.
 46 
 47 M fileDicom_io.c
 48 
 49         . Modified the limit on profile sizes from 200 to 2000 to allow smoothing with 
 50           larger Gaussian kernels (more realsitic for modern hardware).
 51 
 52 M imgPrc_conv_prof.c
 53 
 54         . Added OpnMP pragmas to allow the LSF smoothing to parallelise the LSF 
 55           smoothing algorithm across the rasters of the image being smoothed.
 56 
 57 M imgPrc_lsf.c
 58 
 59         . Two changes associated with situations where the resliced image lies partly
 60           outside the boundaries of the sequence.  First, the value entered into such
 61           regions can be set by calling seq_set_outside (default is zero).  Second, 
 62           an additional set of reslice functions have been added (seq_slicex_mask, 
 63           seq_slicey_mask, seq_slicez_mask)that return a vector of two images; the 
 64           resliced image and a mask image.  The mask image has voxel values of 1 where
 65           the resliced image voxel came from within the sequence, and 0 where it was
 66           outside the boundaries.
 67 
 68 M imgSeq_slice.c
 69 M imgSeq_slice.h
 70 
 71 -------------------------------------------------------------------------------------
 72 ID:6 nat 7/7/2011 (solaris/linux)
 73 
 74         . Software optimised to improve execution of edge detection on large images
 75           by a factor of 1000. Some debug reporting removed from linked lists to assist,
 76           and geometry hull calculation modified to work better assuming data is
 77           viewed along the z axis.
 78 
 79           Minor modifications to geometry read write and conic structure to
 80           accomodate occluding boundaries.
 81 
 82 M sysLst_dd.c
 83 M geomEdge_link.c 
 84 M geomEdge_gen.c
 85 M geomImg_canny.c
 86 M geomGen_hull.c
 87 M geomImg_canny.h
 88 M geomEdge_gen.h
 89 M geomEdge_link.h
 90 M geomLine_inter.c
 91 M geom_CurveDef.h
 92 M fileGeom3_read.c
 93 M visPgh_hist.c
 94 
 95 --------------------------------------------------------------------------------------
 96 ID:5 pab 14/12/2010 (linux 2.6.22)
 97 
 98         . Modified ralloc so that, if the attempted allocation is not possible (you 
 99           have run out of memory) it returns NULL rather than crashing TINA.
100 
101 M sysMem_ralloc.c
102 
103 --------------------------------------------------------------------------------------
104 ID:4 pab 23/06/2010 (linux 2.6.22)
105 
106         . Added OpenMP pragma to the averaging applied when loading a sequence with 
107           downsampling.
108 
109 M imgPrc_sample.c
110 
111 --------------------------------------------------------------------------------------
112 ID:3 pab 04/03/2010 (linux 2.6.22)
113 
114         . Modifications to the libraries to support the new manual landmarking tool,
115           developed as part of the collaborative project with the Max Planck Institute
116           for Evolutionary Biology.
117 
118         . Added new functionality to support colour Imrects.  Note that the colour data
119           is stored as three new fields in the Imrect structure (similar to the existing
120           data field).  However, in order to preserve backward compatibility whilst not
121           allocating huge amounts of unneeded memory, the new fields are not automatically
122           allocated when the Imrect is allocated: a call to im_col_alloc must be used to 
123           do that at the point where the new fields are needed.
124 
125 M imgGen_alloc.c
126 M imgGen_alloc.h
127 M imgGen_get.c
128 M imgGen_get.h
129 M img_GenDef.h
130 
131         . Added two new functions (imf_downsample and imf_downsample_average) that allow
132           downsampling of an image by an integer factor, with or without averaging of the 
133           pixels.  These are used by the updated version of the sequence tool, allowing 
134           sequences to be downsampled during loading to preserve memory.
135 
136 M imgPrc_sample.c
137 M imgPrc_sample.h
138 
139         . Modification to the sequence tool to allow images to be downsampled on loading
140           (for the case where the sequence is too large to fit in available memory).  A
141           new field (downsample) has been added to the sequence structure to indicate
142           this.
143 
144 M imgSeq_alloc.c
145 M img_SeqDef.h
146 
147         . Modified the DICOM reader to read the slice thickness field as well as the 
148           slice spacing field.  If the slice spacing is absent, then the slice thickness
149           can be substituted to give the voxel dimension in the inter-slice direction
150           (this arose as a problem with the micro-CT images provided by the Max Planck 
151           Institute).
152 
153 M fileDicom_io.c
154 
155 ----------------------------------------------------------------------------------------
156 ID:2 pab 06/11/2009 (linux 2.6.22)
157 
158         . Fixed the includes to use library-level, rather than file-level, headers.
159 
160 M visPgh_analysis2.c
161 
162 ----------------------------------------------------------------------------------------
163 ID:1 pab 04/11/2009 (linux 2.6.22)
164 
165         . First RC of TINA 6
166           Changes compared to TINA 5:
167 
168             - extensive rewrite of build system.
169             - all widget library interfaces except GTK+ 2 removed.
170 
171 EOF-------------------------------------------------------------------------------------

~ [ source navigation ] ~ [ diff markup ] ~ [ identifier search ] ~ [ freetext search ] ~ [ file search ] ~

This page was automatically generated by the LXR engine.
Visit the LXR main site for more information.