This page lists the bugs, bug fixes, and minor changes to Jim version 3.0. Any minor changes should not affect core functionality of Jim unless otherwise indicated. Fixed versions of Jim have a minor release indicated by a letter such as "3.0_j". You can see which version you currently have installed from the Help -> About menu. Comparison with the table below will indicate whether a bug you have noted needs to be reported, or has already been fixed.
If you want to install the latest release of Jim, you can either do a full re-install, or you can simply use this link to download the file xinapse.jar. Click right on the link above, and save to a temporary location before using it to replace the file called xinapse.jar in the install directory. If you don't know which is the install directory, do:
find / -name "xinapse.jar" -print
xinapse.jar.
xinapse.jar file before replacing it with the
updated version.
Important Note: if you do a full re-install of the latest version of Jim 3.0, and have performed any customisation of the installation (such as editing the Unix shell script Jim), then please be sure and make backup copies of these changed files in another directory before installing Jim again. If your license.dat file is in the Jim installation directory, then please also ensure that you make a backup copy of license.dat.
Key: type of issue
| Type | Bug/Enhancement Description | Fix Description | Bug ID | Status | Version / date |
|---|---|---|---|---|---|
| b | "Go to Slice" dialog does not function when a new image is loaded into Jim's display. | Fixed. Dialog is now re-enabled when a new image is loaded.30-052 | F | 3.0_aa, 01 Sep 2006 | |
| B | AIF lead/lag feature does not function correctly. This bug affects both the DCE-MRI and the Perfusion toolkits. The time-shift of the data for the tissue response relative to the AIF does not happen. That shift in time was happening for both the tissue response and the AIF and either the AIF or the tissue response were being corrupted (depending on whether you set a lead or a lag). Results where a zero lead/lag was entered are unaffected by this bug. | Code has been reworked to provide the correct shift in time between the AIF and the tissue response when a lead/lag is non-zero. Tooltips have been updated to make it clear that the "time of contrast arrival" is the time at which the signal changes (due to contrast agent) either in the tissue or in the vessel where the AIF is measured (whichever comes first). | 30-051 | F | 3.0_z, 15 May 2006 |
| B |
Update to bug IDs 30-041 and 30-010. Incorrect scaling of Philips DICOM (not
ACR-NEMA) images, where Rescale Slope and Rescale Intensity are
used in Philips MR images to pack the images into 8/12
bits. Jim used the Rescale Slope and Rescale Intensity values
incorrectly when rescale the images as the are imported/converted so
that intensties across different scans/patients were not comparable.
This bug should only affect the intensities of Philips DICOM images, and is only an issue if intensities across different scans need to be comparable (e.g., when calculating MTR values). |
The DICOM elements <0x0028, 0x1052> Rescale Intercept and <0x0028, 0x1053> Rescale Slope are now correctly taken into account when converting to UNC/Analyze formats. The intensity in the converted (UNC/Analyze) is now the intensity of the image straight after the FFT, so that intensities for the same coil are comparable. | 30-050 | F | 3.0_y, 18 April 2006 |
| b |
Problem reading saved ROI files under the following circumstances:
|
Incorrect escaping of the backslash character "\" character used in Microsoft Windows path names. | 30-049 | F | 3.0_x, 23 Jan 2006 |
| b | Dicom transfers fail when requestor presentention context ID greater than 127. | Incorrect reading of context ID fixed. Caused by bad conversion from unsigned byte value. | 30-048 | F | 3.0_w, 8 Dec 2005 |
| b | Animated GIF image not exported from Movie correctly when image display area is not square (GIF may appear squashed or stretched). | Fixed. | 30-047 | F | 3.0_v, 22 Nov 2005 |
| b | Formula field of user presets not stored/restored correctly in Algebra Tool. | Fixed. | 30-046 | F | 3.0_u, 21 Nov 2005 |
| b | Attempted display of images with PixelDataType.INT pixels, and a large data range (with minimum intensity close to Integer.MIN and maximum intensity close to Integer.MAX, caused Jim to hang and eventually show an OutOfMemory error. | Fixed. | 30-045 | F | 3.0_t, 12 Oct 2005 |
| b | In Jim Preferences, preferred image class (Analyze, UNC) is not correctly retrieved from the user preferences. | Fixed. | 30-044 | F | 3.0_t, 12 Oct 2005 |
| b | In Image Fitter, user presets do not set the name and scaling factor for the independent variable. | Fixed. | 30-043 | F | 3.0_s, 29 Sep 2005 |
| B | In Image Fitter, when an output variable scaling factor is used, pixels may be excluded in output images even if fitted value falls within the dynamic range of image, leading to "drop-outs" even for a valid fit for that pixel. | Fixed. | 30-042 | F | 3.0_s, 29 Sep 2005 |
| E |
Update to bug ID 30-010. Incorrect scaling of Philips DICOM (not
ACR-NEMA) images, where Rescale Slope and Rescale Intensity are
(incorrectly) used in Philips MR images to pack the images into 8
bits. Jim now uses the Rescale Slope and Rescale Intensity values to
rescale the images as the are imported/converted.
This bug should only affect the intensities of Philips DICOM images, and is only an issue if intensities across different scans need to be comparable (e.g., when calculating MTR values). |
The DICOM elements <0x0028, 0x1052> Rescale Intercept and <0x0028, 0x1053> Rescale Slope are taken into account when converting to UNC/Analyze formats. The output images are appropriately rescaled. | 30-041 | - | 3.0_r, 29 July 2005 |
| e | Automatic finding of AIF in Brain Perfusion tool is disproportionately slow for images with not very many pixels, but with a long time series (large number of time points). | AIF finding has been speeded by allowing the user to set a smaller number of "candidate" pixels. The candidates can be quickly found, and then finding the final AIF pixels from amongst these is quicker. | - | - | 3.0_r, 29 Jul 2005 |
| b | Brain Finder tool failed to find a suitable CSF/Brain threshold when working with images that have distinct grey matter/white matter histogram peaks. | Improvements made to Brain Finder tool, to give better fitting to histogram for images with separate GM/WM peaks. Also, volumes of grey matter and white matter are now separately reported. | 30-040 | F | 3.0_r, 29 Jul 2005 |
| b | Image algebra tool fails when working with very large images. | Fixed. | 30-039 | F | 3.0_r, 29 Jul 2005 |
| e | Keyboard accelerator added to allow contour ROI generation on a key press. | Pressing 'c' or 'C' on the keyboard generates a contour when the ROI Toolkit is active and the image display pane has the input focus. | - | - | 3.0_r, 29 Jul 2005 |
| b |
Brain finder tool failed at the BPF calculation stage when:
|
Brain Finder now operates correctly on imported images. | 30-038 | F | 3.0_r, 29 Jul 2005 |
| b | Masker tool did not work with dispim ROIs. Dispim ROI file could not be opened. | Masker tool will now work with both Jim ROIs and dispim ROIs. | 30-037 | F | 3.0_q, 20 May 2005 |
| B | DCE-MRI analysis tool failed to respond to changes in the user interface to the arterial haematocrit value. The default value (0.45) was always used. | Fixed. | 30-036 | F | 3.0_q, 20 May 2005 |
| B | Image profiles causes java stack trace to be shown when applied to multi-slice Analyze and UNC RGB colour images. Also, image masking undo operations fails on these images when masking is applied to any other slice than slice 1. | Fixed. | 30-035 | F | 3.0_q, 20 May 2005 |
| B | When editing Irregular and CurvedLine ROIs using the eraser, unexpected behaviour sometimes resulted. For example, using the eraser to divide an ROI into two ROIs sometimes resulted in three ROIs being produced, with one ROI being similar to the original undivided ROI. | Interaction with eraser reworked to give consistent results. | 30-034 | F | 3.0_p, 14 Feb 2005 |
| E | Perfusion and DCE-MRI tools have been provided with an option to allow a lead or lag between the measured arterial input function and the actual input to the tissue. If the measured input function is significantly upstream (proximal) compared to the artery that feeds the tissue, then this can be compensated by providing a "lead" time. Similarly, if the measured input function is actually a venous drain, then this can be compensated by providing a "lag" time. | - | - | 3.0_p, 14 Feb 2005 | |
| b | Various tools (Image Resampler, Algebra, Fitter, DCE-MRI and Perfusion tools) would give error messages when valid numerical values were inputted into text fields of the GUIs, but there were spurious space " " characters after the numerical values. | Fixed. | 30-033 | F | 3.0_n, 20 Jan 2005 |
| B | Images with signed byte pixel values were incorrectly displayed in Jim. | Pixel values were incorrectly being interpreted as unsigned byte values. Affected signed byte images only. | 30-032 | F | 3.0_n, 20 Jan 2005 |
| B | Tools that use Levenberg-Marquart fitting, such as the DCE-MRI tool, BrainFinder tool, and Image Fitter tool failed to converge on the proper solution. | Bug in Levenburg-Marquart introduced in version 3.0_l prevented convergence. | 30-031 | F | 3.0_n, 20 Jan 2005 |
| b | Image converter shows message "invalid destination directory" when Jim is started with an explicit image argument. | System property user.dir incorrectly set when Jim started with an explicit image argument. | 30-030 | F | 3.0_n, 20 Jan 2005 |
| b | When using Masker tool with ROIs as the masks, deleted ROIs were used as masks just as for undeleted ROIs. | Masker tool now checks that an ROI is not deleted before using it. | 30-029 | F | 3.0_m, 13 Dec 2004 |
| b | When using Brain Finder Tool to calculate BPF, BPF was sometimes found to be 1.0 or 0.0. | Fitting peaks to intensity histogram (to separate brain from CSF) has been improved to make it more reliable. Please report any further problems. | 30-028 | F | 3.0_l, 24 Nov 2004 |
| b |
When using Interleaver Tool to interleave Analyze images, error
message: interleaver: problem writing output image: patient position is not supported in Analyze images was shown. |
Fixed. | 30-027 | F | 3.0_l, 24 Nov 2004 |
| b | Some DICOM images with data elements having VR's of UNKNOWN and undefined length cannot be Imported/Converted by Jim. | Fixed. | 30-026 | F | 3.0_l, 24 Nov 2004 |
| b | DCE-MRI tool failed to give a result when the computed volume fraction has greater than 1.0. | Although a Ve greater than 1 is physically non-meaningful, this can occur if the user supplies a synthetic AIF that has a wrong scaling, of the relaxivities for plasma/extra-cellular space supplied by the user are very wrong. Although under these circumstances the computed Ve will be wrong, the computed Ktrans would be valid and is now reported. | 30-025 | F | 3.0_l, 24 Nov 2004 |
| b | Some DICOM images with DCT Huffman coded compressed pixel data could not be Imported/Converted by Jim. Those images with more than one Huffman table defined were affected by this bug. | Fixed. Images with more than one Huffman table can now be Imported/Converted. | 30-024 | F | 3.0_k, 1 Nov 2004 |
| b | After using ROI totaliser to calculate total ROI volume, when an ROI is selected, the intensity-related statistics (mean intensity, sd, min and max) are sometimes not shown. This happened because intensity-related statistics do not need to be calculated in order to calculate the ROI volume, and their cached values were zero. | Fixed. Cached statistics are cleared after totalising volumes. | 30-023 | F | 3.0_k, 1 Nov 2004 |
| b | Perfusion tool sometimes failed when ROI used to define arterial input function was a Marker ROI. | Fixed. A single marker can now be used to define a single pixel for an arterial input function. | 30-022 | F | 3.0_k, 1 Nov 2004 |
| b | Info Viewer sometimes fails to pop-up when certain Analyze images are displayed with a selected slice. | Fixed. | 30-021 | F | 3.0_k, 1 Nov 2004 |
| B | Image fitter sometimes failed to converge for all pixels. | Levenburg-Marquart routine failed to re-initialise correctly when using same fitter for multiple pixels. Re-initialisation is now more rigorous. | 30-020 | F | 3.0_k, 1 Nov 2004 |
| b | Poor packing behaviour of Image Fitter tool window when variable name was changed, led to variable value field being very small. Needed user to stretch window again before variable values could be entered. | Packing behaviour fixed. | 30-019 | F | 3.0_j, 14 Oct 2004 |
| b | Some Analyze images with malformed headers could not be displayed. Analyze DB name strings should be null-terminated in the .hdr file. | Now checks for non-terminated strings and takes corrective action. | 30-018 | F | 3.0_j, 14 Oct 2004 |
| B |
Pixel sizes incorrectly set for DICOM images with non-square pixels. This
bug only affects those imported/converted images with different pixel
width and height. N.B. This bug does not affect computed areas or volumes, but could affect computed linear dimensions but only for imported DICOM images with non-square pixels. |
DICOM Pixel Spacing (0x0028, 0x0030) is Y_SPACING\X_SPACING. The spacing was incorrectly interpreted as X_SPACING\Y_SPACING, leading to width and height of pixels being interchanged. | 30-017 | F | 3.0_j, 14 Oct 2004 |
| b | Orthogonal views not shown with image overlay when there are different PixelDataTypes for main image and overlay image, and axial or coronal main orientation selected. | Fixed. | 30-016 | F | 3.0_j, 1 Nov 2004 |
| e | Image Concatenator tool now uses a MultipleImageSelectionPanel rather than a ImageSelectionGroupPanel to select the input images. This was implemented to allow easier processing of large numbers of images to be concatenated. | - | - | 3.0_i, 24 Sep 2004 | |
| B | Image conversion failed to set image pixel sizes correctly for DICOM images where the pixel spacing is encoded as a pair of floating point values. Older DICOM images have their pixel spacing is encoded as a string value. | Fixed. | 30-015 | F | 3.0_i, 24 Sep 2004 |
| b | DCE Analysis tool did not provide area under the curve of the [Gd] values when "Tofts with vp term" analysis was performed. | Fixed. | 30-014 | F | 3.0_i, 24 Sep 2004 |
| B | DCE Analysis tool did not allow correct ROI-based analysis when for images acquired and analysed as inversion-recovery or saturation recovery. | Signal intensity for "M0" image was not being correctly read for the region of interest. Tool now requires ROI file to be explicitly specified for M0 image for (a) the region to be analysed, and (b) the AIF. | 30-013 | F | 3.0_i, 24 Sep 2004 |
| e | When loading duplicate ROIs from an ROI file, then a message warning of the duplicate ROIs pops up for every duplicate ROI that was not loaded to Jim. This could was annoying if there were lots of duplicate ROIs. | The pop up now allows the user to say "OK to All" to suppress further warning messages. | 30-012 | F | 3.0_h, 7 Sep 2004 |
| b | Image Converter and Importer incorrectly set the value of "scan TR" in UNC images. The incorrect value set was equal to the flip angle. | Correct value of scan TR now set in UNC images. | 30-011 | F | 3.0_h, 7 Sep 2004 |
| E |
The individual slices of Philips DICOM (not ACR-NEMA) images may have
different intensity scaling factors so that their intensities cannot
be quantitatively compares. A mechanism to overcome this has been
implemented so that the pixel intensities of slices within a series,
or between series for the same scanner and coil combination are
comparable.
This enhancement only affects the intensities of Philips DICOM images. |
Instead of the DICOM pixel values (stored values, SV) being stored in the converted
images, the intensities straight after the FFT (FP values) are stored in the the
converted images. The scaling factor to convert between the SV and FP
values is:
FP = (SV - si) / ss / 128 where si the value of "scale intercept", DICOM element <0x0029 0x1052>, and ss the value of "scale slope", DICOM element <0x0029 0x1053>. The factor of 128 is included to prevent the FP values overflowing the intensity range of the 16-bit pixels. |
30-010 | - | 3.0_h, 7 Sep 2004 |
| b | Brain and CSF volumes incorrectly reported in cubic mm rather than ml in BrainFinder tool. Although the stated units were ml, the actual volumes reported were in cubic mm. Previous results should be divided by 1000 to get the result in ml. | The volumes reported and the corresponding units are now consistently millilitres (ml). | 30-009 | F | 3.0_h, 7 Sep 2004 |
| e | Modified DICOM image converter to handle Philips images from an Easy Vision with badly-encoded sequences of items. | - | - | F | 3.0_h, 7 Sep 2004 |
| e |
Additional items written to DCEMRI log file when performing
region-of-interest analysis of Ktrans. Additional items are:
|
- | - | - | 3.0_g, 25 Jul 2004 |
| b | Under exceptional circumstances, saving of ROI statistics to ROI file failed, writing the message "error writing statistics" to the ROI file. This prevented the ROIs being read in. | The ROI reader now handles this exceptional case. | 30-008 | F | 3.0_g, 25 Jul 2004 |
| b | When image saving from within Jim failed, a warning message appeared in the status bar, but there was no pop-up error message. | A pop-up message has now been provided. | 30-007 | F | 3.0_f, 22 Jun 2004 |
| b | Image overlay shows overlaid image with incorrect contrast and brightness settings when opacity value adjusted. When user makes changes to contrast and brightness using the sliders for the overlay image, making them different from the auto-contrast settings, these new values are ignored when adjustments are made to the opacity. | Fixed. | 30-006 | F | 3.0_f, 22 Jun 2004 |
| b |
Perfusion and DCE-MRI analysis do not work when the arterial input
function is defined by the user using a single ROI at a single time point (i.e.,
the AIF is not defined by an ROI at every time point).
Java stack trace shown in console window starting: |
Incorrect program logic tried to read an ROI from every slice regardless. | 30-005 | F | 3.0_f, 22 Jun 2004 |
| b |
Image Resampler does not work with Analyze images when cropping,
padding or reorienting images. Reports error: "image
position (patient) is not supported in Analyze images".
|
Exception thrown by Analyze images when trying to set unsupported image position in now caught and handled. Usage text shown by stand-alone Resampler tool running in text-mode has also been improved. | 30-004 | F | 3.0_e, 29 Jun 2004 |
| e | Added ability of Masker Tool to use ROIs to perform masking operations. Previously, only masking using an image as a mask was possible. Also, documentation about ROI combination operations for masking within ROI toolkit improved and cross-referenced to Masker Tool. | This features now allows batch processing of ROI masking using the stand-alone Masker Tool running in text mode. | - | - | 3.0_d, 26 May 2004 |
| B | Normalisation of histogram in Image Statistics display incorrect. Normalisation was incorrectly dividing by the total number of pixels in the image, rather than just the number of pixels in the visible portion of the histogram. | The sum of all pixel counts in normalised histograms now sums to 1.0. | 30-003 | F | 3.0_c, 30 Apr 2004 |
| b | Annotation of position of profile graph is incorrect on Profiles popup dialog. y-position of vertical profiles was always labelled "x", and positions were reported with no significant figures after decimal point. | Fixed. Documentation on MTR task-based help also improved and updated. | 30-002 | F | 3.0_c, 30 Apr 2004 |
| e | Image Resampler, Perfusion and DCE-MRI tools made more compact. | Tools tended not to fit completely on screen when number of screen pixels was low. Made more compact by only displaying those parts that are needed, and removing extraneous space. | - | F | 3.0_b, 08 Apr 2004 |
| b | GIF export sometimes fails when ROIs are displayed on image. | GIF colour compaction is now more conservative to allow for extra colours used when displaying ROIs. | 30-001 | F | 3.0_b, 08 Apr 2004 |
| - | New release. | - | - | - | 3.0_a, 01 Apr 2004 |