spectrIm-QMRS
Version 3.0.1 Alpha
Version 3.0.1 Alpha packaging
Older spectrIm-QMRS documentation treated HLSVDPro as functionality that might need separate enabling or local installation. In version 3.0.1 Alpha, the intended clinical distribution is a self-contained ZIP package: Java, Python, and the required HLSVD-related runtime components are bundled with the application.
The main user-facing change is that clinicians should not normally need to install Java, Python, or HLSVDPro separately on a hospital workstation. This is important in clinical environments where local installation of external runtimes is restricted or prohibited.
HLSVDPro is now part of the spectrIm-QMRS preprocessing workflow: it can be called from the Processing menu and the main preprocessing toolbar. The non-interactive HLSVDPro command uses default settings, while the preferred configurable route is now the version 3.0.1 Alpha HLSVD filter-definition workflow.
Processing > Remove Water (HLSVDPro) applies non-interactive HLSVDPro water removal using the current defaults. The same Processing menu also provides interactive HLSVD, HLSVD filter definition, and the 64-bit HLSVD implementation. For new v3 documentation, the most important configurable entry point is Processing > Define HLSVD Filter....
The new filter-definition window contains HLSVD settings, workflow order, denoising options, and a spectrum preview. It is the interface that should be documented for version 3.0 alpha and refined further for version 3.1. Its design goal is to make HLSVD prefiltering visible and reversible enough for practical work: preview first, apply the ordered cascade to selected spectra, follow the progress, and undo the last workflow application if the chosen parameters were not appropriate.
HLSVDPro can also be represented as a batch-processing command with lower limit, upper limit, and number-of-peaks parameters. This is useful when the same preprocessing should be applied repeatedly across datasets.
The following windows still exist in version 3.0.1 Alpha, but they are retained for compatibility and transitional use. They should not be treated as the primary documentation target for new users. The plan is to remove or replace these legacy HLSVD windows in version 3.1.
Older combined HLSVD analysis window with water range, broad component filtering, component table, and direct apply/quit controls.
Older low-level settings window for acquisition defaults, HLSVD decomposition parameters, legacy filter defaults, execution mode, and runtime paths.
The current Java defaults show how HLSVD and HLSVDPro are initialized when no user preference file exists. These values should be treated as defaults, not universal recommendations; acquisition protocol, field strength, nucleus, residual water behavior, and artifact type can all require different settings.
10241120304.4#5.4 ppmWater only32450 HzVersion 3.0.1 Alpha exposes HLSVD as a configurable preprocessing tool, not only as a fixed water-removal button. The Java defaults include several named filter modes:
The v3 code also defines a saved HLSVD preprocessing cascade with ordered steps. HLSVD is therefore no longer documented only as a water-removal tool: the workflow may contain water removal, extreme broad artifact cleanup, denoising, smiley artifact cleanup, and lipid removal. Each row can be enabled or disabled, and the order column determines the sequence in which enabled rows are applied. The table below is an example workflow and should be adapted to the case under investigation:
| Enabled | Order | HLSVD prefilter | Ranges [ppm] | Hankel | Components | Width [Hz] |
|---|---|---|---|---|---|---|
| Yes | 1 | 3. Water removal | 4.4#5.4 |
30 | 10 | N.A. |
| Yes | 2 | 2. Extreme broad artifact cleanup | N.A. | 180 | 40 | 25 |
| Yes | 3 | 1. Smiley artifact cleanup | -0.2#0.1;9.3#9.6 |
120 | 30 | N.A. |
| Yes | 4 | 4. Lipid removal | 0.0#1.6 |
20 | 10 | 1 |
| Yes | 5 | 5. Denoising | 1.7#4.2 |
160 | 140 | N.A. |
The order column controls the cascade applied by the workflow. The numeric prefix in the prefilter name is the internal filter type, so the display can legitimately show, for example, order 1 with 3. Water removal. In this default configuration only water all rows are enabled for illustration. This does not mean that all rows should be enabled for every dataset.
The reason for separating these rows is that HLSVD first estimates damped complex sinusoids and then subtracts only the components that match the selected filter rule. Water removal, lipid removal, broad component suppression, edge-artifact cleanup, and denoising therefore need different ranges, component counts, Hankel settings, and sometimes different positions in the workflow.
To be expanded: add practical default-setting recommendations for typical SVS and MRSI datasets after more examples have been collected.
For ordinary v3 alpha use, configuration should be minimal. Manual checks are mainly needed when running a developer build, moving the application folder after first launch, replacing bundled Python or HLSVD binaries, or diagnosing a failed HLSVDPro call.
The old interactive HLSVD settings GUI should be considered deprecated documentation-wise. It remains available in version 3.0 alpha for compatibility, but it is planned to be replaced completely in release 3.1 by a redesigned configuration interface. This deprecation concerns the old settings window only, not the new HLSVDPro / denoising functionality.
Legacy HLSVD-64 settings still include DLL paths such as
callfortran_hlsvd_arr.dll, libhlsvd.dll,
and supporting runtime libraries. In the bundled v3 package these
should be placed relative to the spectrIm root and should not
require system-wide installation.
Algorithm background, denoising use cases, and reference to the original Delft/Pijnappel HLSVD paper.
Menu-level documentation of HLSVDPro, interactive HLSVD, HLSVD filter definition, and HLSVD 64-bit commands.
Where HLSVD fits in the larger path from loading a dataset to preprocessing, maps, and TDFDFit quantification.