Apparatus and Method for Analyzing Spectrum
AssigneeJEOL Ltd.
InventorTomoki Nakao
Life of this patent
Estimating the term…
Abstract
A baseline component in an NMR spectrum (observed spectrum) is expressed as a sparse signal in a time space. In a nonlinear fitting under a condition of minimizing an Lp norm (wherein p≤1), the Lp norm functions to emphasize the sparsity of a norm computation target. Using these two properties, an optimum solution of a coefficient vector to be given to a computed vector is searched to minimize the Lp norm for a residual vector in the time space corresponding to a residual spectrum.
Claims9
An apparatus for analyzing a spectrum, comprising:
- a processor that receives an observed spectrum and that analyzes the observed spectrum, wherein
- the processor searches an optimum set of coefficients by changing a set of coefficients to be assigned to a function which defines a computed spectrum such that an evaluation value for a residual vector in a predetermined space corresponding to a residual spectrum between the observed spectrum and the computed spectrum is minimized,
- the computed spectrum is fitted to a waveform-of-interest component in the observed spectrum,
- the predetermined space is a space in which a baseline component in the observed spectrum is expressed as a sparse signal, and
- the evaluation value is an Lp norm of the residual vector in the predetermined space (wherein p≤1).
An apparatus for analyzing a spectrum, comprising:
- a receiving means that receives an observed spectrum;
- and
- a search means that searches an optimum set of coefficients by changing a set of coefficients to be assigned to a function which defines a computed spectrum such that an evaluation value for a residual vector in a predetermined space corresponding to a residual spectrum between the observed spectrum and the computed spectrum is minimized, wherein
- the computed spectrum is fitted to a waveform-of-interest component in the observed spectrum,
- the predetermined space is a space in which a baseline component in the observed spectrum is expressed as a sparse signal, and
- the evaluation value is an Lp norm of the residual vector in the predetermined space (wherein p≤1).
A method for analyzing a spectrum, comprising:
- receiving an NMR spectrum obtained by an NMR measurement;
- and
- searching an optimum set of coefficients by changing a set of coefficients to be assigned to a function which defines a computed spectrum such that an evaluation value for a residual vector in a predetermined space corresponding to a residual spectrum between the NMR spectrum and the computed spectrum is minimized, wherein
- the computed spectrum is fitted to a waveform-of-interest component in the NMR spectrum,
- the predetermined space is a space in which a baseline component in the NMR spectrum is expressed as a sparse signal, and
- the evaluation value is an Lp norm of the residual vector in the predetermined space (wherein p≤1).
Dependencies are inferred from the claim text.