Peak Waveform Inference for Overlapping Asymmetric Chromatography Peaks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing peak detection methods in chromatography face challenges in accurately determining peak positions and areas due to asymmetric peak waveforms caused by tailing, especially when peaks overlap, leading to errors in quantitative analysis, particularly in simultaneous analysis of multiple components.
Innovation Solution
A method and device using adversarial learning with generative adversarial networks (GAN) to infer model functions and shape distributions for peak waveforms, allowing precise peak detection and separation without increasing the number of parameters, by extracting partial waveforms and performing adversarial learning with generation and discriminative models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If model function fitting is used to improve peak detection accuracy, then measurement precision improves, but the method fails when waveforms have asymmetric shapes due to tailing or overlapping peaks
Solution Approach 1:
The invention transforms the waveform representation by applying logarithmic transformation to the amplitude values and reversing the time axis. This parameter transformation converts asymmetric peak waveforms with tailing into symmetric shapes that can be accurately modeled by Gaussian functions, thereby enabling precise peak detection while handling asymmetric real-world chromatogram data
Solution Approach 2:
The invention inverts the time axis by setting t' = -t, which reverses the direction of the waveform. This inversion, combined with logarithmic amplitude transformation, converts right-skewed asymmetric peaks into left-skewed or symmetric forms that fit Gaussian modeling better, allowing the model function fitting approach to work effectively on previously unsuitable asymmetric data
2Measurement precision
If the number of parameters in the model function is increased to improve fitting accuracy for asymmetric peaks, then measurement precision improves, but device complexity increases
Solution Approach 1:
Instead of adding more parameters to the Gaussian function to model asymmetric peaks, the invention transforms the data parameters (logarithmic amplitude and reversed time) so that the simple two-parameter Gaussian model can accurately fit the transformed asymmetric peak data, maintaining model simplicity while achieving high fitting accuracy
Solution Approach 2:
The invention replaces the complex mechanical approach of adding more parameters to the model function with a mathematical transformation approach. By transforming the coordinate system and amplitude scale, the simple Gaussian model becomes sufficient to represent complex asymmetric peak shapes without increasing model complexity
Data Source
AI summary
A waveform information inference device according to one mode of the present invention includes: a waveform extraction unit (31) configured to extract a partial waveform to be modeled from a signal waveform acquired based on actual measurement using a predetermined analysis device; and an adversarial learning unit (32) configured to acquire a model function corresponding to the partial waveform, or the model function and shape distribution information in the function by performing adversarial learning using two mutually adversarial models which are a generation model and a discriminative model using the partial waveform obtained by the waveform extraction unit as an input. The present invention can acquire a precision peak model function and its shape parameter distribution information.