Method and device for discriminating cells by electromagnetic wave scattering spectrum
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2026-02-18
- Publication Date
- 2026-04-14
AI Technical Summary
Current methods for distinguishing tumor cells from normal cells, especially under an optical microscope, are challenging due to the difficulty in observing submicrometer-sized structures and protein denaturation, and existing techniques like light scattering spectroscopy and AI analysis have limitations in accuracy and applicability.
The method involves analyzing light scattering spectra using machine learning to differentiate normal and abnormal cells by extracting feature quantities from the spectral data, which are not limited by physical interpretations, allowing for the detection of subtle differences in cell structures that cannot be observed with optical microscopes.
This approach enables high-sensitivity and accurate discrimination of tumor cells, reducing reliance on pathologist expertise and improving diagnostic accuracy beyond conventional optical microscopy.