Elemental classification at single-sample resolution
One micro-laser pulse produces a full elemental spectrum from a single sample. A supervised model turns that spectrum into a class assignment you can act on. Two instruments run the same engine: MicroLIBS™ for high-resolution analytical work, Laser Classifier™ HT for production throughput.
Core Capability
We use micro-laser elemental analysis combined with supervised multivariate modeling to convert spectral data into actionable classification outputs. The method is matrix-agnostic. Seed is where we work today, and it is one of several areas the platform already covers.
Bulk chemical tests provide averages. Optical inspection evaluates appearance. Neither captures classification at the resolution of the individual sample.
Elemental signatures reveal statistically significant discrimination between classes, within-lot variability, and hidden structure not visible optically. That is what makes an evidence-based go/no-go decision possible.
Seed Classification & QC
The Laser Classifier™ platform performs rapid elemental analysis of individual seeds using a single micro-laser pulse per seed. Each seed generates a characteristic elemental signature, processed by a supervised classification model to assign the seed to a defined class.
Conventional methods, whether chemical or optical, lack single-seed resolution. By operating at the individual seed scale, within-lot variability becomes measurable and decision criteria become quantifiable.
Applications
The classifier does not know what it is looking at. It measures elemental emission and a supervised model assigns a class. That is why the same platform has been applied to seed, agricultural products, diseased plant tissue, insect vectors, fossil resin and human tissue.
The core commercial application. Seeds that are visually identical can carry different genotypes, and no optical or morphological inspection separates them. Elemental signatures do.
In an ornamental species program run under a commercial agreement, the platform was applied to single versus double seed classification on customer material, with the model built and cross-validated on the customer's own defined classes. Work of this kind is covered by confidentiality, so results are discussed under agreement rather than published.
Quantify the fraction of off-type or anomalous seed in a lot, compare lots against a reference model, and set decision thresholds that are statistical rather than visual.
Run on customer material under commercial agreement.
Elemental discrimination of genome dosage in intact seed, aimed at breeding programs that need ploidy state early and without consuming the seed. Internal feasibility and proof-of-concept work is active.
Provisional patent filed June 2026. Not yet offered as a service.
Variety and geographic origin from elemental signature. Arabica separated from Robusta in coffee, and quality control plus origin identification in handmade cigars.
Optics and Photonics Journal 2017; Applied Optics 2015.
Huanglongbing-infected citrus identified from phloem samples, and bacterial infection detected inside the insect vector itself. Two different biological matrices, same supervised pipeline.
Applied Optics 2018; Scientific Reports 2019.
Classification of amber samples by LIBS combined with chemometric methods. Evidence that the approach is not limited to living tissue.
Revista Cubana de Fisica 2023.
Fast detection of prostate malignant tissue using multipulsed LIBS. The most demanding classification problem the group has published, and the clearest proof that the method generalizes.
Revista Cubana de Fisica 2022.
Six domains, one measurement principle and one modeling approach. The published record behind each of these is in Scientific Foundation below.
Process
From sample intake to classification output, traceable at every step.
Seeds are logged and tracked under controlled conditions. Each seed is evaluated individually, so every classification result corresponds to a specific seed.
A precisely controlled micro-laser pulse interacts with the seed surface, generating a transient plasma emission with multi-element spectral information.
Raw spectral data are processed using ProLIBSpector™, performing signal normalization, feature extraction, supervised model application, and class assignment with confidence scores.
Transforms analytical data into structured decision support.
Instruments
Both instruments run ProLIBSpector™ and produce the same class outputs from the same supervised models. What differs is how a sample gets in front of the laser, and how many samples per hour that allows.
High spatial resolution
MicroLIBS™ Analysis System
High spatial resolution single-sample analysis. Built for micro-analytical work where position on the sample matters: embryo-level measurement, sub-seed tissue mapping, and protocol development.
Production throughput
Laser Classifier™ HT
Automated platform for production-scale classification. Samples are loaded in standard SBS plates, one sample per well, and the stage indexes through them without individual handling.
The pulse ablates a nanogram-scale volume and the seed stays intact and plantable, which is what makes the measurement usable ahead of planting rather than instead of it. The fair way to show that is at two scales, because the number only means something next to the seed it was made on.
Technology
Laser Classifier™ is based on laser-induced breakdown spectroscopy (LIBS), adapted for single-sample surface verification. A focused pulse ablates a nanogram-scale volume, the resulting micro-plasma emits, and the spectrum resolves elemental composition with little or no sample preparation. The platform builds on decades of laser spectroscopy development and classification research by the same group.
Our Lab & Equipment
Scientific Foundation
The same measurement principle and the same supervised classification approach, applied to biological matrices that have almost nothing in common with each other: coffee bean, cigar leaf, citrus phloem, insect vector, fossil resin, prostate tissue. Each one peer-reviewed. That breadth is the argument for the platform, and it is why the application areas above are not a roadmap.
LIBS-based bacterial infection identification in insect vectors of Huanglongbing (HLB) disease.
Laser-Induced Breakdown Spectroscopy (LIBS) as a novel technique for detecting bacterial infection in insects
Rapid identification of HLB-infected citrus plants through phloem analysis using LIBS.
Rapid identification of Huanglongbing-infected citrus plants using laser-induced breakdown spectroscopy of phloem samples
Spectral differentiation between Arabica and Robusta coffee varieties.
Laser-Induced Breakdown Spectroscopy (LIBS) Applied in the Differentiation of Arabica and Robusta Coffee
Quality control and geographic origin identification of handmade cigars.
Laser-Induced Breakdown Spectroscopy (LIBS) Quality Control and Origin Identification of Handmade Manufactured Cigars
Fast detection of prostate malignant tissue using multipulsed LIBS.
Fast Detection of Prostate Malignant Tissue by Multipulsed Laser-Induced Breakdown Spectroscopy (LIBS)
Analysis and classification of amber samples using LIBS and chemometric methods.
Analysis of Amber Samples by LIBS and Chemometrics Methods
Every study above used LIBS combined with supervised classification models. That is the same core methodology running on Laser Classifier™ today, on both instruments, for seed and for everything else.
Get Started
Laser Classifier™ is introduced through a structured pilot engagement designed to evaluate seed-level classification performance using your defined samples. This is a formal technical evaluation, not a demonstration or free trial.
