BIRALAB

Healthcare and biomedical diagnostics

Non-invasive screening for skin cancer and metabolic disease, and clinical decision support.

The problem

Screening for skin cancer and metabolic disease in Vietnam still relies on biopsy or biochemical assays. Both are invasive, involve waiting for results, and are concentrated in central hospitals. Patients in provincial areas often present late, when treatment options have already narrowed. The barrier is not clinical knowledge but cost and distance to equipment.

How we solve it

Raman spectroscopy reads the molecular fingerprint of tissue without cutting a sample and without reagents. BIRALAB builds deep learning models for spectral classification, compresses them to run directly on the measuring device, and adds an explainable AI layer that identifies which wavenumber regions drove the conclusion. Clinicians receive a result together with its evidence rather than an unverifiable number. Because inference happens on-device, data never leaves the healthcare facility.

Core technologyHow we solve itHealthcare and biomed…RamanEdge AIXAI

BIRALAB does not place patient data on any web infrastructure. Research on human samples is carried out on separate, approved infrastructure, and only anonymised data cleared by an ethics committee is ever released openly.

Results so far

The figures below cover only the 86 publications in this research area that appear in the BIRALAB catalogue, using the same method applied to every area.

Publications in catalogue
86
Time span
2002–2026
Citations
2,506
Co-author countries
20

Most cited

HIFCF: An effective hybrid model between picture fuzzy clustering and intuitionistic fuzzy recommender systems for medical diagnosis

2015 · 187 citations

Citation counts and co-author countries come from OpenAlex, updated 9 August 2026; 85 of 86 publications were matched.

These are catalogue figures, not performance measurements of any particular system. A publication may be assigned to more than one area, so the areas add up to more than the number of records in the catalogue. An h-index per area is deliberately not shown: the measure only means something for a closed set, and computing it on a subset yields a number that is lower than reality and easy to misread.

View statistics for the whole catalogue →

Case study

The first case study is being prepared with a partner organisation. It will be published once all parties agree and, where human data is involved, once ethics approval is in place.

Related publications

View all publications

Next step

Bring us your problem

For companies with a concrete testing problem.

Bring us your problem

Propose a research collaboration

For research groups and hospitals seeking joint research.

Propose a research collaboration