What is it about?
This study investigates how near-infrared (NIR) spectroscopy and artificial intelligence can estimate adulteration levels in paprika powder. We analyzed mixtures containing different proportions of paprika components and compared traditional chemometric models with deep learning techniques. Partial Least Squares Regression (PLSR) achieved the best results, with R² values up to 0.981. These findings demonstrate the potential of NIR spectroscopy for rapid, non-destructive assessment of paprika authenticity and quality.
Featured Image
Photo by Diana Polekhina on Unsplash
Why is it important?
Food adulteration threatens product authenticity, consumer confidence, and fair trade. Paprika powder can be adulterated with lower-value plant materials, making reliable quality control essential. Conventional analytical methods may require extensive sample preparation and laboratory resources. Our research demonstrates that near-infrared spectroscopy combined with chemometrics can estimate adulteration levels efficiently, offering a promising approach to strengthen food authentication, improve quality control, and support transparency in the spice industry.
Perspectives
These findings open opportunities for developing portable NIR-based systems for rapid paprika authentication. Future research should validate the models using commercial samples, different paprika varieties, and additional adulterants. Integrating optimized chemometric algorithms into affordable handheld spectrometers could support on-site quality control and supply chain monitoring. Further studies should also explore model robustness under real industrial conditions and the potential application of this approach to other powdered spices and food products.
PhD Jimy Oblitas
Universidad Privada del Norte
Read the Original
This page is a summary of: Adulterant estimation in paprika powder using deep learning and chemometrics through near-infrared spectroscopy, Neural Computing and Applications, May 2024, Springer Science + Business Media,
DOI: 10.1007/s00521-024-09830-8.
You can read the full text:
Contributors
The following have contributed to this page
