PhotoMetrix UVC: Smartphone-Based Colorimetric Analysis of Milk Adulterants

A smartphone-based system for digital image colorimetric analysis using partial least squares regression. The method was validated by determining four adulterants in raw milk samples, achieving coefficients of determination higher than 0.99. The research was authored by Adilson Ben da Costa and published on Papers with Code.

Use Cases

• Detecting chemical adulterants in food products based on the colorimetric analysis described.
• Building regression models for quantitative analysis using RGB histogram data from digital images.
• Developing point-of-use analytical applications for smartphones based on the described hardware and software method.

Strengths

• Regression models for all four adulterants showed a coefficient of determination (R2Cal) higher than 0.99.
• Root mean square error of prediction (RMSEP) values for the adulterants ranged from 0.28 to 0.48 g L-1.
• Recovery rates in spiked milk samples ranged from 90 to 120%.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect bias inherent to a single proof-of-concept study on milk adulteration.

Provenance

Source

Adilson Ben da Costa

Collection Method

Images captured with a UVC in a 3D-printed chamber, converted to RGB histograms, and analyzed with PLS regression models within a smartphone app.