Smartphone Video Imaging Combined with Machine Learning: A Cost-Effective Method for Authenticating Whey Protein...

Area of Science:

• Analytical Chemistry
• Food Science
• Machine Learning

Background:

• Rising demand for whey protein supplements necessitates cost-effective quality control methods.
• Current authentication techniques can be expensive and time-consuming.
• Ensuring the authenticity of sports nutrition products is crucial for consumer trust.

Purpose of the Study:

• To develop a rapid, low-cost method for authenticating sports whey protein supplements.
• To utilize smartphone video imaging (SVI) combined with machine learning for quality characterization.
• To assess the performance of SVI against established techniques like hyperspectral imaging (HSI).

Main Methods:

• Smartphone video imaging (SVI) was employed, using a smartphone screen to illuminate samples and its camera to capture color changes.
• Video data was processed into spectral data, and machine learning models established relationships between video data and sample characteristics.
• The method was validated across tasks including brand identification, fat and energy content quantification, and adulterant detection.

Main Results:

• SVI achieved high accuracy in whey protein concentrate (WPC) brand identification (0.933) and adulterant detection (0.96).
• Excellent coefficients of determination were obtained for quantifying fat content (0.897), energy levels (0.906), and milk powder adulteration (0.963).
• Performance was comparable to hyperspectral imaging (HSI), despite SVI's significantly lower equipment cost.

Conclusions:

• The combination of smartphones and machine learning provides a cost-effective and practical preliminary screening tool for whey protein supplement authenticity.
• SVI offers a viable alternative for rapid quality assessment in the food supply chain.
• This technology can empower consumers and manufacturers with accessible product verification.