Validation of the multivariate models for correlation between texture profile analysis, digital images, and NIR...

Area of Science:

• Food Science and Technology
• Analytical Chemistry
• Spectroscopy

Background:

• Texture Profile Analysis (TPA) is crucial for food quality assessment.
• Traditional TPA methods can be time-consuming and destructive.
• Developing rapid, non-destructive methods for TPA is highly desirable.

Purpose of the Study:

• To evaluate near-infrared (NIR) spectroscopy and digital imaging techniques for predicting TPA parameters.
• To develop a multivariate multiproduct calibration model for diverse flour types.
• To validate the predictive models using established parameters of merit.

Main Methods:

• Utilized NIR spectroscopy, conventional digital images, and thermal/X-ray filtered images.
• Employed partial least squares (PLS) regression for multivariate analysis.
• Incorporated multiple flour types into single calibration models for broader applicability.

Main Results:

• Successfully determined key TPA parameters including hardness, adhesiveness, springiness, chewiness, gumminess, cohesiveness, and resilience.
• Variable Importance in Projection (VIP-scores) revealed shared predictive variables for hardness, gumminess, and cohesion.
• Identified common variables for adhesiveness and elasticity, as well as for chewability and resilience.

Conclusions:

• NIR spectroscopy and digital imaging, coupled with PLS, provide a robust, non-destructive method for TPA determination.
• The developed multivariate multiproduct model demonstrates versatility across different flour types.
• This approach offers significant potential for efficient food quality control and product development.