The Changes in Bell Pepper Flesh as a Result of Lacto-Fermentation Evaluated Using Image Features and Machine...

Area of Science:

• Food Science
• Computer Science
• Agricultural Science

Background:

• Food processing extends shelf life but requires quality control.
• Nondestructive methods like image analysis and machine learning offer solutions for food quality assessment.
• Distinguishing between fresh and processed foods is crucial for consumers and industry.

Purpose of the Study:

• To develop an innovative method for distinguishing fresh from lacto-fermented red bell peppers.
• To utilize image texture analysis and machine learning algorithms for this classification task.
• To evaluate the effectiveness of different color spaces and algorithms in identifying processing-induced changes.

Main Methods:

• Fresh and lacto-fermented red bell pepper samples were imaged using a digital camera.
• Texture parameters were extracted from images converted into various color channels (Lab, RGB, XYZ).
• Machine learning algorithms (Lazy, Functions, Trees, Bayes, Meta, Rules) were employed to build classification models.

Main Results:

• Classification models achieved up to 99% accuracy.
• The highest accuracy was observed using the Lab color space with the IBk algorithm, RGB with SMO, and XYZ with IBk/SMO.
• Significant differences in image features between fresh and processed peppers were confirmed.

Conclusions:

• Image texture analysis combined with machine learning effectively differentiates fresh and lacto-fermented red bell peppers.
• This approach provides a reliable, nondestructive method for evaluating structural changes in processed foods.
• The developed models demonstrate the potential for automated quality control in food processing.