Classification of tough and tender beef by image texture analysis

Area of Science:

• Food Science
• Image Analysis
• Meat Science

Background:

• Beef tenderness is a critical quality attribute influencing consumer satisfaction.
• Objective methods for predicting beef tenderness are needed to complement subjective sensory evaluations.
• Fresh beef image texture may correlate with cooked beef tenderness.

Purpose of the Study:

• To extract texture features from fresh beef images.
• To classify fresh beef steaks into tough and tender groups based on cooked tenderness.
• To evaluate the effectiveness of image texture analysis for predicting beef tenderness.

Main Methods:

• Collected fresh beef short loin steaks from crossbred steers.
• Used a wavelet-based decomposition method for texture feature extraction from images.
• Employed a rotational leave-one-out cross-validation scheme with 90 sample images.
• Classified steaks into tough and tender groups using sensory evaluation scores.

Main Results:

• Achieved a correct classification rate of 83.3% in cross-validation.
• Texture features demonstrated significant ability to differentiate between tough and tender beef.
• Image texture analysis provided a promising objective measure for beef tenderness prediction.

Conclusions:

• Fresh beef image texture features are valuable indicators for predicting cooked beef tenderness.
• While not sufficient alone for multi-level tenderness grading, texture features can be part of a larger prediction model.
• This approach offers a potential objective tool for the meat industry to assess beef quality.