CBD: Coffee Beans Dataset.

Area of Science:

• Agricultural Science
• Computer Science
• Data Science

Background:

• Coffee bean quality assessment is crucial but challenging due to subjective factors and defects.
• High-quality datasets are essential for developing accurate automated coffee bean classification and grading systems.
• Current methods often lack the comprehensive data needed for robust machine learning models.

Purpose of the Study:

• To introduce the Coffee Beans Dataset (CBD), a novel, high-resolution image dataset for coffee bean classification.
• To provide a standardized resource for training and evaluating machine learning and deep learning models in coffee grading.
• To facilitate advancements in automated coffee quality assessment and classification accuracy.

Main Methods:

• Collected 450 high-resolution images of coffee beans across 9 distinct grades (A, AA, AAA, AB, C, PB-I, PB-II, BITS, BULK).
• Ensured 50 images per class, sourced from Wayanad, Kerala, to capture regional quality variations.
• Utilized the EfficientNet-B0 model for initial testing and validation of the dataset's efficacy.

Main Results:

• The Coffee Beans Dataset (CBD) comprises 450 images across 9 grades, offering diverse coffee bean quality representation.
• Testing the EfficientNet-B0 model on the CBD resulted in a classification accuracy of 100%.
• Demonstrated the dataset's capability to support high-performance automated coffee bean grading.

Conclusions:

• The CBD is a valuable resource for researchers and industry professionals in coffee quality monitoring.
• The dataset facilitates the development of advanced machine learning algorithms for precise coffee bean classification.
• The 100% accuracy achieved highlights the potential for significant improvements in automated coffee grading systems.