Comparison of Coffee Bean Roasting Level Classification Using ResNet50 and VGG16
Abstract
The classification of coffee bean roasting levels is an important aspect of ensuring coffee product quality. This study compares the performance of two deep learning architectures, ResNet50 and VGG16, in classifying coffee bean images into three roasting levels: light, medium, and dark. The dataset consists of 1,800 images with a resolution of 224×224 pixels, divided into training, validation, and testing sets. Both models were trained with identical configurations using transfer learning and partial fine-tuning. The evaluation results show a very small accuracy difference of only 0.01 point, with ResNet50 slightly outperforming VGG16. This indicates that both models are equally reliable for roast level classification. However, ResNet50 is more time-efficient, requiring only about 10 minutes of training compared to over 25 minutes for VGG16. This difference is suspected to be related to the complexity of VGG16’s architecture. The study concludes that ResNet50 offers high efficiency with competitive accuracy. Further research is recommended to optimize VGG16’s architecture to improve computational efficiency without compromising accuracy.
Keywords: Machine learning; Restnet50; VGG16; Coffee bean roasting
Abstrak
Klasifikasi tingkat roasting biji kopi merupakan aspek penting dalam penjaminan mutu produk kopi. Penelitian ini membandingkan performa dua arsitektur deep learning, ResNet50 dan VGG16, dalam mengklasifikasikan citra biji kopi pada tiga tingkat roasting: light, medium, dan dark. Dataset berisi 1.800 citra beresolusi 224×224 piksel, dibagi menjadi data latih, validasi, dan uji. Kedua model dilatih dengan konfigurasi identik menggunakan transfer learning dan fine-tuning parsial. Hasil evaluasi menunjukkan selisih akurasi sangat tipis, hanya 0,01 poin, dengan ResNet50 sedikit unggul. Hal ini menunjukkan kedua model sama-sama andal untuk klasifikasi tingkat roasting. Namun, ResNet50 lebih efisien secara waktu, hanya memerlukan sekitar 10 menit pelatihan dibandingkan VGG16 yang lebih dari 25 menit. Perbedaan ini diduga terkait kompleksitas arsitektur VGG16. Disimpulkan bahwa ResNet50 menawarkan efisiensi tinggi dengan akurasi kompetitif. Penelitian lanjutan disarankan mengevaluasi optimasi arsitektur VGG16 untuk meningkatkan efisiensi komputasi tanpa mengorbankan akurasi.
Keywords
References
L. Poisson, I. Blank, A. Dunkel, and T. Hofmann, “The Chemistry of Roasting-Decoding Flavor Formation,” in The Craft and Science of Coffee, Elsevier Inc., 2017, pp. 273–309. doi: 10.1016/B978-0-12-803520-7.00012-8.
K. F. Ting, J. C. Chen, and T. L. Chen, “Effect of coffee roasting on the cupping quality of coffee,” Coffee Sci., vol. 19, 2024, doi: 10.25186/.v19i.2174.
“The Ultimate Guide to Understanding Coffee Roast Levels,” coffeebros.com. Accessed: Jul. 17, 2025. [Online]. Available: https://coffeebros.com/blogs/coffee/the-ultimate-guide-to-understanding-coffee-roast-levels?srsltid=AfmBOoqzWtMDSq7ov_uDv-klLX1iqz35Jb_3CqoGbBMmuoWz97PvPAZH
C. da C. Garcia et al., “Relative importance and interaction of roasting variables in coffee roasting process,” Coffee Sci., vol. 13, no. 3, pp. 379–388, 2018, doi: 10.25186/cs.v13i3.1483.
W. B. Sunarharum, M. Nurminah, and N. G. Purba, “Effect of Different Roasting Levels and Manual Brewing Techniques on the Sensory Profile of Liberica Coffee with Honey Process,” J. Coffee Sustain., vol. 01, no. 01, pp. 40–48, 2024, doi: 10.21776/ub.jcs.2024.01.01.05.
N. Klinger, “ANN and CNN: Analyzing Differences and Similarities,” viso.ai. Accessed: Jul. 17, 2025. [Online]. Available: https://viso.ai/deep-learning/ann-and-cnn-analyzing-differences-and-similarities/
S. D. Astuti et al., “Electronic nose coupled with artificial neural network for classifying of coffee roasting profile,” Sens. Bio-Sensing Res., vol. 43, p. 100632, 2024, doi: 10.1016/j.sbsr.2024.100632.
A. F. Ihsani and M. Ichwan, “Kombinasi Arsitektur VGG16 dan DenseNet121 Untuk Klasifikasi Kualitas Biji Kopi Berdasarkan Tingkat Kematangan,” vol. 8, no. 1, pp. 1–12, 2023.
M. N. Tama, A. B. Saptomo, D. K. Baroroh, A. P. Rifai, N. H. Tho, and A. Info, “Comparative Evaluation of Convolutional Neural Network Full Learning Model with Transfer Learning ( VGG-16 ) for Coffee Bean Roasting Level Classification,” vol. 6, no. 2, pp. 332–343, 2025, doi: 10.59395/ijadis.v6i2.1398.
I. M. Pakaya, R. Radi, and B. Purwantana, “Classification of Roasting Level of Coffee Beans Using Convolutional Neural Network with MobileNet Architecture for Android Implementation,” J. Tek. Pertan. Lampung (Journal Agric. Eng., vol. 13, no. 3, pp. 924-933, 2024, doi: 10.23960/jtep-l.v13i3.924-932.
T. Firmansyah et al., “Klasfikasi Tingkat Kematangan Roasting Biji Kopi Berbasis Deep Learning dengan Arsitektur MobileNet,” J. Inf. Syst. Res., vol. 6, no. 2, pp. 1433–1443, 2025, doi: 10.47065/josh.v6i2.6811.
S. Ontoum, T. Khemanantakul, P. Sroison, T. Triyason, and B. Watanapa, “Coffee Roast Intelligence,” 2022, [Online]. Available: http://arxiv.org/abs/2206.01841
M. A. Alrasyid, B. Rohmatullah, R. Damayanti, and D. F. Al-Riza, “ResNet-50 to classify the types of Indonesian local coffee beans,” Proc. 1ST UNHAS Int. Conf. Agric. Technol. (UICAT 2021), vol. 2596, no. 1, p. 050008, 2023.
O. Osakwe, “The Significance of Discovery Screening and Structure Optimization Studies,” Soc. Asp. Drug Discov. Dev. Commer., pp. 109-128, 2016.
S. T. Musharraf, “Everything you need to know about VGG16- A Deep Learning Architecture,” linkedin.com. Accessed: Jul. 17, 2025. [Online]. Available: https://www.linkedin.com/pulse/everything-you-need-know-vgg16-deep-learning-syed-talal-musharraf-pjkkf/
T. Sharma, “Detailed Explanation of Resnet CNN Model,” medium.com. Accessed: Jul. 17, 2025. [Online]. Available: https://medium.com/@sharma.tanish096/detailed-explanation-of-residual-network-resnet50-cnn-model-106e0ab9fa9e
How To Cite This :
Refbacks
- There are currently no refbacks.
