How Deep Learning is Improving Coffee Harvesting: A Review of CNN-VGG19 for Automated Coffee Cherry Maturity Classification

on

Coffee cherry maturity plays a fundamental role in determining the physical quality, chemical composition, and sensory characteristics of coffee beans. Harvesting cherries at the appropriate ripeness stage contributes directly to flavor development, processing efficiency, and market value, whereas immature or overripe cherries may reduce beverage quality and economic returns. Despite its importance, maturity assessment in many coffee-producing regions continues to rely on manual visual inspection, a practice that is labor-intensive, time-consuming, and susceptible to human inconsistency. As coffee production expands and quality standards become increasingly stringent, there is growing demand for automated inspection technologies capable of providing faster, more objective, and highly accurate maturity classification.

Recent advances in computer vision and Convolutional Neural Networks (CNN) have significantly improved image-based agricultural inspection by enabling machines to recognize subtle visual differences among biological objects. Building upon these developments, the reviewed study proposes a CNN model based on the Visual Geometry Group-19 (VGG19) architecture to classify coffee cherries into four maturity categories: unripe, semi-ripe, ripe, and overripe. By introducing a more detailed maturity classification than many previous approaches, the research offers valuable insights for agricultural engineers, artificial intelligence researchers, computer vision specialists, precision agriculture practitioners, coffee producers, and technology developers interested in intelligent quality inspection systems.

Bibliographic Information

Item Information
Article Title Classifying four maturity categories of coffee cherry using CNN-VGG19
Authors Dominic Olango Cagadas, Dwi Sudarno Putra, Kristine Mae Paboreal Dunque, and Meri Azmi
Journal Teknomekanik
Volume & Issue Volume 7, Issue 2
Publication Year 2024
Pages 176–184
DOI https://doi.org/10.24036/teknomekanik.v7i2.31072
Publisher Universitas Negeri Padang
License Creative Commons Attribution 4.0 International (CC BY 4.0)

1. Research Background

  • Coffee quality depends heavily on harvesting cherries at the proper maturity stage. Coffee cherries undergo progressive physiological changes during ripening that directly influence bean composition, processing characteristics, and beverage quality. Harvesting fruit too early or too late can reduce sweetness, alter flavor balance, lower processing yield, and ultimately decrease the commercial value of coffee. Consequently, accurate maturity identification has become an essential component of quality control throughout the coffee production chain.
  • Traditional harvesting practices still rely primarily on manual inspection. In many coffee-producing regions, workers determine cherry maturity through visual observation of fruit color before harvesting. While experienced farmers can achieve acceptable results, manual inspection is inherently labor-intensive, time-consuming, and susceptible to inconsistency, particularly during large-scale harvesting operations where thousands of cherries must be evaluated within a limited time.
  • Existing harvesting methods involve important operational trade-offs. Strip picking enables rapid harvesting by removing cherries simultaneously regardless of maturity, but often produces mixed batches containing unripe, semi-ripe, ripe, and overripe fruit. Selective picking delivers better quality because only mature cherries are harvested; however, it requires repeated harvesting rounds, greater labor input, and significantly higher production costs.
  • Previous agricultural studies have demonstrated the importance of coffee cherry maturity. Earlier research has shown that immature cherries tend to produce undesirable sensory characteristics, whereas properly ripened cherries generate superior aroma, sweetness, and cup quality. These findings emphasize the need for reliable maturity classification methods capable of supporting consistent harvesting decisions and improving overall coffee quality.
  • Deep learning has become an increasingly effective solution for agricultural image analysis. Recent advances in deep learning have transformed computer vision applications by enabling automatic extraction of complex visual features from digital images. Among these approaches, Convolutional Neural Networks have consistently demonstrated excellent performance in object detection, image recognition, and agricultural classification tasks without relying on manually engineered image features.
  • Previous coffee classification models still present opportunities for improvement. Earlier computer vision studies successfully classified coffee cherries into broad maturity groups, but many focused on only three categories and reported accuracy levels that could be further improved. Such limitations reduce their ability to represent the gradual ripening process observed under practical harvesting conditions and motivate continued development of more accurate classification systems.
  • The literature identifies a clear research gap. Existing image-based maturity classification systems rarely combine a transfer-learning CNN architecture with four maturity categories that distinguish unripe, semi-ripe, ripe, and overripe cherries simultaneously. This limitation indicates the need for a more comprehensive classification framework capable of improving recognition performance while representing actual coffee maturation stages more realistically.
  • This study proposes a CNN-VGG19 based maturity classification framework. To address the identified gap, the researchers developed a CNN-VGG19 model utilizing transfer learning to classify coffee cherries into four maturity categories. The proposed framework aims to improve recognition accuracy while supporting automated inspection that may reduce labor requirements, increase harvesting consistency, and assist local coffee farmers through intelligent image-based quality assessment.

2. Research Objectives

  • To develop an intelligent image classification model based on Convolutional Neural Networks (CNN) and the VGG19 architecture for automatically identifying the maturity level of coffee cherries from digital images.
  • To classify coffee cherries into four distinct maturity categories—unripe, semi-ripe, ripe, and overripe—thereby providing a more comprehensive maturity assessment than many previously reported image classification systems.
  • To improve the recognition accuracy of coffee cherry maturity classification by employing transfer learning and deep feature extraction through the CNN-VGG19 model.
  • To evaluate the classification performance of the proposed model using standard machine learning performance indicators, including accuracy, precision, recall, F1-score, specificity, and confusion matrix analysis.
  • To investigate whether image processing and deep learning techniques can provide a practical alternative to conventional manual inspection methods used by coffee farmers during harvesting.
  • To contribute an automated quality inspection approach capable of supporting faster, more objective, and more consistent coffee cherry maturity assessment in agricultural production.

3. Why This Research Matters

  • Supports precision agriculture. Automated maturity classification enables more objective harvesting decisions by reducing dependence on subjective visual inspection. Such technology represents an important step toward precision agriculture, where digital technologies improve production efficiency through data-driven decision making.
  • Improves coffee quality consistency. More accurate identification of coffee cherry maturity helps ensure that harvested cherries possess similar ripeness characteristics. Greater harvesting consistency can improve post-harvest processing, reduce quality variation, and contribute to better coffee flavor and higher commercial value.
  • Reduces labor-intensive inspection. Manual maturity evaluation requires experienced workers and repeated visual inspection throughout the harvesting season. Automated image classification offers opportunities to decrease inspection time, reduce labor requirements, and improve operational efficiency, particularly for large-scale coffee production.
  • Demonstrates the value of artificial intelligence in agriculture. The proposed CNN-VGG19 framework illustrates how artificial intelligence and deep learning can address practical agricultural challenges beyond traditional engineering applications. The study contributes to the growing adoption of intelligent decision-support systems within modern farming.
  • Advances computer vision applications. By extending maturity classification to four categories, the research demonstrates the capability of transfer-learning CNN models to recognize subtle visual differences among biological objects. This contributes to the broader development of computer vision technologies for agricultural inspection and automated quality assessment.
  • Encourages digital transformation in agricultural engineering. Intelligent image classification systems have the potential to become components of future smart farming platforms that integrate cameras, mobile devices, cloud computing, and automated decision support. Such developments align with ongoing efforts to digitalize agricultural production through Industry 4.0 technologies.
  • Provides broader engineering relevance. Although developed specifically for coffee cherries, the proposed deep learning framework demonstrates principles that can be adapted to other agricultural products requiring automated maturity detection, grading, defect inspection, or quality classification using computer vision techniques.

4. Research Methodology

  • Research Type

    This study employed an experimental quantitative research approach in the field of deep learning and computer vision. The research focused on developing an image classification model capable of automatically recognizing the maturity level of coffee cherries using digital image processing and a pretrained Convolutional Neural Network architecture. Rather than relying on conventional image analysis techniques, the study utilized transfer learning to improve feature extraction and classification performance.

  • Image Dataset

    The experimental dataset consisted of 356 coffee cherry images collected from actual coffee farms and complementary image sources. Before model training, experienced coffee farmers manually categorized each image into one of four maturity classes according to the external color of the fruit: unripe, semi-ripe, ripe, and overripe. This expert labeling process established the ground-truth dataset used throughout model development and evaluation.

  • Image Pre-processing

    All collected images were resized and standardized to satisfy the input requirements of the CNN-VGG19 architecture. The preprocessing stage ensured consistent image dimensions before feature extraction while reducing unnecessary computational complexity. Image preparation also improved model robustness by allowing the network to learn discriminative maturity characteristics from standardized visual inputs.

  • Deep Learning Model

    The proposed classification framework employed the Visual Geometry Group-19 (VGG19) convolutional neural network through a transfer learning strategy. The pretrained model contains multiple convolutional layers for hierarchical feature extraction, max-pooling layers for dimensionality reduction, and fully connected layers for final classification. To adapt the architecture to the present study, the original output layer was modified so that the model could classify four coffee cherry maturity categories instead of the original ImageNet classes.

  • Transfer Learning Strategy

    Rather than training the entire network from the beginning, the researchers froze the initial VGG19 feature extraction layers and fine-tuned the classification layers using the coffee cherry dataset. This transfer learning strategy enabled the model to benefit from previously learned visual representations while reducing training time and improving classification performance despite the relatively limited dataset size.

  • Classification Framework

    The classification workflow consisted of several sequential stages, including image acquisition, preprocessing, CNN-VGG19 feature extraction, model training, prediction, and maturity classification. Following feature extraction, the network generated probability scores corresponding to four maturity classes, allowing each input image to be assigned automatically to the most probable maturity category.

  • Performance Evaluation

    Model performance was evaluated using several widely accepted classification metrics, including overall accuracy, precision, recall, F1-score, specificity, and confusion matrix analysis. These complementary indicators provided a comprehensive assessment of the model's ability to distinguish among the four maturity categories while identifying both correctly and incorrectly classified samples.

  • Validation Procedure

    The effectiveness of the proposed CNN-VGG19 model was assessed by comparing prediction results between training and validation datasets. Training and validation accuracy curves, together with corresponding loss curves, were analyzed to evaluate learning behavior and model convergence. In addition, confusion matrix analysis was performed to examine classification performance for each maturity category individually and to identify potential misclassification patterns.

5. Key Findings

CNN-VGG19 Successfully Classified Four Coffee Cherry Maturity Categories

The primary outcome of this study is the successful implementation of a CNN-VGG19 model capable of distinguishing coffee cherries into four maturity categories: unripe, semi-ripe, ripe, and overripe. Unlike several previous computer vision studies that focused on only three maturity levels, the proposed framework provides a more detailed representation of the natural ripening process, making it better suited for practical harvesting conditions where intermediate maturity stages are commonly encountered.

The experimental results demonstrate that transfer learning using the VGG19 architecture can effectively extract visual features associated with subtle differences in coffee cherry color and appearance. This capability enables the proposed model to perform reliable multi-class classification while supporting automated quality inspection in agricultural applications.

High Classification Accuracy Demonstrates Strong Recognition Capability

The CNN-VGG19 model achieved an overall classification accuracy of approximately 90%, indicating that most coffee cherry images were correctly assigned to their respective maturity categories. The classification report also showed consistently high precision, recall, and F1-score values across the evaluated classes, suggesting balanced performance rather than high accuracy being dominated by only one maturity category.

These findings indicate that the proposed deep learning model possesses strong discriminative capability for recognizing visual maturity characteristics. Such performance suggests that automated image classification can provide reliable support for maturity assessment while reducing the dependence on subjective human observation during harvesting.

Training and Validation Curves Indicate Stable Learning Performance

Analysis of the training process revealed a steady increase in both training and validation accuracy throughout model learning, accompanied by a corresponding decrease in training and validation loss. This pattern indicates that the network progressively learned meaningful image features while improving prediction capability over successive training iterations.

The consistency between the accuracy and loss curves suggests that the transfer learning strategy enabled effective model optimization without substantial learning instability. These results provide additional evidence that the CNN-VGG19 architecture was well suited for the coffee cherry maturity classification task investigated in this study.

Confusion Matrix Reveals Reliable Classification Across All Maturity Classes

The confusion matrix demonstrated that the majority of coffee cherry images were correctly classified into their corresponding maturity categories. Most prediction errors occurred between neighboring maturity stages, particularly between unripe and semi-ripe cherries or between ripe and overripe cherries. Such misclassifications are understandable because adjacent maturity levels often exhibit highly similar visual characteristics.

Despite these relatively minor classification errors, the confusion matrix confirms that the proposed CNN-VGG19 model maintains reliable prediction capability across all four maturity classes. The balanced classification performance illustrates the effectiveness of deep feature extraction in distinguishing subtle differences among biologically similar objects.

Transfer Learning Improved Deep Learning Efficiency

Rather than developing a convolutional neural network entirely from scratch, the researchers employed transfer learning by adapting the pretrained VGG19 architecture to the coffee cherry classification problem. This strategy enabled the model to utilize previously learned visual representations while focusing subsequent training on maturity-specific image characteristics.

The successful implementation of transfer learning demonstrates that pretrained convolutional neural networks can achieve high classification performance even when relatively limited agricultural datasets are available. This finding is particularly important for agricultural engineering applications where collecting very large labeled image datasets may be difficult and time-consuming.

The Proposed Model Shows Strong Potential for Automated Coffee Quality Inspection

Beyond achieving high predictive performance, the study demonstrates that deep learning can serve as an effective decision-support tool for coffee harvesting. Automated maturity recognition offers opportunities to reduce manual inspection effort while improving consistency in harvesting decisions and minimizing variability associated with human judgment.

The overall findings suggest that CNN-VGG19-based image classification could become an important component of future intelligent agricultural systems. By integrating computer vision with precision farming technologies, similar approaches may contribute to more efficient harvesting practices, improved coffee quality management, and broader adoption of artificial intelligence within modern agricultural production.

6. Scientific Contribution

  • Introduces a four-category coffee cherry maturity classification framework. Unlike many previous studies that classified coffee cherries into only three maturity levels, this research develops a CNN-VGG19 model capable of distinguishing four maturity categories: unripe, semi-ripe, ripe, and overripe. This finer classification provides a more realistic representation of coffee ripening during harvesting operations.
  • Demonstrates the effectiveness of transfer learning for agricultural image classification. The study confirms that integrating a pretrained VGG19 architecture with transfer learning can successfully extract discriminative visual features from coffee cherry images, enabling accurate classification without requiring an extremely large training dataset.
  • Advances computer vision applications in precision agriculture. By combining image processing, deep learning, and automated maturity recognition, the research expands the application of computer vision technologies for agricultural quality assessment. The proposed framework illustrates how artificial intelligence can support objective and efficient crop inspection.
  • Provides comprehensive model evaluation. Rather than relying solely on overall accuracy, the study evaluates classification performance using precision, recall, F1-score, specificity, and confusion matrix analysis. This comprehensive assessment strengthens confidence in the reliability and robustness of the proposed CNN-VGG19 model.
  • Bridges agricultural expertise and artificial intelligence. The research integrates expert knowledge from coffee farmers with modern deep learning techniques by using manually labelled maturity classes as ground-truth data for supervised learning. This interdisciplinary approach demonstrates how domain expertise and machine learning can be effectively combined to address practical agricultural problems.
  • Provides a foundation for future intelligent harvesting systems. The proposed maturity classification framework establishes a technical basis for developing automated harvesting assistance systems, smart quality inspection devices, and intelligent agricultural monitoring platforms capable of supporting precision farming applications.

7. Industrial Implications

  • Improves harvesting efficiency. Automated maturity classification can reduce the time required for manual inspection, allowing coffee producers to evaluate large quantities of cherries more rapidly while maintaining consistent quality standards throughout harvesting operations.
  • Enhances coffee quality management. More accurate maturity identification enables producers to separate cherries according to ripeness before post-harvest processing, contributing to improved product consistency, better bean quality, and potentially higher commercial value.
  • Reduces dependence on manual labor. Coffee harvesting often requires experienced workers capable of distinguishing subtle maturity differences. Intelligent image classification systems may assist farmers by providing objective decision support, thereby reducing labor intensity and minimizing inconsistencies caused by human judgment.
  • Supports precision agriculture technologies. The proposed CNN-VGG19 framework can be integrated with cameras, smartphones, drones, or automated harvesting equipment to provide real-time maturity assessment. Such integration supports the continued digital transformation of agricultural production through data-driven quality inspection.
  • Encourages the adoption of artificial intelligence in agriculture. The successful implementation of deep learning for coffee maturity recognition demonstrates the practical value of artificial intelligence in solving real agricultural problems. Similar approaches may be extended to fruit grading, crop monitoring, disease detection, and automated quality control for other agricultural commodities.
  • Supports digital engineering and Industry 4.0 initiatives. Intelligent computer vision systems represent an important component of modern digital agriculture, where machine learning, automation, and sensor technologies work together to improve productivity, quality assurance, and sustainable food production.
  • Offers opportunities for commercialization. The proposed image classification model could serve as the foundation for portable inspection devices or mobile applications that assist coffee farmers in determining harvest readiness. Such technologies have the potential to improve operational decision-making while increasing accessibility to intelligent agricultural tools.

8. Research Limitations

  • The study utilized a dataset consisting of 356 labelled coffee cherry images. Although sufficient to demonstrate the feasibility of the proposed approach, larger and more diverse datasets could further improve model generalization under varying agricultural conditions.
  • Image classification was performed using photographs collected under controlled experimental conditions. Variations in illumination, background complexity, camera angle, and environmental conditions encountered during field deployment may influence classification performance and should be investigated further.
  • The proposed model focused exclusively on visual characteristics represented by external fruit appearance. Other maturity indicators, such as internal fruit properties, chemical composition, or multispectral information, were beyond the scope of the present study.
  • Although the CNN-VGG19 model achieved high classification accuracy, the study evaluated only one deep learning architecture. Comparative investigations involving more recent convolutional neural network models may provide additional insight into potential performance improvements.
  • The research emphasized image classification accuracy rather than deployment in operational harvesting environments. Future implementation within real agricultural production systems will require consideration of computational efficiency, hardware requirements, and real-time processing capability.
  • The study concentrated on coffee cherries harvested in a specific production context. Additional validation using different coffee varieties, cultivation regions, and environmental conditions would strengthen the general applicability of the proposed classification framework.

9. Future Research Opportunities

  • Investigate the performance of the proposed CNN-VGG19 framework using substantially larger and more diverse image datasets collected from multiple coffee-producing regions, varieties, cultivation systems, and harvesting seasons to improve model robustness and generalization capability.
  • Compare the proposed VGG19 architecture with more recent deep learning models, such as EfficientNet, ResNet, DenseNet, MobileNet, and Vision Transformers to determine the most suitable architecture for coffee maturity classification under practical agricultural conditions.
  • Develop lightweight deep learning models that can be deployed on smartphones, embedded devices, or edge-computing platforms, enabling real-time coffee cherry maturity assessment directly in the field without requiring high-performance computing resources.
  • Evaluate the robustness of automated maturity classification under varying environmental conditions, including changes in illumination, weather, shadows, fruit occlusion, camera viewpoints, and complex natural backgrounds commonly encountered during harvesting operations.
  • Integrate computer vision with Internet of Things (IoT) technologies, cloud computing, and wireless sensor networks to develop intelligent decision-support systems capable of monitoring coffee plantations and providing continuous harvest recommendations.
  • Investigate multimodal classification approaches by combining RGB images with hyperspectral imaging, multispectral sensing, thermal imaging, or chemical analysis to improve maturity prediction accuracy beyond visual appearance alone.
  • Extend the proposed methodology to support additional agricultural quality inspection tasks, including defect detection, disease identification, fruit grading, yield estimation, and automated post-harvest sorting using integrated deep learning frameworks.
  • Develop autonomous harvesting systems that integrate CNN-based maturity recognition with robotic manipulators, machine vision, and intelligent navigation technologies to enable selective harvesting of optimally ripened coffee cherries.
  • Conduct comparative economic analyses to evaluate the cost-effectiveness, productivity improvement, labor reduction, and return on investment associated with deploying AI-based coffee maturity classification systems in commercial coffee plantations.
  • Perform long-term validation studies involving coffee farmers, cooperatives, and agricultural industries to assess user acceptance, operational reliability, and scalability of intelligent coffee maturity inspection systems under real production environments.

10. Potential for Public Policy Citation (Overton)

This study demonstrates meaningful potential for citation in public policy documents because it addresses one of the fundamental challenges in agricultural production: improving crop quality through intelligent digital technologies. Automated coffee cherry maturity classification aligns closely with current government initiatives promoting digital agriculture, artificial intelligence adoption, and precision farming. The proposed CNN-VGG19 framework provides evidence that computer vision can support more objective harvesting decisions while reducing labor dependency and improving product consistency.

The research could contribute to national agricultural digitalization strategies, smart farming roadmaps, agricultural modernization programs, food quality improvement initiatives, and innovation policies encouraging the adoption of artificial intelligence in the agricultural sector. The study may also provide useful technical references for government agencies, agricultural research institutions, coffee development authorities, and extension services seeking evidence-based approaches to enhance crop quality management.

Nevertheless, the article is less likely to be cited directly within regulatory standards because its primary contribution focuses on the development and evaluation of a deep learning model rather than regulatory frameworks or policy implementation. Wider policy influence would likely require additional studies demonstrating successful deployment under commercial farming conditions, economic feasibility, and measurable impacts on agricultural productivity and sustainability.

11. Who Should Read This Paper?

  • Researchers in artificial intelligence and deep learning.
  • Computer vision researchers.
  • Agricultural engineers.
  • Precision agriculture researchers.
  • Coffee scientists and coffee quality researchers.
  • Machine learning engineers.
  • Graduate students studying artificial intelligence, computer vision, or agricultural engineering.
  • Coffee producers and plantation managers.
  • Agricultural technology developers.
  • Engineers developing intelligent inspection systems.
  • Government agencies responsible for agricultural innovation and digital transformation.
  • Start-ups developing AI-based agricultural solutions.

12. Final Thoughts

This study provides a valuable contribution to the growing application of artificial intelligence within agricultural engineering by demonstrating how deep learning can improve the efficiency and consistency of coffee cherry maturity assessment. Rather than relying solely on conventional visual inspection, the proposed CNN-VGG19 framework offers an automated image classification approach capable of distinguishing four maturity stages with high classification performance. The integration of transfer learning further illustrates how existing deep learning architectures can be effectively adapted to specialized agricultural problems without requiring extremely large datasets.

One of the principal strengths of the research lies in its practical orientation. The proposed system addresses a real challenge faced by coffee producers by reducing the dependence on subjective manual inspection while supporting more consistent harvesting decisions. The comprehensive evaluation using multiple performance metrics also provides confidence in the reliability of the developed classification model and demonstrates the suitability of CNN-VGG19 for image-based maturity recognition.

From an engineering perspective, the significance of this work extends beyond coffee production. Similar computer vision frameworks can be adapted for automated fruit grading, crop monitoring, quality inspection, and intelligent agricultural robotics. As digital agriculture continues to evolve through artificial intelligence, edge computing, and smart sensing technologies, research such as this provides an important foundation for developing practical decision-support systems capable of improving agricultural productivity and sustainability.

Overall, this article represents a meaningful contribution to precision agriculture and computer vision research. By successfully combining transfer learning, image processing, and deep learning for agricultural quality inspection, the study demonstrates how engineering innovation can address practical challenges in food production while creating opportunities for future intelligent farming technologies.

Suggested Citation

Teknomekanik (UNP) Style

Cagadas, D. O., Putra, D. S., Dunque, K. M. P., & Azmi, M. (2024). Classifying four maturity categories of coffee cherry using CNN-VGG19. Teknomekanik, 7(2), 176–184. https://doi.org/10.24036/teknomekanik.v7i2.31072

APA (7th Edition)

Cagadas, D. O., Putra, D. S., Dunque, K. M. P., & Azmi, M. (2024). Classifying four maturity categories of coffee cherry using CNN-VGG19. Teknomekanik, 7(2), 176–184. https://doi.org/10.24036/teknomekanik.v7i2.31072

IEEE Style

D. O. Cagadas, D. S. Putra, K. M. P. Dunque, and M. Azmi, "Classifying four maturity categories of coffee cherry using CNN-VGG19," Teknomekanik, vol. 7, no. 2, pp. 176–184, 2024, doi: 10.24036/teknomekanik.v7i2.31072.

Harvard Style

Cagadas, D.O., Putra, D.S., Dunque, K.M.P. & Azmi, M., 2024. Classifying four maturity categories of coffee cherry using CNN-VGG19. Teknomekanik, 7(2), pp.176–184. Available at: https://doi.org/10.24036/teknomekanik.v7i2.31072.

Vancouver Style

Cagadas DO, Putra DS, Dunque KMP, Azmi M. Classifying four maturity categories of coffee cherry using CNN-VGG19. Teknomekanik. 2024;7(2):176-184. doi: 10.24036/teknomekanik.v7i2.31072.

Chicago (Author–Date)

Cagadas, Dominic Olango, Dwi Sudarno Putra, Kristine Mae Paboreal Dunque, and Meri Azmi. 2024. "Classifying Four Maturity Categories of Coffee Cherry Using CNN-VGG19." Teknomekanik 7 (2): 176–184. https://doi.org/10.24036/teknomekanik.v7i2.31072.

MLA (9th Edition)

Cagadas, Dominic Olango, et al. "Classifying Four Maturity Categories of Coffee Cherry Using CNN-VGG19." Teknomekanik, vol. 7, no. 2, 2024, pp. 176–184. Crossref, https://doi.org/10.24036/teknomekanik.v7i2.31072.

Editorial Note

Editorial Note: This blog post is an independent scholarly review intended for educational and scientific communication purposes. It summarizes and discusses the published article in the author's own words while providing full attribution to the original publication, consistent with the principles of the CC BY 4.0 license.

Comments

Post a Comment