Jihoon Song
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Jihoon Song has a total of 7 co-authors. Here are the top co-authors by number of shared publications:
- Soo Chung
Department of Biosystems Engineering, Seoul National University, Seoul, Republic of Korea
2 shared publications - Chanwoo Park
Department of Biosystems Engineering, Seoul National University, Seoul 08826, Republic of Korea
1 shared publication - Daesik Son
Department of Biosystems Engineering, Seoul National University, Seoul 08826, Republic of Korea
1 shared publication - Jung-Sun Gloria Kim
Department of Biosystems Engineering Seoul National University Seoul Republic of Korea
1 shared publication - Junseung Bae
Department of Biosystems Engineering, Seoul National University, Seoul 08826, Republic of Korea
1 shared publication - Myungjin Ko
Department of Biosystems Engineering, Seoul National University, Seoul 08826, Republic of Korea
1 shared publication
Latest Publications
- Open Access
Capillary Flow Profile Analysis on Paper-Based Microfluidic Chips for Classifying Astringency Intensity
Astringency, a complex oral sensation resulting from interactions between mucin and polyphenols, remains difficult to quantify in portable field settings. Therefore, quantifying the aggregation through interactions can enable the classification of the astringency intensity, and assessing the capillary action driven by the surface tension offers an effective approach for this purpose. This study successfully replicates tannic acid (TA)–mucin aggregation on a paper-based microfluidic chip and utilizes machine learning (ML) to analyze the resulting capillary flow dynamics. Aggregates formed by mixing mucin with TA solutions at three concentrations showed that higher TA levels led to greater aggregation, consequently reducing the capillary flow rates. The flow dynamics were consistently recorded using a smartphone mounted within a custom 3D-printed frame equipped with a motorized sample loading system, ensuring standardized experimental conditions. Among eight trained ML models, the support vector machine (SVM) demonstrated the highest classification accuracy at 95.2% in distinguishing the astringency intensity levels. Furthermore, fitting the flow data to a theoretical capillary flow equation allowed for the extraction of a single coefficient as an input feature, which achieved comparable classification performance, validating the simplified feature extraction strategy. This method was also feasible even with only a portion of the initial data. This approach is simple and cost-effective and can potentially be developed into a portable system, making it useful for field analysis of various liquid samples.
Open AccessCapillary Flow Profile Analysis on Paper-Based Microfluidic Chips for Classifying Astringency Intensity
Astringency, a complex oral sensation resulting from interactions between mucin and polyphenols, remains difficult to quantify in portable field settings. Therefore, quantifying the aggregation through interactions can enable the classification of the astringency intensity, and assessing the capillary action driven by the surface tension offers an effective approach for this purpose. This study successfully replicates tannic acid (TA)–mucin aggregation on a paper-based microfluidic chip and utilizes machine learning (ML) to analyze the resulting capillary flow dynamics. Aggregates formed by mixing mucin with TA solutions at three concentrations showed that higher TA levels led to greater aggregation, consequently reducing the capillary flow rates. The flow dynamics were consistently recorded using a smartphone mounted within a custom 3D-printed frame equipped with a motorized sample loading system, ensuring standardized experimental conditions. Among eight trained ML models, the support vector machine (SVM) demonstrated the highest classification accuracy at 95.2% in distinguishing the astringency intensity levels. Furthermore, fitting the flow data to a theoretical capillary flow equation allowed for the extraction of a single coefficient as an input feature, which achieved comparable classification performance, validating the simplified feature extraction strategy. This method was also feasible even with only a portion of the initial data. This approach is simple and cost-effective and can potentially be developed into a portable system, making it useful for field analysis of various liquid samples.
- Open Access
- 3 June 2025
- Published by MDPI AG in Applied Sciences
- Vol. 15 (11) , 6278
Comparison of Image Preprocessing Strategies for Convolutional Neural Network-Based Growth Stage Classification of Butterhead Lettuce in Industrial Plant Factories
The increasing need for scalable and efficient crop monitoring systems in industrial plant factories calls for image-based deep learning models that are both accurate and robust to domain variability. This study investigates the feasibility of CNN-based growth stage classification of butterhead lettuce (Lactuca sativa L.) using two data types: raw images and images processed through GrabCut–Watershed segmentation. A ResNet50-based transfer learning model was trained and evaluated on each dataset, and cross-domain performance was assessed to understand generalization capability. Models trained and tested within the same domain achieved high accuracy (Model 1: 99.65%; Model 2: 97.75%). However, cross-domain evaluations revealed asymmetric performance degradation—Model 1-CDE (trained on raw images, tested on preprocessed images) achieved 82.77% accuracy, while Model 2-CDE (trained on preprocessed images, tested on raw images) dropped to 34.15%. Although GrabCut–Watershed offered clearer visual inputs, it limited the model’s ability to generalize due to reduced contextual richness and oversimplification. In terms of inference efficiency, Model 2 recorded the fastest model-only inference time (0.037 s/image), but this excluded the segmentation step. In contrast, Model 1 achieved 0.055 s/image without any additional preprocessing, making it more viable for real-time deployment. Notably, Model 1-CDE combined the fastest inference speed (0.040 s/image) with stable cross-domain performance, while Model 2-CDE was both the slowest (0.053 s/image) and least accurate. Grad-CAM visualizations further confirmed that raw image-trained models consistently attended to meaningful plant structures, whereas segmentation-trained models often failed to localize correctly in cross-domain tests. These findings demonstrate that training with raw images yields more robust, generalizable, and deployable models. The study highlights the importance of domain consistency and preprocessing trade-offs in vision-based agricultural systems and lays the groundwork for lightweight, real-time AI applications in smart farming.
- 3 June 2025
- Published by MDPI AG in Applied Sciences
- Vol. 15 (11) , 6278
Open AccessComparison of Image Preprocessing Strategies for Convolutional Neural Network-Based Growth Stage Classification of Butterhead Lettuce in Industrial Plant Factories
The increasing need for scalable and efficient crop monitoring systems in industrial plant factories calls for image-based deep learning models that are both accurate and robust to domain variability. This study investigates the feasibility of CNN-based growth stage classification of butterhead lettuce (Lactuca sativa L.) using two data types: raw images and images processed through GrabCut–Watershed segmentation. A ResNet50-based transfer learning model was trained and evaluated on each dataset, and cross-domain performance was assessed to understand generalization capability. Models trained and tested within the same domain achieved high accuracy (Model 1: 99.65%; Model 2: 97.75%). However, cross-domain evaluations revealed asymmetric performance degradation—Model 1-CDE (trained on raw images, tested on preprocessed images) achieved 82.77% accuracy, while Model 2-CDE (trained on preprocessed images, tested on raw images) dropped to 34.15%. Although GrabCut–Watershed offered clearer visual inputs, it limited the model’s ability to generalize due to reduced contextual richness and oversimplification. In terms of inference efficiency, Model 2 recorded the fastest model-only inference time (0.037 s/image), but this excluded the segmentation step. In contrast, Model 1 achieved 0.055 s/image without any additional preprocessing, making it more viable for real-time deployment. Notably, Model 1-CDE combined the fastest inference speed (0.040 s/image) with stable cross-domain performance, while Model 2-CDE was both the slowest (0.053 s/image) and least accurate. Grad-CAM visualizations further confirmed that raw image-trained models consistently attended to meaningful plant structures, whereas segmentation-trained models often failed to localize correctly in cross-domain tests. These findings demonstrate that training with raw images yields more robust, generalizable, and deployable models. The study highlights the importance of domain consistency and preprocessing trade-offs in vision-based agricultural systems and lays the groundwork for lightweight, real-time AI applications in smart farming.