Education 🎓📚
Jeon is a “triple alumnus” of Kyungnam University, Changwon. He earned his B.S. in Computer Engineering (2009‑2016) with early research on adaptive fuzzy binarization, followed by an M.S. (2016‑2018) focused on trade‑off and internal‑ensemble strategies for semantic segmentation 🖥️. He then pursued a Ph.D. (2018‑2022), crafting a dissertation that models long‑term potentiation to segment depth images and detect objects of similar color—melding neuroscience inspiration with machine learning innovation 🧠🤖.
Professional Development 🚀
Since March 2022, Jeon has served as a postdoctoral fellow and lecturer in the Robot Vision Laboratory, spearheading the Smart Manufacturing ICT Project while guiding student capstones. Earlier stints as head teaching assistant saw him deliver TensorFlow tutorials and assist AI courses, cementing his reputation as an engaging educator 🎤. Beyond academia, he collaborates with industry partners on predictive‑maintenance apps, VR systems for rural migrants, and real‑time edge‑AI pipelines that judge vehicle loads or monitor cucumber diseases 🥒. His toolbox spans C/C++/C#, Java, Python, web stacks, and the full deep‑learning triad—TensorFlow, PyTorch, and Keras. Active on GitHub and ResearchGate, he shares reproducible code, datasets, and pre‑prints, nurturing an open‑science ethos 🌐. Multilingual in English and Japanese 🗣️, Jeon frequently presents at ICEIC, iFUZZY, and ICAIIC, expanding his global network while refining proposals for smart‑factory funding and cross‑disciplinary AI initiatives.
Research Focus🔍
Jeon’s core pursuit is applied deep learning for intelligent perception. He designs lightweight CNNs, attention‑enhanced U‑Nets, and ViT hybrids to tackle semantic segmentation, object detection, and human‑pose estimation 👁️🗨️. Agriculture is a recurring theme: his Att‑NestedUNet and DM‑YOLOv8 series detect weeds, pests, and crop diseases in real time, promising pesticide reductions and yield gains 🌾. In remote sensing, he crafts small‑target detectors for KOMPSAT‑3A imagery and lightweight optical payloads 🛰️, aiding environmental monitoring. Parallel biomedical studies segment skin lesions and diagnose otologic disorders, while industrial projects predict tool wear and production‑line failures ⚙️. Jeon’s philosophy fuses domain knowledge with edge‑friendly AI, optimizing inference speed without sacrificing accuracy. He experiments with neuro‑inspired mechanisms—long‑term potentiation, multiple‑path feature aggregation—to push beyond conventional CNN limits, always validating on tough, imbalanced datasets. The unifying thread is robust, resource‑aware vision that moves from lab to field seamlessly.
Awards & Honors 🏅🎉
Jeon’s innovation has been repeatedly recognized. He clinched the 1st Prize in South Korea’s National R&D Real Challenge Program (2018) 🚀, followed by back‑to‑back Excellent Paper Awards from the Korea Institute of Intelligent Systems (2018, 2019) and a 2017 accolade from the Korea Institute of Information Scientists and Engineers 📜. His marine‑garbage detection and vegetation‑classification studies each earned KIIS honors, while his bronze prize at the Korea Institute of Information Technology (2022) celebrated advances in lightweight AI for smart factories 🏭. Collectively, these distinctions underscore both scholarly rigor and real‑world impact, reinforcing his stature as a rising star in applied computer vision ✨
Publication Top Notes
1. Plant Leaf Recognition Using a Convolution Neural Network
Authors: WS Jeon, SY Rhee
Journal: International Journal of Fuzzy Logic and Intelligent Systems, Vol. 17(1), pp. 26–34
Year: 2017
Citations: 215
Summary:
This study presents a convolutional neural network (CNN)-based method for the automatic recognition of plant leaves. The authors demonstrate that CNNs can effectively extract hierarchical features from leaf images, outperforming traditional machine learning techniques. The proposed model is tested on standard leaf image datasets, achieving high accuracy and robustness. This work significantly contributes to smart agriculture and plant species identification.
2. Fingerprint Pattern Classification Using Convolution Neural Network
Authors: WS Jeon, SY Rhee
Journal: International Journal of Fuzzy Logic and Intelligent Systems, Vol. 17(3), pp. 170–176
Year: 2017
Citations: 63
Summary:
The paper explores fingerprint pattern classification using CNNs to overcome limitations of feature-based techniques. The model successfully distinguishes between fingerprint classes (e.g., whorl, loop, arch), demonstrating the CNN’s capability in biometric recognition tasks. This application shows potential in security and authentication systems, offering a data-driven approach without relying heavily on handcrafted features.
3. Comparative Analysis of Generalized Intersection over Union and Error Matrix for Vegetation Cover Classification Assessment
Authors: H Choi, HJ Lee, HJ You, SY Rhee, WS Jeon
Journal: Sensors and Materials, Vol. 31(11), pp. 3849–3858
Year: 2019
Citations: 21
Summary:
This paper evaluates two metrics—Generalized Intersection over Union (GIoU) and Error Matrix—for assessing vegetation classification from remote sensing data. Using classification outputs of land cover maps, the study highlights how GIoU provides a more nuanced understanding of spatial accuracy compared to traditional methods. It aids in improving evaluation frameworks for environmental monitoring.
4. Analysis of Deep Learning Applicability for KOMPSAT-3A Satellite Image Classification
Authors: SY Rhee, WS Jeon, H Choi
Journal: Journal of the Korean Society for Geospatial Information Science, Vol. 26(4), pp. 69–76
Year: 2018
Citations: 9
Summary:
The research investigates the applicability of deep learning, particularly CNNs, in classifying high-resolution images from the KOMPSAT-3A satellite. The paper demonstrates how deep learning can enhance land use and land cover classification, suggesting improvements in geospatial data interpretation and practical applications in remote sensing.
5. Sugar Beets and Weed Detection Using Semantic Segmentation
Authors: XZ Hu, WS Jeon, SY Rhee
Conference: 2022 International Conference on Fuzzy Theory and Its Applications (iFUZZY), pp. 1–4
Year: 2022
Citations: 7
Summary:
This paper presents a semantic segmentation approach using deep learning to differentiate between sugar beets and weeds in field images. The approach applies a CNN-based segmentation model to pixel-wise classify crops and unwanted plants, facilitating precision agriculture and automated weeding systems. The study supports the development of smart farming technologies.
Conclusion
Dr. Wang‑Su Jeon stands out as a highly deserving recipient of the Best Researcher Award. His work not only advances the frontiers of computer vision and AI but also addresses pressing societal needs with practical, deployable solutions. His blend of academic brilliance, industry collaboration, and social responsibility defines a new generation of researcher-leaders. He is not just a researcher but a catalyst for technological change—both in South Korea and on the global stage.