HSHa-Yoon Song

Ha-Yoon Song

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Ha-Yoon Song has a total of 13 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

    • 17 November 2026
    • Published by MDPI AG in Agronomy
      • Vol. 16  (22) , 2321
      • https://doi.org/10.3390/agronomy16222321
    Open Access

    From Bench to Conveyor: A Four-Generation Deployment Study of an Automated Grading System for Pine Mushroom

    by Jeon-Seong Kang, Jonghun Jeong, Beom-Joon Park, Junwon Yoon, Ha-Yoon Song, Hyun-Joon Chung, Dong-Geol Choi

    This is a longitudinal deployment study and not a new model. We report four development cycles (2025–2026) of a grading system for pine mushroom (Tricholoma matsutake), the highest-value non-timber forest product traded in the Republic of Korea, as it moved from a bench dataset to a sorting conveyor, and we separate image-based agreement with recorded inspector grades, engineering measurements, functional field demonstrations, and a belt-level experiment. All figures were re-derived by re-running the delivered models on their test partitions. Generation 1 combined a dual-view CNN ensemble with a morphometric rule (96.70% agreement, 8.5 FPS). Generation 2 reached 30 FPS by removing a backbone, because three networks on one GPU at batch 1 contended rather than overlapped. Generation 3 moved to three cameras; on new data, the earlier architecture fell to 70.0%, and, in a first field trial, frozen–thawed specimens defeated segmentation while the classifiers kept emitting grades. Generation 4 agreed with the inspector on 60 of 61 specimens (98.36%; 95% CI 91.3–99.7%) of a cleaned internal split that was also used for checkpoint selection, at 55 FPS inference throughput; a second trial demonstrated PLC-driven sorting, and a belt-level experiment that fed the 61 test specimens through a laboratory conveyor sorter sent 60 (98.36%) to the lane of their inspector grade—the same result as on the stored images. The re-analysis showed that the Generation 3 pipeline agreed on only 32.8% of the later Generation 4 collection; that a speed optimisation decorrelated the stipe thickness estimate (r=−0.32); and that, in Generation 4, both classifiers yielded identical predictions, one had collapsed onto a single camera, and the rule alone agreed on 52.5%. The accuracies are not a progress curve, the indicators are unvalidated image-derived proxies, and the system is a single-site prototype; we state the evaluations that a commercial claim would require.

  • Performance Comparison of Thermal Human Detection Models for Small Reconnaissance Robots

    by Jeon-Seong Kang, Beom-Joon Park, Junwon Yoon, Ha-Yoon Song, Hyun-Joon Chung, Duck-Hyun Suh, Taejoon Yoon, Jooyoung Kim

  • Open Access

    Capability-Conditioned Relay Placement for Geometrically Feasible Chains in Multi-Robot Teams with Heterogeneous Communication Ranges

    by Beom-Joon Park, Junwon Yoon, Jeon-Seong Kang, Ha-Yoon Song, Duck-Hyun Suh, Taejoon Yoon, Jooyoung Kim, Hyun-Joon Chung

    Multi-robot task allocation can place relay and mission goals in one matrix; whether this controls geometric relay-chain feasibility in execution had not been established. A deployed unified formulation was evaluated in closed-loop Robot Operating System 2 (ROS 2)/Gazebo simulation on six identical robots with synthetic ranges and no simulated radio, as component-level assurance for human-supervised systems, not a mission-success estimate. As deployed, a mission task displaced a relay in 20/20 trials; a cost-divisor capability model assigned hops beyond holders’ ranges. After two established repairs, capability-conditioned relay placement (CCRP) chose relay positions and holders jointly, bounding each hop by both holders’ ranges. On the mixed-range fleet, CCRP formed a geometrically feasible chain in 10/10 trials, in 16.5 s (median) against 21.3 s for equal-hop placement in one session (in-session gains 1.1–10.05 s, not all significant). Across five arrangement–target cells, plans were reproduced in 100/100 trials. CCRP’s median formation time was lower in all four cells whose holders differed, significantly in two; in one, the capability-blind fall-through held no chain. CCRP produced geometrically feasible plans and, on the mixed-range fleet, faster chain formation; stalls beside stationary robots remained outside its distance objective. No transfer to untested maps, radios or hardware is claimed.

  • Open Access

    RipenessGAN: Growth Day Embedding-Enhanced GAN for Stage-Wise Jujube Ripeness Data Generation

    by Jeon-Seong Kang, Junwon Yoon, Beom-Joon Park, Junyoung Kim, Sung Chul Jee, Ha-Yoon Song, Hyun-Joon Chung

    RipenessGAN is a novel Generative Adversarial Network (GAN) designed to generate synthetic images across different ripeness stages of jujubes (green fruit, white ripe fruit, semi-red fruit, and fully red fruit), aiming to provide balanced training data for diverse applications beyond classification accuracy. This study addresses the problem of data imbalance by augmenting each ripeness stage using our proposed Growth Day Embedding mechanism, thereby enhancing the performance of downstream classification models. The core innovation of RipenessGAN lies in its ability to capture continuous temporal transitions among discrete ripeness classes by incorporating fine-grained growth day information (0–56 days) in addition to traditional class labels. The experimental results show that RipenessGAN produces synthetic data with higher visual quality and greater diversity compared to CycleGAN. Furthermore, the classification models trained on the enriched dataset exhibit more consistent and accurate performance. We also conducted comprehensive comparisons of RipenessGAN against CycleGAN and class-conditional diffusion models (DDPM) under strictly controlled and fair experimental settings, carefully matching model architectures, computational resources, training conditions, and evaluation metrics. The results indicate that although diffusion models yield highly realistic images and CycleGAN ensures stable cycle-consistent generation, RipenessGAN provides superior practical benefits in training efficiency, temporal controllability, and adaptability for agricultural applications. This research demonstrates the potential of RipenessGAN to mitigate data imbalance in agriculture and highlights its scalability to other crops.


  • Open Access

    A Transfer Learning Approach for Diverse Motion Augmentation Under Data Scarcity

    by Junwon Yoon, Jeon-Seong Kang, Ha-Yoon Song, Beom-Joon Park, Kwang-Woo Jeon, Hyun-Joon Chung, Jang-Sik Park

    Motion-capture data provide high accuracy but are difficult to obtain, necessitating dataset augmentation. To our knowledge, no prior study has investigated few-shot generative models for motion-capture data that address both quality and diversity. We tackle the diversity loss that arises with extremely small datasets (n ≤ 10) by applying transfer learning and continual learning to retain the rich variability of a larger pretraining corpus. To assess quality, we introduce MFMMD (Motion Feature-Based Maximum Mean Discrepancy)—a metric well-suited for small samples—and evaluate diversity with the multimodality metric. Our method embeds an Elastic Weight Consolidation (EWC)-based regularization term in the generator’s loss and then fine-tunes the limited motion-capture set. We analyze how the strength of this term influences diversity and uncovers motion-specific characteristics, revealing behavior that differs from that observed in image-generation tasks. The experiments indicate that the transfer learning pipeline improves generative performance in low-data scenarios. Increasing the weight of the regularization term yields higher diversity in the synthesized motions, demonstrating a marked uplift in motion diversity. These findings suggest that the proposed approach can effectively augment small motion-capture datasets with greater variety, a capability expected to benefit applications that rely on diverse human-motion data across modern robotics, animation, and virtual reality.