YBYou-Suk Bae

You-Suk Bae

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


You-Suk Bae has a total of 18 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Traceable Cybersecurity Conformity Pre-Assessment of Electric Vehicle Charging Equipment Using Cross-Standard Gap Analysis and a Local LLM

    by Jin-Hyeok Kang, Seung-Hwan Lee, You-Suk Bae

    Electric vehicle charging equipment now links vehicles, payment and roaming services, cloud platforms, and the grid, but its cybersecurity requirements are fragmented across standards. This study combines a cross-standard gap analysis with a traceable conformity pre-assessment framework for a local large language model; the framework is a design-validation proof of concept, not a production-ready conformity-assessment system. Using EN 18031 as the reference axis, IEC 62443-4-2, ETSI EN 303 645, and NIST IR 8473 requirements were mapped on a five-level correspondence scale by a single researcher and consolidated into integrated requirements with applicability, evidence, and acceptance rules. A clause-aware retrieval-augmented generation pipeline (semantic retrieval only) was compared with no-retrieval and fixed-length baselines on 80 synthetic design-gold cases, reported on the 44 holdout cases not used for tuning. Better retrieval did not ensure trustworthy conformity reasoning: retrieved clause text in the judgment prompt lowered verdict accuracy (0.89 without, 0.77 with), the most accurate configuration (0.93) used retrieval only for citations, and a rule-based classifier with no model or corpus reached 0.66 on the same cases. Citation extraction, citation verification, and human review were each necessary. Extraction guarantees citation existence, but exact-gold-clause agreement remained at most 0.38. External validity on real manufacturer documents is untested; the framework complements rather than replaces accredited certification.

  • Open Access

    A Semi-Automatic Labeling Framework for PCB Defects via Deep Embeddings and Density-Aware Clustering

    by Sang-Jeong Lee, Sung-Bal Seo, You-Suk Bae

    (1) Background. Printed circuit board (PCB) inspection is increasingly constrained by the cost and latency of reliable labels, owing to tiny/low-contrast defects embedded in complex backgrounds and severe class imbalance. (2) Methods. We proposed a semi-automatic labeling pipeline that converts anomaly detection proposals into class labels via small margin cropping from images, interchangeable embeddings (HOG, ResNet-50, ViT-B/16), clustering (k-means/GMM/HDBSCAN), and cluster-level verification using representative montages. (3) Results. On 9354 cropped defects spanning 10 categories (imbalance IR ≈ 1542, Gini ≈ 0.642), ResNet-50 + HDBSCAN achieved NMI ≈ 0.290, AMI ≈ 0.283, and purity ≈ 0.624 with ~47 clusters; ViT + HDBSCAN was comparable (NMI ≈ 0.281, AMI ≈ 0.274, ~44 clusters). With a fixed taxonomy, k-means (K = 10) yielded the strongest ARI (0.169 with ResNet-50; 0.158 with ViT). Macro-purity exceeded micro-purity, indicating many small, homogeneous clusters suitable for one-shot acceptance/rejection, enabling an upper-bound ~200× reduction in operator decisions relative to per-image labeling. (4) Conclusions. The workflow provides an auditable, resource-flexible path from normal-only localization to scalable supervision, prioritizing labeling productivity over detector state-of-the-art and directly addressing the industrial bottleneck in the development lifecycle for PCB inspection.


  • Open Access

    Hybrid Deep-Ensemble Network with VAE-Based Augmentation for Imbalanced Tabular Data Classification

    by Sang-Jeong Lee, You-Suk Bae

    Background: Severe class imbalance limits reliable tabular AI in manufacturing, finance, and healthcare. Methods: We built a modular pipeline comprising correlation-aware seriation; a hybrid convolutional neural network (CNN)–transformer–Bidirectional Long Short-Term Memory (BiLSTM) encoder; variational autoencoder (VAE)-based minority augmentation; and deep/tree ensemble heads (XGBoost and Support Vector Machine, SVM). We benchmarked the Synthetic Minority Oversampling Technique (SMOTE) and ADASYN under identical protocols. Focal loss and ensemble weights were tuned per dataset. The primary metric was the Area Under the Precision–Recall Curve (AUPRC), with receiver operating characteristic area under the curve (ROC AUC) as complementary. Synthetic-data fidelity was quantified by train-on-synthetic/test-on-real (TSTR) utility, two-sample discriminability (ROC AUC of a real-vs-synthetic classifier), and Maximum Mean Discrepancy (MMD2). Results: Across five datasets (SECOM, CREDIT, THYROID, APS, and UCI), augmentation was data-dependent: VAE led on APS (+3.66 pp AUPRC vs. SMOTE) and was competitive on CREDIT (+0.10 pp vs. None); the SMOTE dominated SECOM; no augmentation performed best for THYROID and UCI. Positional embedding (PE) with seriation helped when strong local correlations were present. Ensembles typically favored XGBoost while benefiting from the hybrid encoder. Efficiency profiling and a slim variant supported latency-sensitive use. Conclusions: A data-aware recipe emerged: prefer VAE when fidelity is high, the SMOTE on smoother minority manifolds, and no augmentation when baselines suffice; apply PE/seriation selectively and tune per dataset for robust, reproducible deployment.


  • Open Access

    Stop-band Energy Constraint for Orthogonal Tunable Wavelet Units in Convolutional Neural Networks for Computer Vision problems

    by An D. Le, Hung Nguyen, Sungbal Seo, You-Suk Bae, Truong Q. Nguyen

    This work introduces a stop-band energy constraint for filters in orthogonal tunable wavelet units with a lattice structure, aimed at improving image classification and anomaly detection in CNNs, especially on texture-rich datasets. Integrated into ResNet-18, the method enhances convolution, pooling, and downsampling operations, yielding accuracy gains of 2.48% on CIFAR-10 and 13.56% on the Describable Textures dataset. Similar improvements are observed in ResNet-34. On the MVTec hazelnut anomaly detection task, the proposed method achieves competitive results in both segmentation and detection, outperforming existing approaches.

  • Open Access

    Biorthogonal Tunable Wavelet Unit with Lifting Scheme in Convolutional Neural Network

    by An Le, Hung Nguyen, Sungbal Seo, You-Suk Bae, Truong Nguyen

    This work introduces a novel biorthogonal tunable wavelet unit constructed using a lifting scheme that relaxes both the orthogonality and equal filter length constraints, providing greater flexibility in filter design. The proposed unit enhances convolution, pooling, and downsampling operations, leading to improved image classification and anomaly detection in convolutional neural networks (CNN). When integrated into an 18-layer residual neural network (ResNet-18), the approach improved classification accuracy on CIFAR-10 by 2.12% and on the Describable Textures Dataset (DTD) by 9.73%, demonstrating its effectiveness in capturing fine-grained details. Similar improvements were observed in ResNet-34. For anomaly detection in the hazelnut category of the MVTec Anomaly Detection dataset, the proposed method achieved competitive and wellbalanced performance in both segmentation and detection tasks, outperforming existing approaches in terms of accuracy and robustness.

  • Detection of Foreign Object in Cutting Surfaces using Artificial Intelligence in Plastic Processing

    by Dae-Cheol Kim, Jeong-Min Park, Su-Min Lee, In-Hee Hong, Seong-Kyu Kim, Won-Gok Lee, You-Seok Bae

    뿌리산업(소성가공산업)의 노후된 운영 방식은 전문인력 고령화, 생산현장 노하우 전수 어려움, 비숙련인력의 생산성 저하, 육안검사 기반 시스템의 비 정확성 및 품질 저하 등의 문제를 야기한다. 4차산업혁명과 맞물려, 인공지능을 적용하여 스마트 팩토리 구축으로 위의 문제들을 해결하는 것이 추세이다. 본 논문에서는 소성가공분야 단조 공정의 결함 발생을 줄이기 위해 객체검출 알고리즘 Faster- RCNN, SSD, Yolo의 특성을 분석하고, 알고리즘을 사용하여 절삭 면 이물질 검출 기법을 구현하였다. 또한, 소성가공 단조 공정에서 Faster-RCNN, SSD, Yolo를 사용한 절삭 면 이물질 검출 결과의 정확도와 mAP(mean Average Precision)을 비교하여 소성가공산업 분야에 최적화된 스마트 팩토리 구축 가능성을 입증하였다.


  • Recognition and Analysis of Ready-Mixed Concrete Slump using Image Processing and Deep-Learning

    by Geun-Oh Choi, Seong-Kyu Kim, Won-Gok Lee, Jong-Se Lee, You-Suk Bae

    레미콘의 슬럼프 품질관리는 현재 샘플링으로 검사되어 전량검사보다 정확도가 떨어지며 현재의 방법으로는 시간도 오래 걸린다. 본 논문에서는 슬럼프 콘을 이용한 현재의 방식 대신 카메라를 통해 획득한 비디오 영상을 기반으로 콘크리트 슬럼프의 질기를 파악하는 방법을 기술하였다. 이 방법은 LBP(Local Binaray Pattern), GLCM(Gray-Level Co-occurrence Matrix)의 기본적인 영상처리와 Optical Flow를 사용하여 영상에서의 속력을 계산한 뒤 LSTM(Long Short-Term Memory Recurrent Neural Network)을 활용하여 구현하였으며 비디오 데이터를 통해 슬럼프 질기를 분석하였다.