arXiv subjects

Seungho Choi

Publications and source records attributed to Seungho Choi.

4 recordsLinked to original sources

arXiv2026

DiTTo: Scalable Order-aware All-in-One Image Restoration Agent

Real-world images rarely suffer from a single degradation, and the order in which degradations are removed substantially affects the final restoration quality, motivating agent-based image restoration (IR), where a vision-language model schedules a pool of pre-built restoration-experts. However, existing training-based agents require $\mathcal{O}((N^{\mathbf{D}})^{2})$ restoration-expert calls per image to construct the Optimal Restoration-action Trajectory Dataset (ORTD), where $N^{\mathbf{D}}$ denotes the number of degradation types in the universe $\mathbf{D}$, and couple agent training to a fixed restoration-expert pool, preventing extension to newly introduced restoration-experts without full retraining. To overcome these efficiency and extensibility bottlenecks, we propose \textbf{DiTTo}, a novel order-aware image restoration agent framework consisting of the DiTTo Simulator and the DiTTo Agent. The DiTTo Simulator combines $\cup$S-IR for single-step restoration-action simulation and AiO-IQA for per-action quality prediction, reducing ORTD construction to $\mathcal{O}(N^{\mathbf{D}})$ simulator calls per image; the DiTTo Agent is trained by SFT on the simulator-generated ORTD, followed by \textbf{Order-aware Restoration Alignment (ORA)} that aligns degradation identification, restoration-action-ordering, and output format along independent axes. This enables \textbf{plug-and-play scalable extensibility}: adding a new restoration-expert requires updating only the lightweight ORA stage. On the MiO-100 evaluation set with up to five concurrent degradations, our DiTTo Agent achieves state-of-the-art multi-degradation restoration quality among previous agent-based IR methods.

cs.CV↗

arXiv2026

FRAMER: Frequency-Aligned Self-Distillation with Adaptive Modulation Leveraging Diffusion Priors for Real-World Image Super-Resolution

Real-image super-resolution (Real-ISR) seeks to recover HR images from LR inputs with mixed, unknown degradations. While diffusion models surpass GANs in perceptual quality, they under-reconstruct high-frequency (HF) details due to a low-frequency (LF) bias and a depth-wise "low-first, high-later" hierarchy. We introduce FRAMER, a plug-and-play training scheme that exploits diffusion priors without changing the backbone or inference. At each denoising step, the final-layer feature map teaches all intermediate layers. Teacher and student feature maps are decomposed into LF/HF bands via FFT masks to align supervision with the model's internal frequency hierarchy. For LF, an Intra Contrastive Loss (IntraCL) stabilizes globally shared structure. For HF, an Inter Contrastive Loss (InterCL) sharpens instance-specific details using random-layer and in-batch negatives. Two adaptive modulators, Frequency-based Adaptive Weight (FAW) and Frequency-based Alignment Modulation (FAM), reweight per-layer LF/HF signals and gate distillation by current similarity. Across U-Net and DiT backbones (e.g., Stable Diffusion 2, 3), FRAMER consistently improves PSNR/SSIM and perceptual metrics (LPIPS, NIQE, MANIQA, MUSIQ). Ablations validate the final-layer teacher and random-layer negatives.

cs.CV↗

arXiv2025

Quantifying Fire Risk Index in Chemical Industry Using Statistical Modeling Procedure

Fire incident reports contain detailed textual narratives that capture causal factors often overlooked in structured records, while financial damage amounts provide measurable outcomes of these events. Integrating these two sources of information is essential for uncovering interpretable links between descriptive causes and their economic consequences. To this end, we develop a data-driven framework that constructs a composite Risk Index, enabling systematic quantification of how specific keywords relate to property damage amounts. This index facilitates both the identification of high-impact terms and the aggregation of risks across semantically related clusters, thereby offering a principled measure of fire-related financial risk. Using more than a decade of Korean fire investigation reports on the chemical industry classified as Special Buildings (2013 through 2024), we employ topic modeling and network-based embedding to estimate semantic similarities from interactions among words and subsequently apply Lasso regression to quantify their associations with property damage amounts, thereby estimate fire risk index. This approach enables us to assess fire risk not only at the level of individual terms but also within their broader textual context, where highly interactive related words provide insights into collective patterns of hazard representation and their potential impact on expected losses. The analysis highlights several domains of risk, including hazardous chemical leakage, unsafe storage practices, equipment and facility malfunctions, and environmentally induced ignition. The results demonstrate that text-derived indices provide interpretable and practically relevant insights, bridging unstructured narratives with structured loss information and offering a basis for evidence-based fire risk assessment and management.

stat.AP↗

arXiv2025

HanjaBridge: Resolving Semantic Ambiguity in Korean LLMs via Hanja-Augmented Pre-Training

Large language models (LLMs) often show poor performance in low-resource languages like Korean, partly due to unique linguistic challenges such as homophonous Sino-Korean words that are indistinguishable in Hangul script. To address this semantic ambiguity, we propose HanjaBridge, a novel meaning-injection technique integrated into a continual pre-training (CPT) framework. Instead of deterministically mapping a word to a single Hanja (Chinese character), HanjaBridge presents the model with all possible Hanja candidates for a given homograph, encouraging the model to learn contextual disambiguation. This process is paired with token-level knowledge distillation to prevent catastrophic forgetting. Experimental results show that HanjaBridge significantly improves Korean language understanding, achieving a 21\% relative improvement on the KoBALT benchmark. Notably, by reinforcing semantic alignment between Korean and Chinese through shared Hanja, we observe a strong positive cross-lingual transfer. Furthermore, these gains persist even when Hanja augmentation is omitted at inference time, ensuring practical efficiency with no additional run-time cost.

cs.CL↗