arXiv subjects

Sungho Kim

Publications and source records attributed to Sungho Kim.

At least 19 recordsLinked to original sources

arXiv2026

GaussVLA: Geometry-Aware Spatial Reasoning for Vision-Language-Action Model

Vision-Language-Action (VLA) models encode visual observations as flat 2D patch tokens that carry no intrinsic geometric structure, and augmenting them with dense monocular depth injects per-pixel scalar values that encode neither surface orientation nor geometric confidence. This leaves the policy with limited structured spatial reasoning for action prediction. We propose GaussVLA, a Mamba-based VLA that incorporates two custom modules: Gaussian Spatial Tokenizer (GST) to lift frozen semantic and depth features into compact 3D Gaussian tokens, pools geometrically salient regions with learned queries, and \emph{Depth-Aware Chain-of-Thought (DA-CoT)} that performs structured, non-autoregressive geometric reasoning under language and flow-time conditioning. Across both simulation and real-world evaluations, GaussVLA demonstrates strong spatial-manipulation performance while remaining parameter-efficient. On LIBERO, it achieves 93.5% average success and 100.0% success on the Spatial suite with only 200M parameters, improving over SpatialVLA by 19.7% relative average success while remaining significantly more parameter-efficient.

cs.RO↗

arXiv2026

Rabinowitz Floer homology for Legendrian submanifolds in prequantization bundles

Let $Y$ be a prequantization bundle over an integral symplectic manifold $(\Sigma,\omega)$. Let $L$ be a closed monotone Lagrangian submanifold that admits a Legendrian lift $\mathcal{L}$ in $Y$. Under the assumption that the minimal Maslov number $N_L$ of $L$ is greater than 2, we define the Rabinowitz Floer homology of $\mathcal{L}$. We then establish an isomorphism between the $\mathbb{Z}_d$-equivariant Rabinowitz Floer homology of $\mathcal{L}$ and the quantum homology of $L$, where $d$ is the degree of the covering map $\mathcal{L}\to L$. Under a more restrictive condition on $N_L$, we show that this map is a ring isomorphism. Using this isomorphism, we compute the quantum homology ring of Lagrangian spheres in quadrics and two-step flag manifolds. Furthermore, we investigate the implications of the quantum invertibility of $\omega$ for the vanishing of the quantum homology of $L$ and the obstructions to topologically simple fillings of $\mathcal{L}$. We also show that if $(\Sigma,\omega)$ admits a polarization and $L$ is disjoint from the Lagrangian trace, the quantum homology of $L$ vanishes.

math.SG↗

arXiv2026

GST-VLA: Structured Gaussian Spatial Tokens for 3D Depth-Aware Vision-Language-Action Models

VLA models encode visual observations as 2D patch tokens with no intrinsic geometric structure. We introduce GST-VLA with two contributions. First, the Gaussian Spatial Tokenizer (GST) converts frozen dense depth and frozen semantic patch features into $N_g{=}128$ anisotropic 3D Gaussian primitives, each parameterized by a metric residual mean $\mu \in \mathbb{R}^3$, log-scale covariance $\log \sigma \in \mathbb{R}^3$, and learned opacity $\alpha \in (0,1)$. The covariance eigenstructure encodes local surface orientation, and opacity provides per-primitive geometric confidence, both inaccessible from scalar depth. Spatial attention pooling with learned queries concentrates the fixed token budget on geometrically salient regions rather than distributing uniformly. Second, 3D Depth-Aware Chain-of-Thought (DA-CoT) reasoning supervises four structured intermediate spatial thoughts, covering 3D object grounding, grasp affordance contact geometry, pairwise metric distances, and coarse SE(3) waypoints, as explicit generation targets in the training loss. A cross-attention sublayer at every VLM transformer block provides direct access to the raw 256-primitive Gaussian field during DA-CoT generation. A 300M-parameter flow-matching action expert with mixture-of-experts feedforward sublayers decodes 7-DoF delta action chunks via conditional ODE integration, conditioned on both VLM hidden states and DA-CoT outputs through dual cross-attention. Trained with composite $\mathcal{L}_\mathrm{flow} + \mathcal{L}_\mathrm{CoT} + \mathcal{L}_\mathrm{depth}$ across three progressive stages, GST-VLA achieves 96.4% on LIBERO (+2.0%), and 80.2% on SimplerEnv (+5.4%). Ablations isolate the contribution of each GST component, each DA-CoT thought, and each training stage, confirming independent and synergistic gains concentrated on precision demanding tasks.

cs.CV↗

arXiv2025

VFM-VLM: Vision Foundation Model and Vision Language Model based Visual Comparison for 3D Pose Estimation

Vision Foundation Models (VFMs) and Vision Language Models (VLMs) have revolutionized computer vision by providing rich semantic and geometric representations. This paper presents a comprehensive visual comparison between CLIP based and DINOv2 based approaches for 3D pose estimation in hand object grasping scenarios. We evaluate both models on the task of 6D object pose estimation and demonstrate their complementary strengths: CLIP excels in semantic understanding through language grounding, while DINOv2 provides superior dense geometric features. Through extensive experiments on benchmark datasets, we show that CLIP based methods achieve better semantic consistency, while DINOv2 based approaches demonstrate competitive performance with enhanced geometric precision. Our analysis provides insights for selecting appropriate vision models for robotic manipulation and grasping, picking applications.

cs.CV↗

arXiv2025

Explainable Parkinsons Disease Gait Recognition Using Multimodal RGB-D Fusion and Large Language Models

Accurate and interpretable gait analysis plays a crucial role in the early detection of Parkinsons disease (PD),yet most existing approaches remain limited by single-modality inputs, low robustness, and a lack of clinical transparency. This paper presents an explainable multimodal framework that integrates RGB and Depth (RGB-D) data to recognize Parkinsonian gait patterns under realistic conditions. The proposed system employs dual YOLOv11-based encoders for modality-specific feature extraction, followed by a Multi-Scale Local-Global Extraction (MLGE) module and a Cross-Spatial Neck Fusion mechanism to enhance spatial-temporal representation. This design captures both fine-grained limb motion (e.g., reduced arm swing) and overall gait dynamics (e.g., short stride or turning difficulty), even in challenging scenarios such as low lighting or occlusion caused by clothing. To ensure interpretability, a frozen Large Language Model (LLM) is incorporated to translate fused visual embeddings and structured metadata into clinically meaningful textual explanations. Experimental evaluations on multimodal gait datasets demonstrate that the proposed RGB-D fusion framework achieves higher recognition accuracy, improved robustness to environmental variations, and clear visual-linguistic reasoning compared with single-input baselines. By combining multimodal feature learning with language-based interpretability, this study bridges the gap between visual recognition and clinical understanding, offering a novel vision-language paradigm for reliable and explainable Parkinsons disease gait analysis. Code:https://github.com/manaralnaasan/RGB-D_parkinson-LLM

cs.CV↗

arXiv2025

VLM6D: VLM based 6Dof Pose Estimation based on RGB-D Images

The primary challenge in computer vision is precisely calculating the pose of 6D objects, however many current approaches are still fragile and have trouble generalizing from synthetic data to real-world situations with fluctuating lighting, textureless objects, and significant occlusions. To address these limitations, VLM6D, a novel dual-stream architecture that leverages the distinct strengths of visual and geometric data from RGB-D input for robust and precise pose estimation. Our framework uniquely integrates two specialized encoders: a powerful, self-supervised Vision Transformer (DINOv2) processes the RGB modality, harnessing its rich, pre-trained understanding of visual grammar to achieve remarkable resilience against texture and lighting variations. Concurrently, a PointNet++ encoder processes the 3D point cloud derived from depth data, enabling robust geometric reasoning that excels even with the sparse, fragmented data typical of severe occlusion. These complementary feature streams are effectively fused to inform a multi task prediction head. We demonstrate through comprehensive experiments that VLM6D obtained new SOTA performance on the challenging Occluded-LineMOD, validating its superior robustness and accuracy.

cs.CV↗

arXiv2024

Soft Segmented Randomization: Enhancing Domain Generalization in SAR-ATR for Synthetic-to-Measured

Synthetic aperture radar technology is crucial for high-resolution imaging under various conditions; however, the acquisition of real-world synthetic aperture radar data for deep learning-based automatic target recognition remains challenging due to high costs and data availability issues. To overcome these challenges, synthetic data generated through simulations have been employed, although discrepancies between synthetic and real data can degrade model performance. In this study, we introduce a novel framework, soft segmented randomization, designed to reduce domain discrepancy and improve the generalize ability of synthetic aperture radar automatic target recognition models. The soft segmented randomization framework applies a Gaussian mixture model to segment target and clutter regions softly, introducing randomized variations that align the synthetic data's statistical properties more closely with those of real-world data. Experimental results demonstrate that the proposed soft segmented randomization framework significantly enhances model performance on measured synthetic aperture radar data, making it a promising approach for robust automatic target recognition in scenarios with limited or no access to measured data.

cs.CV↗

arXiv2024

Diffusion Deepfake

Recent progress in generative AI, primarily through diffusion models, presents significant challenges for real-world deepfake detection. The increased realism in image details, diverse content, and widespread accessibility to the general public complicates the identification of these sophisticated deepfakes. Acknowledging the urgency to address the vulnerability of current deepfake detectors to this evolving threat, our paper introduces two extensive deepfake datasets generated by state-of-the-art diffusion models as other datasets are less diverse and low in quality. Our extensive experiments also showed that our dataset is more challenging compared to the other face deepfake datasets. Our strategic dataset creation not only challenge the deepfake detectors but also sets a new benchmark for more evaluation. Our comprehensive evaluation reveals the struggle of existing detection methods, often optimized for specific image domains and manipulations, to effectively adapt to the intricate nature of diffusion deepfakes, limiting their practical utility. To address this critical issue, we investigate the impact of enhancing training data diversity on representative detection methods. This involves expanding the diversity of both manipulation techniques and image domains. Our findings underscore that increasing training data diversity results in improved generalizability. Moreover, we propose a novel momentum difficulty boosting strategy to tackle the additional challenge posed by training data heterogeneity. This strategy dynamically assigns appropriate sample weights based on learning difficulty, enhancing the model's adaptability to both easy and challenging samples. Extensive experiments on both existing and newly proposed benchmarks demonstrate that our model optimization approach surpasses prior alternatives significantly.

cs.CV↗

arXiv2023

Rabinowitz Floer homology for prequantization bundles and Floer Gysin sequence

Let $Y$ be a prequantization bundle over a closed spherically monotone symplectic manifold $\Sigma$. Adapting an idea due to Diogo and Lisi, we study a split version of Rabinowitz Floer homology for $Y$ in the following two settings. First, $\Sigma$ is a symplectic hyperplane section of a closed symplectic manifold $X$ satisfying a certain monotonicity condition; in this case, $X \setminus \Sigma$ is a Liouville filling of $Y$. Second, the minimal Chern number of $\Sigma$ is greater than one, which is the case where the Rabinowitz Floer homology of the symplectization $\mathbb{R} \times Y$ is defined. In both cases, we construct a Gysin-type exact sequence connecting the Rabinowitz Floer homology of $X\setminus\Sigma$ or $\mathbb{R} \times Y$ and the quantum homology of $\Sigma$. As applications, we discuss the invertibility of a symplectic hyperplane section class in quantum homology, the isotopy problem for fibered Dehn twists, the orderability problem for prequantization bundles, and the existence of translated points. We also provide computational results based on the exact sequence that we construct.

math.SG↗

arXiv2020

NTIRE 2020 Challenge on Real Image Denoising: Dataset, Methods and Results

This paper reviews the NTIRE 2020 challenge on real image denoising with focus on the newly introduced dataset, the proposed methods and their results. The challenge is a new version of the previous NTIRE 2019 challenge on real image denoising that was based on the SIDD benchmark. This challenge is based on a newly collected validation and testing image datasets, and hence, named SIDD+. This challenge has two tracks for quantitatively evaluating image denoising performance in (1) the Bayer-pattern rawRGB and (2) the standard RGB (sRGB) color spaces. Each track ~250 registered participants. A total of 22 teams, proposing 24 methods, competed in the final phase of the challenge. The proposed methods by the participating teams represent the current state-of-the-art performance in image denoising targeting real noisy images. The newly collected SIDD+ datasets are publicly available at: https://bit.ly/siddplus_data.

cs.CV↗

arXiv2020

The 1st Agriculture-Vision Challenge: Methods and Results

The first Agriculture-Vision Challenge aims to encourage research in developing novel and effective algorithms for agricultural pattern recognition from aerial images, especially for the semantic segmentation task associated with our challenge dataset. Around 57 participating teams from various countries compete to achieve state-of-the-art in aerial agriculture semantic segmentation. The Agriculture-Vision Challenge Dataset was employed, which comprises of 21,061 aerial and multi-spectral farmland images. This paper provides a summary of notable methods and results in the challenge. Our submission server and leaderboard will continue to open for researchers that are interested in this challenge dataset and task; the link can be found here.

cs.CV↗

arXiv2016

Resistive Switching Characteristics of Al/Si3N4/p-Si MIS-Based Resistive Switching Memory Devices

In this study, we proposed and demonstrated a self-rectifying property of silicon nitride (Si3N4)-based resistive random access memory device by employing p-type silicon (p-Si) as bottom electrode. The RRAM devices consisted of Al/Si3N4/p-Si are fabricated by a low presure chemical vapor deposition and exhibited an intrinsic diode property with non-linear current-voltage (I-V) behavior. In addition, compared to conventional metal/insulator/metal (MIM) structure of Al/Si3N4/Ti RRAM cells, operating current in whole bias regions for proposed metal/insulator/semiconductor (MIS) cells has been dramatically lowered because introduced p-Si bottom electrode efficiently suppresses the current in both low and high resistive states. As a result, the results mean that by employing p-Si as bottom electrode the Si3N4-based RRAM cells can be applied to selector-free RRAM cells.

cond-mat.mtrl-sci↗

arXiv2016

Resistive Switching Phenomena of HfO2 Films Grown by MOCVD for Resistive Switching Memory Devices

The resistive switching phenomena of HfO2 films grown by metalorganic chemical vapor deposition was studied for the application of ReRAM devices. In the fabricated Pt/HfO2/TiN memory cells, the bipolar resistive switching characteristics were observed, and the set and reset states were measured to be as low as 7 uA and 4 uA, respectively, at VREAD = 1 V. Regarding the resistive switching performance, the stable RS performance was observed under 40 repetitive dc cycling test with the small variations of set/reset voltages and currents, and good retention characteristics over 105 s in both LRS and HRS. These results show the possibility of MOCVD grown HfO2 films as a promising resistive switching materials for ReRAM applications.

cond-mat.mtrl-sci↗

arXiv2015

Site occupancy and magnetic properties of Al-substituted M-type strontium hexaferrite

We use first-principles total-energy calculations based on density functional theory to study the site occupancy and magnetic properties of Al-substituted $M$-type strontium hexaferrite SrFe$_{12-x}$Al$_{x}$O$_{19}$ with $x=0.5$ and $x=1.0$. We find that the non-magnetic Al$^{3+}$ ions preferentially replace Fe$^{3+}$ ions at two of the majority spin sites, $2a$ and $12k$, eliminating their positive contribution to the total magnetization causing the saturation magnetization $M_s$ to be reduced as Al concentration $x$ is increased. Our formation probability analysis further provides the explanation for increased magnetic anisotropy field when the fraction of Al is increased. Although Al$^{3+}$ ions preferentially occupy the $2a$ sites at a low temperature, the occupation probability of the $12k$ site increases with the rise of the temperature. At a typical annealing temperature ($> 700\,^{\circ}{\rm C}$) Al$^{3+}$ ions are much more likely to occupy the $12k$ site than the $2a$ site. Although this causes the magnetocrystalline anisotropy $K_1$ to be reduced slightly, the reduction in $M_s$ is much more significant. Their combined effect causes the anisotropy field $H_a$ to increase as the fraction of Al is increased, consistent with recent experimental measurements.

cond-mat.mtrl-sci↗

arXiv2014

Structural, elastic and thermal properties of cementite (Fe$_3$C) calculated using Modified Embedded Atom Method

Structural, elastic and thermal properties of cementite (Fe$_3$C) were studied using a Modified Embedded Atom Method (MEAM) potential for iron-carbon (Fe-C) alloys. Previously developed Fe and C single element potentials were used to develop an Fe-C alloy MEAM potential, using a statistically-based optimization scheme to reproduce structural and elastic properties of cementite, the interstitial energies of C in bcc Fe as well as heat of formation of Fe-C alloys in L$_{12}$ and B$_1$ structures. The stability of cementite was investigated by molecular dynamics simulations at high temperatures. The nine single crystal elastic constants for cementite were obtained by computing total energies for strained cells. Polycrystalline elastic moduli for cementite were calculated from the single crystal elastic constants of cementite. The formation energies of (001), (010), and (100) surfaces of cementite were also calculated. The melting temperature and the variation of specific heat and volume with respect to temperature were investigated by performing a two-phase (solid/liquid) molecular dynamics simulation of cementite. The predictions of the potential are in good agreement with first-principles calculations and experiments.

cond-mat.mtrl-sci↗

arXiv2013

Theory of magnetic enhancement in strontium hexaferrite through Zn-Sn pair substitution

We study the site occupancy and magnetic properties of Zn-Sn substituted M-type Sr-hexaferrite SrFe$_{12-x}$(Zn$_{0.5}$Sn$_{0.5}$)$_x$O$_{19}$ with x = 1 using first-principles total-energy calculations. We find that in a ground-state configuration Zn-Sn ions preferentially occupy $4f_1$ and $4f_2$ sites unlike the model previously suggested by Ghasemi et al. [J. Appl. Phys, \textbf{107}, 09A734 (2010)], where Zn$^{2+}$ and Sn$^{4+}$ ions occupy the $2b$ and $4f_2$ sites. Density-functional theory calculations show that our model has a lower total energy by more than 0.2 eV per unit cell compared to Ghasemi's model. More importantly, the latter does not show an increase in saturation magnetization ($M_s$) compared to the pure $M$-type Sr-hexaferrite, in disagreement with the experiment. On the other hand, our model correctly predicts a rapid increase in $M_s$ as well as a decrease in magnetic anisotropy compared to the pure $M$-type Sr-hexaferrite, consistent with experimental measurements.

cond-mat.mtrl-sci↗

arXiv2012

The effects of Vanadium on the strength of a bcc Fe Σ3(111)[1-10] grain boundary

The effects of micro-alloying element, vanadium, on a bcc Fe Σ3(111)[1-10] symmetric tilt grain boundary strength are studied using density functional theory calculations. The lowest energy configuration of the grain boundary structure are obtained from the first-principles calculations. The substitutional and interstitial point defect formation energies of vanadium in the grain boundary are compared. The substitutional defect is prefered to interstitial one. The segregation energies of vanadium onto the grain boundary and its fractured surfaces are computed. The cohesive energy calculation of the grain boundary with and without vanadium show that vanadium strengthen the bcc iron Σ3(111)[1-10] grain boundary.

cond-mat.mtrl-sci↗

arXiv2011

B+-tree Index Optimization by Exploiting Internal Parallelism of Flash-based Solid State Drives

Previous research addressed the potential problems of the hard-disk oriented design of DBMSs of flashSSDs. In this paper, we focus on exploiting potential benefits of flashSSDs. First, we examine the internal parallelism issues of flashSSDs by conducting benchmarks to various flashSSDs. Then, we suggest algorithm-design principles in order to best benefit from the internal parallelism. We present a new I/O request concept, called psync I/O that can exploit the internal parallelism of flashSSDs in a single process. Based on these ideas, we introduce B+-tree optimization methods in order to utilize internal parallelism. By integrating the results of these methods, we present a B+-tree variant, PIO B-tree. We confirmed that each optimization method substantially enhances the index performance. Consequently, PIO B-tree enhanced B+-tree's insert performance by a factor of up to 16.3, while improving point-search performance by a factor of 1.2. The range search of PIO B-tree was up to 5 times faster than that of the B+-tree. Moreover, PIO B-tree outperformed other flash-aware indexes in various synthetic workloads. We also confirmed that PIO B-tree outperforms B+-tree in index traces collected inside the Postgresql DBMS with TPC-C benchmark.

cs.DB↗