arXiv subjects

Yechan Kim

Publications and source records attributed to Yechan Kim.

14 recordsLinked to original sources

arXiv2026

G-MAD: A Game-Based Data Generation Framework for Multi-View RGB-T Aerial Object Detection

This work introduces G-MAD, an open-source framework that uses Arma3 to generate synchronized multi-view RGB-T data for aerial object detection. G-MAD addresses key limitations of real-world aerial dataset construction, including limited viewpoint control, imperfect RGB-T alignment and high annotation cost. The framework supports structured scenario specification, controllable multi-view camera placement, simultaneous visible/thermal capture, and automatic bounding box annotation using engine-level geometric metadata. These capabilities enable controlled studies of viewpoint variation, multi-modal fusion, and synthetic-to-real transfer in aerial object detection. Besides, using G-MAD, we construct and release AMOD, a new large-scale multi-view aerial RGB-T object detection benchmark. The source code and the dataset are available at https://unique-chan.github.io/G-MAD-Project.

cs.CV↗

arXiv2026

Wave-envelope dark matter beyond the monochromatic paradigm

Ultralight dark matter searches widely assume that signals are monochromatic, with a single frequency set by the mass. This assumption is generally violated in the presence of field mixing, even when the constituent fields have similar frequencies. Instead, dark matter signals can exhibit a two-timescale structure with intrinsic slow modulation. We demonstrate that mixing between ultralight wave dark matter fields induces a parametric structure, leading to a scenario we refer to as wave-envelope dark matter, in which a slow-beating envelope emerges alongside the primary oscillation. This results in distinctive features such as slow modulation and characteristic sideband structures in the frequency spectrum, beyond the conventional monochromatic expectation. As a representative example, we briefly discuss implications for neutrino observables.

hep-ph↗

arXiv2025

BCS superconductivity in the presence of wave dark matter

In the established era of dark matter, condensed matter Hamiltonians-including those of superconductors-may require extension to account for the surrounding Galactic environment. We show that if dark matter is wave-like and couples weakly to electrons, superconducting parameters such as the gap and critical temperature become dynamical quantities that oscillate in time. This modifies the Bardeen-Cooper-Schrieffer framework and produces distinctive temporal signatures whose sensitivity increases with longer measurement durations. Our results illustrate how condensed matter systems, traditionally treated as isolated from their cosmological environment, may acquire new dynamical degrees of freedom from their cosmic embedding. This, in turn, offers a novel window into the dark sector.

hep-ph↗

arXiv2025

Unlocking Robust Semantic Segmentation Performance via Label-only Elastic Deformations against Implicit Label Noise

While previous studies on image segmentation focus on handling severe (or explicit) label noise, real-world datasets also exhibit subtle (or implicit) label imperfections. These arise from inherent challenges, such as ambiguous object boundaries and annotator variability. Although not explicitly present, such mild and latent noise can still impair model performance. Typical data augmentation methods, which apply identical transformations to the image and its label, risk amplifying these subtle imperfections and limiting the model's generalization capacity. In this paper, we introduce NSegment+, a novel augmentation framework that decouples image and label transformations to address such realistic noise for semantic segmentation. By introducing controlled elastic deformations only to segmentation labels while preserving the original images, our method encourages models to focus on learning robust representations of object structures despite minor label inconsistencies. Extensive experiments demonstrate that NSegment+ consistently improves performance, achieving mIoU gains of up to +2.29, +2.38, +1.75, and +3.39 in average on Vaihingen, LoveDA, Cityscapes, and PASCAL VOC, respectively-even without bells and whistles, highlighting the importance of addressing implicit label noise. These gains can be further amplified when combined with other training tricks, including CutMix and Label Smoothing.

cs.CV↗

arXiv2025

NSegment : Label-specific Deformations for Remote Sensing Image Segmentation

Labeling errors in remote sensing (RS) image segmentation datasets often remain implicit and subtle due to ambiguous class boundaries, mixed pixels, shadows, complex terrain features, and subjective annotator bias. Furthermore, the scarcity of annotated RS data due to the high cost of labeling complicates training noise-robust models. While sophisticated mechanisms such as label selection or noise correction might address the issue mentioned above, they tend to increase training time and add implementation complexity. In this paper, we propose NSegment-a simple yet effective data augmentation solution to mitigate this issue. Unlike traditional methods, it applies elastic transformations only to segmentation labels, varying deformation intensity per sample in each training epoch to address annotation inconsistencies. Experimental results demonstrate that our approach improves the performance of RS image segmentation over various state-of-the-art models.

cs.CV↗

arXiv2025

Pallet Detection And Localisation From Synthetic Data

The global warehousing industry is experiencing rapid growth, with the market size projected to grow at an annual rate of 8.1% from 2024 to 2030 [Grand View Research, 2021]. This expansion has led to a surge in demand for efficient pallet detection and localisation systems. While automation can significantly streamline warehouse operations, the development of such systems often requires extensive manual data annotation, with an average of 35 seconds per image, for a typical computer vision project. This paper presents a novel approach to enhance pallet detection and localisation using purely synthetic data and geometric features derived from their side faces. By implementing a domain randomisation engine in Unity, the need for time-consuming manual annotation is eliminated while achieving high-performance results. The proposed method demonstrates a pallet detection performance of 0.995 mAP50 for single pallets on a real-world dataset. Additionally, an average position accuracy of less than 4.2 cm and an average rotation accuracy of 8.2{\deg} were achieved for pallets within a 5-meter range, with the pallet positioned head-on.

cs.CV↗

arXiv2025

Oscillating scalar potential and its implications for cosmic neutrino background searches

We propose a novel mechanism in which an external oscillatory wave modulates the mass-squared term of a scalar potential, periodically switching its sign. As a result of this "potential oscillation," the vacuum transitions between symmetry-broken and symmetry-restored phases. This repeated toggling leads to a time-varying vacuum state with rich phenomenological consequences, driven by the scalar field's couplings to other sectors. As a concrete illustration, we demonstrate how these oscillations can open a new avenue for probing the cosmic neutrino background.

hep-ph↗

arXiv2024

Dark gauge-mediated supersymmetry breaking with a massless dark photon

We study dark gauge-mediated supersymmetry breaking (dark GMSB) in a theory with a new unbroken $U(1)_{D}$ local symmetry and massless dark photon. Messenger fields charged under both Standard Model and dark gauge symmetries produce new soft supersymmetry-breaking terms due to gauge kinetic mixing between $U(1)_Y$ hypercharge and $U(1)_D$. We show that large kinetic mixing induces significant distortions to the superpartner spectra relative to conventional GMSB. Notably, shifts in the Higgs soft masses impact the conditions for electroweak symmetry breaking, lowering the $\mu$ parameter and yielding a relatively light Higgsino that may be accessible at the LHC. Furthermore, for very simple messenger representations, a very light bino-dark photino mixed state is present in the spectrum, which may be probed through exotic Higgs boson decays at future Higgs factories. We also examine the cosmological and phenomenological consequences of the messengers, the lightest of which is absolutely stable and carries fractional electric charge.

hep-ph↗

arXiv2024

NBBOX: Noisy Bounding Box Improves Remote Sensing Object Detection

Data augmentation has shown significant advancements in computer vision to improve model performance over the years, particularly in scenarios with limited and insufficient data. Currently, most studies focus on adjusting the image or its features to expand the size, quality, and variety of samples during training in various tasks including object detection. However, we argue that it is necessary to investigate bounding box transformations as a data augmentation technique rather than image-level transformations, especially in aerial imagery due to potentially inconsistent bounding box annotations. Hence, this letter presents a thorough investigation of bounding box transformation in terms of scaling, rotation, and translation for remote sensing object detection. We call this augmentation strategy NBBOX (Noise Injection into Bounding Box). We conduct extensive experiments on DOTA and DIOR-R, both well-known datasets that include a variety of rotated generic objects in aerial images. Experimental results show that our approach significantly improves remote sensing object detection without whistles and bells and it is more time-efficient than other state-of-the-art augmentation strategies.

cs.CV↗

arXiv2024

Investigating Long-term Training for Remote Sensing Object Detection

Recently, numerous methods have achieved impressive performance in remote sensing object detection, relying on convolution or transformer architectures. Such detectors typically have a feature backbone to extract useful features from raw input images. A common practice in current detectors is initializing the backbone with pre-trained weights available online. Fine-tuning the backbone is typically required to generate features suitable for remote-sensing images. While the prolonged training could lead to over-fitting, hindering the extraction of basic visual features, it can enable models to gradually extract deeper insights and richer representations from remote sensing data. Striking a balance between these competing factors is critical for achieving optimal performance. In this study, we aim to investigate the performance and characteristics of remote sensing object detection models under very long training schedules, and propose a novel method named Dynamic Backbone Freezing (DBF) for feature backbone fine-tuning on remote sensing object detection under long-term training. Our method addresses the dilemma of whether the backbone should extract low-level generic features or possess specific knowledge of the remote sensing domain, by introducing a module called 'Freezing Scheduler' to manage the update of backbone features during long-term training dynamically. Extensive experiments on DOTA and DIOR-R show that our approach enables more accurate model learning while substantially reducing computational costs in long-term training. Besides, it can be seamlessly adopted without additional effort due to its straightforward design. The code is available at https://github.com/unique-chan/dbf.

cs.CV↗

arXiv2024

Refined approaches in second leptogenesis for the baryon-lepton asymmetry discrepancy

The temperature-dependent mass of the heavy neutrino can lead to the second leptogenesis occurring below the electroweak scale, potentially explaining the large discrepancy between baryon and lepton asymmetries. We investigate this scenario further, exploring the intricate interplay of the weak interaction processes within this framework. It includes notable shifts in the dominant decay channels of heavy neutrinos around the electroweak symmetry breaking, along with the resonance behavior of the scattering processes near the $W/Z$ mass. The $CP$ asymmetry can also vary over cosmic history due to the temperature-dependent mass, allowing the $B-L$ asymmetry generation to be amplified in the late epoch. These findings elucidate how such alterations in the dynamics of second leptogenesis contribute to addressing the observed discrepancies in baryon-lepton asymmetry.

hep-ph↗

arXiv2023

Second leptogenesis: Unraveling the baryon-lepton asymmetry discrepancy

We propose a novel scenario to explain the matter-antimatter asymmetry by twofold leptogenesis, wherein heavy Majorana neutrinos exhibit temperature-dependent masses and engage in $CP$-violating decays. This scenario envisages two distinct phases of leptogenesis: one occurring above the electroweak scale and the other below it. The sphaleron process converts the first lepton asymmetry to baryon asymmetry, but not the second one due to its decoupling. This mechanism potentially explains the significant discrepancy between baryon and lepton asymmetries, as suggested by recent observations of Helium-4. Furthermore, our model implies that the present masses of Majorana neutrinos are lighter than the electroweak scale, offering a tangible avenue for experimental verification in various terrestrial settings.

hep-ph↗

arXiv2023

Dirac-Majorana neutrino type oscillation induced by a wave dark matter

Some properties of a neutrino may differ significantly depending on whether it is Dirac or Majorana type. The type is determined by the relative size of Dirac and Majorana masses, which may vary if they arise from an oscillating scalar dark matter. We show that the change can be significant enough to convert the neutrino type between Dirac and Majorana periodically while satisfying constraints on the dark matter. This neutrino type oscillation predicts periodic modulations in the event rates in various neutrino phenomena including the neutrinoless double beta decay. As the energy density and, thus, the oscillation amplitude of the dark matter evolves in the cosmic time scale, the neutrino masses change accordingly, which provides an interesting link between the present-time neutrino physics to the early universe cosmology including the leptogenesis.

hep-ph↗

arXiv2020

Imbalanced Image Classification with Complement Cross Entropy

Recently, deep learning models have achieved great success in computer vision applications, relying on large-scale class-balanced datasets. However, imbalanced class distributions still limit the wide applicability of these models due to degradation in performance. To solve this problem, in this paper, we concentrate on the study of cross entropy which mostly ignores output scores on incorrect classes. This work discovers that neutralizing predicted probabilities on incorrect classes improves the prediction accuracy for imbalanced image classification. This paper proposes a simple but effective loss named complement cross entropy based on this finding. The proposed loss makes the ground truth class overwhelm the other classes in terms of softmax probability, by neutralizing probabilities of incorrect classes, without additional training procedures. Along with it, this loss facilitates the models to learn key information especially from samples on minority classes. It ensures more accurate and robust classification results on imbalanced distributions. Extensive experiments on imbalanced datasets demonstrate the effectiveness of the proposed method.

cs.CV↗