arXiv subjects

Hiroshi Watanabe

Publications and source records attributed to Hiroshi Watanabe.

At least 37 records · Page 2Linked to original sources

arXiv2025

Anisotropic Non-Fermi Liquid and Dynamical Planckian Scaling of a Quasi-Kagome Kondo Lattice System

At the quantum critical point of correlated materials, a non-Fermi liquid state appears where electron correlations continuously develop to very low temperatures. The relaxation time of the interacted electrons, namely quasiparticles, is scaled with the Planckian time, $\hbar/k_{\rm B}T$. However, there is a debate over whether heavy-fermion systems can obey the Planckian time. In the optical conductivity spectra, the Drude response will appear as the scaling of $\hbarω/k_{\rm B}T$ as the dynamical Planckian scaling (DPS). Here, we report the non-Fermi liquid behavior in the Drude response of a candidate for such materials, the quasi-kagome Kondo lattice CeRhSn. Even though the material shows a strong valence fluctuation, renormalized Drude responses observed at the photon energy below 100 meV are characterized by non-Fermi-liquid-like scattering rate $1/τ$. The heavy carriers' Drude response only for the Ce quasi-kagome plane obeyed DPS below 80 K, suggesting the anisotropic quantum criticality with the strong $c \textrm{-} f$ hybridization.

cond-mat.str-el↗

arXiv2025

Seed Selection for Human-Oriented Image Reconstruction via Guided Diffusion

Conventional methods for scalable image coding for humans and machines require the transmission of additional information to achieve scalability. A recent diffusion-based approach avoids this by generating human-oriented images from machine-oriented images without extra bitrate. However, it utilizes a single random seed, which may lead to suboptimal image quality. In this paper, we propose a seed selection method that identifies the optimal seed from multiple candidates to improve image quality without increasing the bitrate. To reduce the computational cost, selection is performed based on intermediate outputs obtained from early steps of the reverse diffusion process. Experimental results demonstrate that our proposed method outperforms the baseline, which uses a single random seed without selection, across multiple evaluation metrics.

cs.CV↗

arXiv2025

SR-NeRV: Improving Embedding Efficiency of Neural Video Representation via Super-Resolution

Implicit Neural Representations (INRs) have garnered significant attention for their ability to model complex signals in various domains. Recently, INR-based frameworks have shown promise in neural video compression by embedding video content into compact neural networks. However, these methods often struggle to reconstruct high-frequency details under stringent constraints on model size, which are critical in practical compression scenarios. To address this limitation, we propose an INR-based video representation framework that integrates a general-purpose super-resolution (SR) network. This design is motivated by the observation that high-frequency components tend to exhibit low temporal redundancy across frames. By offloading the reconstruction of fine details to a dedicated SR network pre-trained on natural images, the proposed method improves visual fidelity. Experimental results demonstrate that the proposed method outperforms conventional INR-based baselines in reconstruction quality, while maintaining a comparable model size.

eess.IV↗

arXiv2025

Structure-Preserving Patch Decoding for Efficient Neural Video Representation

Implicit neural representations (INRs) are the subject of extensive research, particularly in their application to modeling complex signals by mapping spatial and temporal coordinates to corresponding values. When handling videos, mapping compact inputs to entire frames or spatially partitioned patch images is an effective approach. This strategy better preserves spatial relationships, reduces computational overhead, and improves reconstruction quality compared to coordinate-based mapping. However, predicting entire frames often limits the reconstruction of high-frequency visual details. Additionally, conventional patch-based approaches based on uniform spatial partitioning tend to introduce boundary discontinuities that degrade spatial coherence. We propose a neural video representation method based on Structure-Preserving Patches (SPPs) to address such limitations. Our method separates each video frame into patch images of spatially aligned frames through a deterministic pixel-based splitting similar to PixelUnshuffle. This operation preserves the global spatial structure while allowing patch-level decoding. We train the decoder to reconstruct these structured patches, enabling a global-to-local decoding strategy that captures the global layout first and refines local details. This effectively reduces boundary artifacts and mitigates distortions from naive upsampling. Experiments on standard video datasets demonstrate that our method achieves higher reconstruction quality and better compression performance than existing INR-based baselines.

cs.CV↗

arXiv2025

Classification in Japanese Sign Language Based on Dynamic Facial Expressions

Sign language is a visual language expressed through hand movements and non-manual markers. Non-manual markers include facial expressions and head movements. These expressions vary across different nations. Therefore, specialized analysis methods for each sign language are necessary. However, research on Japanese Sign Language (JSL) recognition is limited due to a lack of datasets. The development of recognition models that consider both manual and non-manual features of JSL is crucial for precise and smooth communication with deaf individuals. In JSL, sentence types such as affirmative statements and questions are distinguished by facial expressions. In this paper, we propose a JSL recognition method that focuses on facial expressions. Our proposed method utilizes a neural network to analyze facial features and classify sentence types. Through the experiments, we confirm our method's effectiveness by achieving a classification accuracy of 96.05%.

cs.CV↗

arXiv2025

On equivalence of entropy and viscosity solutions to degenerate parabolic equations and applications

Here, we consider anisotropic degenerate parabolic-hyperbolic equations and degenerate quasilinear Hamilton-Jacobi equations. We prove the equivalence of two notions of entropy and viscosity solutions of two equations, and apply it to obtain a large-time behavior of viscosity solutions to quasilinear Hamilton-Jacobi equations, and entropy solutions to degenerate parabolic-hyperbolic equations in a periodic setting.

math.AP↗

arXiv2025

Guided Diffusion for the Extension of Machine Vision to Human Visual Perception

Image compression technology eliminates redundant information to enable efficient transmission and storage of images, serving both machine vision and human visual perception. For years, image coding focused on human perception has been well-studied, leading to the development of various image compression standards. On the other hand, with the rapid advancements in image recognition models, image compression for AI tasks, known as Image Coding for Machines (ICM), has gained significant importance. Therefore, scalable image coding techniques that address the needs of both machines and humans have become a key area of interest. Additionally, there is increasing demand for research applying the diffusion model, which can generate human-viewable images from a small amount of data to image compression methods for human vision. Image compression methods that use diffusion models can partially reconstruct the target image by guiding the generation process with a small amount of conditioning information. Inspired by the diffusion model's potential, we propose a method for extending machine vision to human visual perception using guided diffusion. Utilizing the diffusion model guided by the output of the ICM method, we generate images for human perception from random noise. Guided diffusion acts as a bridge between machine vision and human vision, enabling transitions between them without any additional bitrate overhead. The generated images then evaluated based on bitrate and image quality, and we compare their compression performance with other scalable image coding methods for humans and machines.

cs.CV↗

arXiv2025

Possibility of BCS-BEC crossover in $κ$-type organic superconductors

The realization of BCS-BEC crossover in superconductors, which smoothly connects Bardeen-Cooper-Schrieffer (BCS) theory with Bose-Einstein Condensation (BEC) in fermion systems, is an intriguing recent topic in strongly correlated electron systems. The organic superconductor $κ$-(BEDT-TTF)$_4$Hg$_{2.89}$Br$_8$ ($κ$-HgBr) under pressure is one of the leading candidates, owing to its unique metallic spin-liquid nature and tunable electron correlation. We theoretically investigate the extended Hubbard model for $κ$-HgBr and discuss the possibility of the BCS-BEC crossover by systematically calculating superconducting correlation function, coherence length, superfluid weight, and chemical potential. Our findings show that the BCS-BEC crossover can be observed when competing phases, such as Mott insulators and charge and/or spin orders, are suppressed by appropriate hole doping. $κ$-HgBr is just the case because both the Mott insulating phase and magnetic orders are absent due to its nonstoichiometric Hg composition and geometrical frustration. We further propose that other $κ$-type organic superconductors could serve as potential candidates of the BCS-BEC crossover if their band fillings and degree of geometrical frustration are systematically tuned.

cond-mat.supr-con↗

arXiv2025

Adapting Image-to-Video Diffusion Models for Large-Motion Frame Interpolation

With the development of video generation models has advanced significantly in recent years, we adopt large-scale image-to-video diffusion models for video frame interpolation. We present a conditional encoder designed to adapt an image-to-video model for large-motion frame interpolation. To enhance performance, we integrate a dual-branch feature extractor and propose a cross-frame attention mechanism that effectively captures both spatial and temporal information, enabling accurate interpolations of intermediate frames. Our approach demonstrates superior performance on the Fréchet Video Distance (FVD) metric when evaluated against other state-of-the-art approaches, particularly in handling large motion scenarios, highlighting advancements in generative-based methodologies.

cs.CV↗

arXiv2025

Structural Changes and Percolation Transition in Networks after Aging Processes

In social networking services, users constantly change, and the network structure changes simultaneously. As the network structure changes, so does the word-of-mouth within it. To study how information transfer on the network changes with the aging of the network, we investigated the relation of the structure and the percolation in the aged networks. We first prepared the Bianconi-Barabási model as the initial network and observed the time evolution of its properties by repeatedly deleting and adding nodes. We introduced two tunable parameters, the deleting parameter $α$ and the adding parameter $β$, and observed the aging behavior for the various parameter sets. We found that the network did not reach its steady state depending on the parameters. We also found that when the network is stable, there is a parameter region where the degree distribution changes from power-law to exponential decay. We performed the percolation analyses and found that the behavior of the percolation probability changes simultaneously with changes in the network's structure. This study is expected to help control network structure in steady-state aged networks.

physics.soc-ph↗

arXiv2025

Evolution and Instability of Bogoliubov Fermi Surfaces under Zeeman Field

We theoretically investigate the evolution and instability of the Bogoliubov Fermi surface (BFS) in the spherical $j=3/2$ model under a Zeeman field. The applied field induces a pronounced expansion in the BFS with $j_z = \pm 3/2$ component. Such behavior can be detected by spectroscopic techniques such as angle-resolved photoemission spectroscopy (ARPES). Interestingly, the Pauli susceptibility exhibits behavior that appears discontinuous just below the transition temperature at zero field, even though it is a second-order transition. This is due to spontaneous magnetization. Furthermore, the analysis of the bogolon correlations in the superconducting phase suggests the possibility of the chiral $p$- or $f$-wave bogolon pairing instabilities rather than the Pomeranchuk instability. These chiral states coexist with the chiral $d$-wave superconducting state, spontaneously break inversion symmetry, and lead to the disappearance of the torus-shaped BFS structure.

cond-mat.supr-con↗

arXiv2024

Time Step Generating: A Universal Synthesized Deepfake Image Detector

Currently, high-fidelity text-to-image models are developed in an accelerating pace. Among them, Diffusion Models have led to a remarkable improvement in the quality of image generation, making it vary challenging to distinguish between real and synthesized images. It simultaneously raises serious concerns regarding privacy and security. Some methods are proposed to distinguish the diffusion model generated images through reconstructing. However, the inversion and denoising processes are time-consuming and heavily reliant on the pre-trained generative model. Consequently, if the pre-trained generative model meet the problem of out-of-domain, the detection performance declines. To address this issue, we propose a universal synthetic image detector Time Step Generating (TSG), which does not rely on pre-trained models' reconstructing ability, specific datasets, or sampling algorithms. Our method utilizes a pre-trained diffusion model's network as a feature extractor to capture fine-grained details, focusing on the subtle differences between real and synthetic images. By controlling the time step t of the network input, we can effectively extract these distinguishing detail features. Then, those features can be passed through a classifier (i.e. Resnet), which efficiently detects whether an image is synthetic or real. We test the proposed TSG on the large-scale GenImage benchmark and it achieves significant improvements in both accuracy and generalizability.

cs.CV↗

arXiv2024

Problem of nonlinear conductivity within relaxation time approximation in noncentrosymmetric insulators

With the recent advancements in laser technology, there has been increasing interest in nonlinear and nonperturbative phenomena such as nonreciprocal transport, the nonlinear Hall effect, and nonlinear optical responses. When analyzing the nonequilibrium steady state, the relaxation time approximation (RTA) in the quantum kinetic equation has been widely used. However, recent studies have highlighted problems with the use of RTA that require careful consideration. In a study published in Phys. Rev. B, $\textbf{109}$, L180302 (2024), we revealed that the RTA has a flaw in predicting finite linear conductivity even for insulators under weak electric fields, and improved the RTA based on the Redfield equation. In this paper, we further extend our approach to nonlinear responses. This approach provides a simple alternative to RTA and is expected to be useful for the study of nonlinear and nonequilibrium phenomena.

cond-mat.mes-hall↗

arXiv2024

Multi-tunneling effect of nonreciprocal Landau-Zener tunneling: Insights from DC field responses

Recent advancements in laser technology have spurred growing interest in nonlinear and nonequilibrium phenomena. Here, we investigate the geometric aspects of quantum tunneling and the nonreciprocal response, particularly focusing on the shift vector, in noncentrosymmetric insulators under a strong DC electric field. In insulators under a strong electric field, electrons undergoing Bloch oscillations interfere with each other by passing through different paths via Landau-Zener tunneling. We found that the interference effect due to multi-tunneling causes the oscillating nonreciprocal response that is significantly amplified with increasing electric field intensity. We also clarified the role of the shift vector in the interference conditions through an analysis of the nonequilibrium steady state. These results will contribute significantly to advancing a systematic understanding of quantum geometric effects in the nonperturbative regime.

cond-mat.mes-hall↗

arXiv2024

Inter-Feature-Map Differential Coding of Surveillance Video

In Collaborative Intelligence, a deep neural network (DNN) is partitioned and deployed at the edge and the cloud for bandwidth saving and system optimization. When a model input is an image, it has been confirmed that the intermediate feature map, the output from the edge, can be smaller than the input data size. However, its effectiveness has not been reported when the input is a video. In this study, we propose a method to compress the feature map of surveillance videos by applying inter-feature-map differential coding (IFMDC). IFMDC shows a compression ratio comparable to, or better than, HEVC to the input video in the case of small accuracy reduction. Our method is especially effective for videos that are sensitive to image quality degradation when HEVC is applied

cs.CV↗

arXiv2024

Delta-ICM: Entropy Modeling with Delta Function for Learned Image Compression

Image Coding for Machines (ICM) is becoming more important as research in computer vision progresses. ICM is a vital research field that pursues the use of images for image recognition models, facilitating efficient image transmission and storage. The demand for recognition models is growing rapidly among the general public, and their performance continues to improve. To meet these needs, exchanging image data between consumer devices and cloud AI using ICM technology could be one possible solution. In ICM, various image compression methods have adopted Learned Image Compression (LIC). LIC includes an entropy model for estimating the bitrate of latent features, and the design of this model significantly affects its performance. Typically, LIC methods assume that the distribution of latent features follows a normal distribution. This assumption is effective for compressing images intended for human vision. However, employing an entropy model based on normal distribution is inefficient in ICM due to the limitation of image parts that require precise decoding. To address this, we propose Delta-ICM, which uses a probability distribution based on a delta function. Assuming the delta distribution as a distribution of latent features reduces the entropy of image portions unnecessary for machines. We compress the remaining portions using an entropy model based on normal distribution, similar to existing methods. Delta-ICM selects between the entropy model based on the delta distribution and the one based on the normal distribution for each latent feature. Our method outperforms existing ICM methods in image compression performance aimed at machines.

cs.CV↗

arXiv2024

Neural Video Representation for Redundancy Reduction and Consistency Preservation

Implicit neural representation (INR) embed various signals into neural networks. They have gained attention in recent years because of their versatility in handling diverse signal types. In the context of video, INR achieves video compression by embedding video signals directly into networks and compressing them. Conventional methods either use an index that expresses the time of the frame or features extracted from individual frames as network inputs. The latter method provides greater expressive capability as the input is specific to each video. However, the features extracted from frames often contain redundancy, which contradicts the purpose of video compression. Additionally, such redundancies make it challenging to accurately reconstruct high-frequency components in the frames. To address these problems, we focus on separating the high-frequency and low-frequency components of the reconstructed frame. We propose a video representation method that generates both the high-frequency and low-frequency components of the frame, using features extracted from the high-frequency components and temporal information, respectively. Experimental results demonstrate that our method outperforms the existing HNeRV method, achieving superior results in 96 percent of the videos.

cs.CV↗

arXiv2024

Photo-induced phase transition on black samarium monosulfide

To investigate the role of the excitons for the origin of the pressure-induced phase transition (BGT) from the black-colored insulator (BI) to the golden-yellow-colored metal (GM) of samarium monosulfide (SmS), optical reflectivity, Sm $3d$ X-ray absorption spectroscopy (XAS), and X-ray diffraction (XRD) with the creation of excitons by photoexcitation (PE) are reported. In the pump-probe reflectivity measurement, following a huge reflectivity change of about 22 %, three different relaxation times with a vibration component were observed. The fast component with the relaxation time ($τ$) of less than 1 ps is due to the excitation and relaxation of electrons into the conduction band, and the slowest one with $τ> {\rm several} 100$ ps originates from the appearance of the photo-induced (PI) state. The components with $τ\sim 10$ ps and vibration originate from the appearance of the PI state and the interference between the reflection lights at the sample surface and the boundary between the BI and PI states, suggesting that the electronic structure of the PI phase is different from that of the BI state. XAS spectra indicate that the Sm mean valence is shifted from the Sm$^{2+}$ dominant to the intermediate between Sm$^{2+}$ and Sm$^{3+}$ by PE but did not change to that of the GM phase across BGT, consistent with the reflectivity data. The XRD result after PE shows that the PI state has much less lattice contraction than the GM phase. These results suggest that the BGT cannot be achieved solely by creating excitons after PE but requires other effects, such as a lattice contraction.

cond-mat.str-el↗