Kenji Ishikawa
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Kenji Ishikawa has a total of 42 co-authors. Here are the top co-authors by number of shared publications:
Yasuhiro Oikawa
15 shared publications- Kohei Yatabe
Tokyo University of Agriculture and Technology
14 shared publications - Yasuhiro Oikawa
Department of Intermedia Art and Science, Waseda University 2 , 3-4-1 Ohkubo, Shinjuku-ku, Tokyo 169-8555
10 shared publications - Noboru Harada
NTT Corporation Japan
7 shared publications - Risako Tanigawa
Communication Science Laboratories, NTT, Inc 1 ., 3-1 Morinosato-Wakamiya, Atsugi, Kanagawa 243-0198
6 shared publications Takashi Onuma
Photron Limited, 1-105 Kanda-Jinbocho, Chiyoda-ku, Tokyo 101-0051, Japan
5 shared publications
Latest Publications
- 1 May 2026
- Published by Acoustical Society of America (ASA) in The Journal of the Acoustical Society of America
- Vol. 159 (5) , 3858-3869
Sound-field-projection synthesis using latent diffusion model for acousto-optic reconstruction
Acousto-optic sensing (AOS) is a non-contact method for measuring sound using light, suitable for environments where microphones are impractical, such as confined spaces or within airflow. Despite its effectiveness, AOS captures line-integrated sound pressure along the optical path, resulting in signals that cannot be interpreted as those from point-wise microphones. To obtain the sound pressure distribution in three-dimensional (3D) space, volumetric sound-field reconstruction is required. Existing methods require multi-directional line-integrated projection data, typically obtained either by deploying multiple devices or by repeatedly exciting the sound source. The former requires deploying multiple high-cost devices, which is often impractical, whereas the latter is not applicable to sound sources that cannot be reproduced consistently. To overcome these limitations, we propose a task called sound projection synthesis, which synthesizes data in specified directions based on observation data. We achieve this by using a latent diffusion model conditioned on observed projections and view angles to generate new sound-field projections. A model pretrained on natural images was fine-tuned using sound-field data and optimized with a pixel-wise loss. Experiments demonstrate that the model can generate realistic projection data. Combining nine observed views with nine generated views improved 3D reconstruction accuracy compared with using only nine observed views.
- 1 May 2026
- Published by Acoustical Society of America (ASA) in The Journal of the Acoustical Society of America
- Vol. 159 (5) , 3858-3869
Sound-field-projection synthesis using latent diffusion model for acousto-optic reconstruction
Acousto-optic sensing (AOS) is a non-contact method for measuring sound using light, suitable for environments where microphones are impractical, such as confined spaces or within airflow. Despite its effectiveness, AOS captures line-integrated sound pressure along the optical path, resulting in signals that cannot be interpreted as those from point-wise microphones. To obtain the sound pressure distribution in three-dimensional (3D) space, volumetric sound-field reconstruction is required. Existing methods require multi-directional line-integrated projection data, typically obtained either by deploying multiple devices or by repeatedly exciting the sound source. The former requires deploying multiple high-cost devices, which is often impractical, whereas the latter is not applicable to sound sources that cannot be reproduced consistently. To overcome these limitations, we propose a task called sound projection synthesis, which synthesizes data in specified directions based on observation data. We achieve this by using a latent diffusion model conditioned on observed projections and view angles to generate new sound-field projections. A model pretrained on natural images was fine-tuned using sound-field data and optimized with a pixel-wise loss. Experiments demonstrate that the model can generate realistic projection data. Combining nine observed views with nine generated views improved 3D reconstruction accuracy compared with using only nine observed views.
- Open Access
- 1 May 2026
- Published by Acoustical Society of America (ASA) in JASA Express Letters
- Vol. 6 (5)
Differentiable equivalent source method: Joint optimization of equivalent source positions and strengths with implicit geometry constraints
The equivalent source method (ESM) is widely used for sound field reconstruction due to its flexibility. However, optimizing source configurations remains challenging because ESM requires all sources to be located outside the reconstruction region. Identifying optimal source positions while satisfying these geometric constraints is an underexplored challenge. To address this, we propose a differentiable equivalent source method (DiffESM) for joint optimization of source positions and strengths. By employing a neural signed distance function to represent feasible regions of the source positions, DiffESM enables gradient-based optimization while enforcing geometric constraints. Preliminary results demonstrated that DiffESM achieved higher reconstruction accuracy than conventional ESMs.
- 1 May 2026
- Published by Acoustical Society of America (ASA) in JASA Express Letters
- Vol. 6 (5)
Open AccessDifferentiable equivalent source method: Joint optimization of equivalent source positions and strengths with implicit geometry constraints
The equivalent source method (ESM) is widely used for sound field reconstruction due to its flexibility. However, optimizing source configurations remains challenging because ESM requires all sources to be located outside the reconstruction region. Identifying optimal source positions while satisfying these geometric constraints is an underexplored challenge. To address this, we propose a differentiable equivalent source method (DiffESM) for joint optimization of source positions and strengths. By employing a neural signed distance function to represent feasible regions of the source positions, DiffESM enables gradient-based optimization while enforcing geometric constraints. Preliminary results demonstrated that DiffESM achieved higher reconstruction accuracy than conventional ESMs.
- 7 November 2025
- Published by Optica Publishing Group in Optics Express
- Vol. 33 (23) , 48403-48422
Temporal super-resolution with a latent diffusion model for optically measured sound field
This paper proposes a latent-diffusion-model-based method to achieve temporal super-resolution in optical sound-field imaging. Optical sound-field imaging, known for its high spatial resolution, measures sound by detecting small variations in the refractive index of air caused by sound. Due to the extremely high speed of sound propagation, optical imaging systems require exceptionally high frame rates for sound-field measurement, which results in high power consumption as well as extremely time-consuming data transmission and processing during long-duration data acquisition. Therefore, to alleviate the sampling requirements of high-speed cameras, mitigate data transmission and processing demands, reduce imaging power consumption, and enhance the temporal quality of imaging, we propose a latent-diffusion-model-based method to achieve temporal super-resolution. During inference, the measured sound-field images are utilized as conditional constraints, enabling the model to generate intermediate sound-field frames between any two consecutive frames in the temporal sequence. Numerical experiments demonstrate that our method outperforms other deep learning approaches in the temporal super-resolution task on optically measured sound-field data. The experimental results demonstrate the applicability of our model to the real world.
- 7 November 2025
- Published by Optica Publishing Group in Optics Express
- Vol. 33 (23) , 48403-48422
Temporal super-resolution with a latent diffusion model for optically measured sound field
This paper proposes a latent-diffusion-model-based method to achieve temporal super-resolution in optical sound-field imaging. Optical sound-field imaging, known for its high spatial resolution, measures sound by detecting small variations in the refractive index of air caused by sound. Due to the extremely high speed of sound propagation, optical imaging systems require exceptionally high frame rates for sound-field measurement, which results in high power consumption as well as extremely time-consuming data transmission and processing during long-duration data acquisition. Therefore, to alleviate the sampling requirements of high-speed cameras, mitigate data transmission and processing demands, reduce imaging power consumption, and enhance the temporal quality of imaging, we propose a latent-diffusion-model-based method to achieve temporal super-resolution. During inference, the measured sound-field images are utilized as conditional constraints, enabling the model to generate intermediate sound-field frames between any two consecutive frames in the temporal sequence. Numerical experiments demonstrate that our method outperforms other deep learning approaches in the temporal super-resolution task on optically measured sound-field data. The experimental results demonstrate the applicability of our model to the real world.
- Open Access
- 1 September 2025
- Published by Acoustical Society of America (ASA) in The Journal of the Acoustical Society of America
- Vol. 158 (3) , 2163-2175
Data-driven volumetric reconstruction for optically measured sound field using physics-constrained 3D Gaussian splatting
Acousto-optic sensing is a powerful approach to measuring sound at a high resolution; yet, it faces a critical challenge because the measured value is a line integral of the sound. To solve this problem, sound-field reconstruction methods have been proposed. Promising approaches include physical-model-based reconstruction methods, which represent a sound field as a linear combination of basis functions and determine the expansion coefficients. However, they are limited by the choice of basis functions, which means that each model has a suitable sound field, making it difficult to apply a single model to all sound fields. In this paper, a data-driven approach that is applicable to high-complexity sound fields is proposed. A 3D Gaussian splatting (3DGS) scheme for three-dimensional (3D) sound-field reconstruction is leveraged. 3DGS is an advanced and cutting-edge approach in computer vision, which represents a 3D scene as the sum of Gaussian kernels placed in 3D space. In the proposed method, the 3DGS-based volume reconstruction approach, R2-Gaussian, is expanded to handle arbitrary real numbers to represent sound fields and introduces a Helmholtz loss in the optimization. Evaluation experiments were performed with 11 simulated sound fields and 1 measured sound field. The experiments have revealed that the 3DGS-based approach can reconstruct sound fields.
- 1 September 2025
- Published by Acoustical Society of America (ASA) in The Journal of the Acoustical Society of America
- Vol. 158 (3) , 2163-2175
Open AccessData-driven volumetric reconstruction for optically measured sound field using physics-constrained 3D Gaussian splatting
Acousto-optic sensing is a powerful approach to measuring sound at a high resolution; yet, it faces a critical challenge because the measured value is a line integral of the sound. To solve this problem, sound-field reconstruction methods have been proposed. Promising approaches include physical-model-based reconstruction methods, which represent a sound field as a linear combination of basis functions and determine the expansion coefficients. However, they are limited by the choice of basis functions, which means that each model has a suitable sound field, making it difficult to apply a single model to all sound fields. In this paper, a data-driven approach that is applicable to high-complexity sound fields is proposed. A 3D Gaussian splatting (3DGS) scheme for three-dimensional (3D) sound-field reconstruction is leveraged. 3DGS is an advanced and cutting-edge approach in computer vision, which represents a 3D scene as the sum of Gaussian kernels placed in 3D space. In the proposed method, the 3DGS-based volume reconstruction approach, R2-Gaussian, is expanded to handle arbitrary real numbers to represent sound fields and introduces a Helmholtz loss in the optimization. Evaluation experiments were performed with 11 simulated sound fields and 1 measured sound field. The experiments have revealed that the 3DGS-based approach can reconstruct sound fields.
- 1 September 2025
- Published by Acoustical Society of America (ASA) in The Journal of the Acoustical Society of America
- Vol. 158 (3) , 1688-1699
Spherical harmonic-domain acousto-optic tomography for three-dimensional sound-field reconstruction
Acousto-optic tomography (AOT), a technology that reconstructs two- or three-dimensional sound fields from optically measured sound-field projections, has been widely studied as an efficient and high-spatial-resolution method for sound field observations. Recently, physical-model-based approaches have made significant progress, with higher accuracy and fewer sampling requirements than conventional methods. Nevertheless, it remains a challenge to reconstruct three-dimensional outgoing sound waves in the volume surrounding a sound source due to constraints on existing methods both in mathematical formulation and measurement systems. In this paper, we propose spherical harmonic-domain AOT, a three-dimensional reconstruction of sound fields by combining instantaneous sound-field imaging using parallel phase-shifting interferometry with a physics-based reconstruction using spherical wave function expansions. We formulated two problems: one for the sound field radiated by a transducer and the other for the sound field scattered by an object. Through experiments, we demonstrated the successful reconstruction of the two types of three-dimensional sound fields: one radiated from two ultrasonic transducers and the other scattered by an ellipsoid.
- 1 September 2025
- Published by Acoustical Society of America (ASA) in The Journal of the Acoustical Society of America
- Vol. 158 (3) , 1688-1699
Spherical harmonic-domain acousto-optic tomography for three-dimensional sound-field reconstruction
Acousto-optic tomography (AOT), a technology that reconstructs two- or three-dimensional sound fields from optically measured sound-field projections, has been widely studied as an efficient and high-spatial-resolution method for sound field observations. Recently, physical-model-based approaches have made significant progress, with higher accuracy and fewer sampling requirements than conventional methods. Nevertheless, it remains a challenge to reconstruct three-dimensional outgoing sound waves in the volume surrounding a sound source due to constraints on existing methods both in mathematical formulation and measurement systems. In this paper, we propose spherical harmonic-domain AOT, a three-dimensional reconstruction of sound fields by combining instantaneous sound-field imaging using parallel phase-shifting interferometry with a physics-based reconstruction using spherical wave function expansions. We formulated two problems: one for the sound field radiated by a transducer and the other for the sound field scattered by an object. Through experiments, we demonstrated the successful reconstruction of the two types of three-dimensional sound fields: one radiated from two ultrasonic transducers and the other scattered by an ellipsoid.
- 9 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-12
Acousto-Optic Reconstruction of Exterior Sound Field Based on Concentric Circle Sampling With Circular Harmonic Expansion
Acousto-optic sensing provides an alternative approach to traditional microphone arrays by shedding light on the interaction of light with an acoustic field. Sound field reconstruction is a fascinating and advanced technique used in acousto-optics sensing. Current challenges in sound-field reconstruction methods pertain to scenarios in which the sound source is located within the reconstruction area, known as the exterior problem. Existing reconstruction algorithms, primarily designed for interior scenarios, often exhibit suboptimal performance when applied to exterior cases. This paper introduces a novel technique for exterior sound-field reconstruction. The proposed method leverages concentric circle sampling and a two-dimensional exterior sound-field reconstruction approach based on circular harmonic extensions. To evaluate the efficacy of this approach, both numerical simulations and practical experiments are conducted. The results highlight the superior accuracy of the proposed method when compared to conventional reconstruction methods, all while utilizing a minimal amount of measured projection data.
- 9 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-12
Acousto-Optic Reconstruction of Exterior Sound Field Based on Concentric Circle Sampling With Circular Harmonic Expansion
Acousto-optic sensing provides an alternative approach to traditional microphone arrays by shedding light on the interaction of light with an acoustic field. Sound field reconstruction is a fascinating and advanced technique used in acousto-optics sensing. Current challenges in sound-field reconstruction methods pertain to scenarios in which the sound source is located within the reconstruction area, known as the exterior problem. Existing reconstruction algorithms, primarily designed for interior scenarios, often exhibit suboptimal performance when applied to exterior cases. This paper introduces a novel technique for exterior sound-field reconstruction. The proposed method leverages concentric circle sampling and a two-dimensional exterior sound-field reconstruction approach based on circular harmonic extensions. To evaluate the efficacy of this approach, both numerical simulations and practical experiments are conducted. The results highlight the superior accuracy of the proposed method when compared to conventional reconstruction methods, all while utilizing a minimal amount of measured projection data.
- Open Access
- 1 June 2025
- Published by Acoustical Society of America (ASA) in JASA Express Letters
- Vol. 5 (6)
Three-dimensional sound field reconstruction from optical projections using physics-informed neural networks
The implicit representation by physics-informed neural networks (PINNs) serves as an effective solution for a key challenge faced by optical sound measurements. Since optical sound measurements observe line integral of the sound pressure along the optical path, reconstruction is necessary to determine the sound pressure at each point in the three-dimensional field. In this paper, we expand the PINNs-based reconstruction method into three-dimensional reconstruction and demonstrate its effectiveness for optically measured sound fields. Furthermore, we propose a reconstruction approach which can estimate solutions well outside the bounds of the data used for training.
- 1 June 2025
- Published by Acoustical Society of America (ASA) in JASA Express Letters
- Vol. 5 (6)
Open AccessThree-dimensional sound field reconstruction from optical projections using physics-informed neural networks
The implicit representation by physics-informed neural networks (PINNs) serves as an effective solution for a key challenge faced by optical sound measurements. Since optical sound measurements observe line integral of the sound pressure along the optical path, reconstruction is necessary to determine the sound pressure at each point in the three-dimensional field. In this paper, we expand the PINNs-based reconstruction method into three-dimensional reconstruction and demonstrate its effectiveness for optically measured sound fields. Furthermore, we propose a reconstruction approach which can estimate solutions well outside the bounds of the data used for training.
- Open Access
- 1 May 2025
- Published by Acoustical Society of America (ASA) in JASA Express Letters
- Vol. 5 (5)
How the shape of the musical triangle influences its sound
Musical triangles, known for their triangular shape with one open end, have fascinated people with their twinkling sound. However, the acoustic factors behind the triangular shape remain unexplained. We discovered that their triangular shape induces a phenomenon conducive to sustained sound. We measured two-dimensional sound fields around musical triangles using an acousto-optic imaging method. Through our analysis, we found that acoustic resonance occurs in the area of semi-open triangular air created by the musical triangle. Consequently, the resonance produces a louder and longer sound. We believe that the resonances are the acoustical reason for the triangular shape.
- 1 May 2025
- Published by Acoustical Society of America (ASA) in JASA Express Letters
- Vol. 5 (5)
Open AccessHow the shape of the musical triangle influences its sound
Musical triangles, known for their triangular shape with one open end, have fascinated people with their twinkling sound. However, the acoustic factors behind the triangular shape remain unexplained. We discovered that their triangular shape induces a phenomenon conducive to sustained sound. We measured two-dimensional sound fields around musical triangles using an acousto-optic imaging method. Through our analysis, we found that acoustic resonance occurs in the area of semi-open triangular air created by the musical triangle. Consequently, the resonance produces a louder and longer sound. We believe that the resonances are the acoustical reason for the triangular shape.
- Open Access
SoundSil-DS: Deep Denoising and Segmentation of Sound-field Images with Silhouettes
Development of optical technology has enabled imaging of two-dimensional (2D) sound fields. This acousto-optic sensing enables understanding of the interaction between sound and objects such as reflection and diffraction. Moreover, it is expected to be used an advanced measurement technology for sonars in self-driving vehicles and assistive robots. However, the low sound-pressure sensitivity of the acousto-optic sensing results in high intensity of noise on images. Therefore, denoising is an essential task to visualize and analyze the sound fields. In addition to denoising, segmentation of sound and object silhouette is also required to analyze interactions between them. In this paper, we propose sound-field-images-with-object-silhouette denoising and segmentation (SoundSil-DS) that jointly perform denoising and segmentation for sound fields and object silhouettes on a visualized image. We developed a new model based on the current state-of-the-art denoising network. We also created a dataset to train and evaluate the proposed method through acoustic simulation. The proposed method was evaluated using both simulated and measured data. We confirmed that our method can applied to experimentally measured data. These results suggest that the proposed method may improve the post-processing for sound fields, such as physical model-based three-dimensional reconstruction since it can remove unwanted noise and separate sound fields and other object silhouettes. Our code is available at https://github.com/nttcslab/soundsil-ds.
Open AccessSoundSil-DS: Deep Denoising and Segmentation of Sound-field Images with Silhouettes
Development of optical technology has enabled imaging of two-dimensional (2D) sound fields. This acousto-optic sensing enables understanding of the interaction between sound and objects such as reflection and diffraction. Moreover, it is expected to be used an advanced measurement technology for sonars in self-driving vehicles and assistive robots. However, the low sound-pressure sensitivity of the acousto-optic sensing results in high intensity of noise on images. Therefore, denoising is an essential task to visualize and analyze the sound fields. In addition to denoising, segmentation of sound and object silhouette is also required to analyze interactions between them. In this paper, we propose sound-field-images-with-object-silhouette denoising and segmentation (SoundSil-DS) that jointly perform denoising and segmentation for sound fields and object silhouettes on a visualized image. We developed a new model based on the current state-of-the-art denoising network. We also created a dataset to train and evaluate the proposed method through acoustic simulation. The proposed method was evaluated using both simulated and measured data. We confirmed that our method can applied to experimentally measured data. These results suggest that the proposed method may improve the post-processing for sound fields, such as physical model-based three-dimensional reconstruction since it can remove unwanted noise and separate sound fields and other object silhouettes. Our code is available at https://github.com/nttcslab/soundsil-ds.
- 25 October 2024
- Published by Optica Publishing Group in Optics Express
- Vol. 32 (23) , 40898-40914
Diffusion-model-based inverse problem processing for optically-measured sound field
This paper proposes a diffusion-model-based method for addressing inverse problems in optical sound-field imaging. Optical sound-field imaging, known for its high spatial resolution, measures sound by detecting small variations in the refractive index of air caused by sound but often suffers from unavoidable noise contamination. Therefore, we present a diffusion model-based approach for sound-field inverse problems, including denoising, noisy sound-field reconstruction and extrapolation. During inference, sound-field degradation is introduced into the inverse denoising process, with range-null space decomposition used as a solver to handle degradation, iteratively generating degraded sound-field information. Numerical experiments show that our method outperforms other deep-learning-based methods in denoising and reconstruction tasks, and obtains effective results in extrapolation task. The experimental results demonstrate the applicability of our model to the real world.
- 25 October 2024
- Published by Optica Publishing Group in Optics Express
- Vol. 32 (23) , 40898-40914
Diffusion-model-based inverse problem processing for optically-measured sound field
This paper proposes a diffusion-model-based method for addressing inverse problems in optical sound-field imaging. Optical sound-field imaging, known for its high spatial resolution, measures sound by detecting small variations in the refractive index of air caused by sound but often suffers from unavoidable noise contamination. Therefore, we present a diffusion model-based approach for sound-field inverse problems, including denoising, noisy sound-field reconstruction and extrapolation. During inference, sound-field degradation is introduced into the inverse denoising process, with range-null space decomposition used as a solver to handle degradation, iteratively generating degraded sound-field information. Numerical experiments show that our method outperforms other deep-learning-based methods in denoising and reconstruction tasks, and obtains effective results in extrapolation task. The experimental results demonstrate the applicability of our model to the real world.