Kousuke Imamura
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Kousuke Imamura has a total of 24 co-authors. Here are the top co-authors by number of shared publications:
Hideo Hashimoto
2 shared publicationsAhmed Swilem
1 shared publication- Akira Kadokura
National Institute of Polar Research Tachikawa Japan
1 shared publication - Dianbo Ma
Graduate School of Natural Science and Technology, Kanazawa University, Kanazawa 9201192, Japan
1 shared publication - Jiashu Zhang
Graduate School of Natural Science and Technology, Kanazawa University, Kakuma-Machi, Kanazawa 920-1192, Ishikawa, Japan
1 shared publication - Kazuo Shiokawa
Institute for Space‐Earth Environmental Research Nagoya University Nagoya Japan
1 shared publication
Latest Publications
- Open Access
- 24 September 2026
- Published by MDPI AG in Electronics
- Vol. 15 (19) , 4394
Multimodal Deep Learning with Automated Myocardial Segmentation for Classification of Myocardial 99mTc-Pyrophosphate Uptake Using Planar and SPECT Imaging
Objective interpretation of myocardial 99mTc-pyrophosphate (PYP) uptake is important in suspected cardiac amyloidosis. We developed a segmentation-informed multimodal framework that classifies myocardial PYP uptake from planar and single-photon emission computed tomography (SPECT) images. The retrospective, class-enriched cohort comprised 49 studies from 44 patients (19 PYP-positive and 25 PYP-negative) with specialist-confirmed image labels. Five consecutive SPECT slices were manually selected under nuclear-medicine-physician supervision; subsequent processing was automated. SPECT Net generated an unthresholded myocardial probability map, which served as a soft, spatially interpretable representation for integration with planar features. Patient-grouped five-fold cross-validation used separate validation and test folds. In PYP-positive patients, cropped nnU-Net achieved a patient-level Dice coefficient of and outperformed three segmentation comparators after Holm correction (). Transformer-style fusion achieved 95.5% accuracy (95% confidence interval: 84.9–98.7%), 89.5% sensitivity, 100.0% specificity, and a 94.4% F1-score. Improvements over planar-only classification (90.9%; McNemar ), localized raw SPECT (86.4%; ), and H/CL assessment (88.6%; ) were not significant. These single-center internal results demonstrate the feasibility of using a soft myocardial segmentation output as an interpretable bridge for multimodal PYP uptake classification but do not establish superiority; multicenter external validation is required.- 24 September 2026
- Published by MDPI AG in Electronics
- Vol. 15 (19) , 4394
Open AccessMultimodal Deep Learning with Automated Myocardial Segmentation for Classification of Myocardial 99mTc-Pyrophosphate Uptake Using Planar and SPECT Imaging
Objective interpretation of myocardial 99mTc-pyrophosphate (PYP) uptake is important in suspected cardiac amyloidosis. We developed a segmentation-informed multimodal framework that classifies myocardial PYP uptake from planar and single-photon emission computed tomography (SPECT) images. The retrospective, class-enriched cohort comprised 49 studies from 44 patients (19 PYP-positive and 25 PYP-negative) with specialist-confirmed image labels. Five consecutive SPECT slices were manually selected under nuclear-medicine-physician supervision; subsequent processing was automated. SPECT Net generated an unthresholded myocardial probability map, which served as a soft, spatially interpretable representation for integration with planar features. Patient-grouped five-fold cross-validation used separate validation and test folds. In PYP-positive patients, cropped nnU-Net achieved a patient-level Dice coefficient of and outperformed three segmentation comparators after Holm correction (). Transformer-style fusion achieved 95.5% accuracy (95% confidence interval: 84.9–98.7%), 89.5% sensitivity, 100.0% specificity, and a 94.4% F1-score. Improvements over planar-only classification (90.9%; McNemar ), localized raw SPECT (86.4%; ), and H/CL assessment (88.6%; ) were not significant. These single-center internal results demonstrate the feasibility of using a soft myocardial segmentation output as an interpretable bridge for multimodal PYP uptake classification but do not establish superiority; multicenter external validation is required. - Open Access
Hierarchical Motion Field Alignment for Robust Optical Flow Estimation
Optical flow estimation is a fundamental and long-standing task in computer vision, facilitating the understanding of motion within visual scenes. In this study, we aim to improve optical flow estimation, particularly in challenging scenarios involving small and fast-moving objects. Specifically, we proposed a learning-based model incorporating two key components: the Hierarchical Motion Field Alignment module, which ensures accurate estimation of objects of varying sizes while maintaining manageable computational complexity, and the Correlation Self-Attention module, which effectively handles large displacements, making the model suitable for scenarios with fast-moving objects. Additionally, we introduced a Multi-Scale Correlation Search layer to enhance the four-dimensional cost volume, enabling the model to address various types of motion. Experimental results demonstrate that our model achieves superior generalization performance and significantly improves the estimation of small, fast-moving objects.
Open AccessHierarchical Motion Field Alignment for Robust Optical Flow Estimation
Optical flow estimation is a fundamental and long-standing task in computer vision, facilitating the understanding of motion within visual scenes. In this study, we aim to improve optical flow estimation, particularly in challenging scenarios involving small and fast-moving objects. Specifically, we proposed a learning-based model incorporating two key components: the Hierarchical Motion Field Alignment module, which ensures accurate estimation of objects of varying sizes while maintaining manageable computational complexity, and the Correlation Self-Attention module, which effectively handles large displacements, making the model suitable for scenarios with fast-moving objects. Additionally, we introduced a Multi-Scale Correlation Search layer to enhance the four-dimensional cost volume, enabling the model to address various types of motion. Experimental results demonstrate that our model achieves superior generalization performance and significantly improves the estimation of small, fast-moving objects.
- Open Access
- 6 July 2021
- Published by American Geophysical Union (AGU) in Journal of Geophysical Research: Space Physics
- Vol. 126 (7)
Spatial Evolution of Wave‐Particle Interaction Region Deduced From Flash‐Type Auroras and Chorus‐Ray Tracing
In‐situ observations of spatial variations of the wave‐particle interaction region require a large number of satellite probes. As an alternative, flash‐type auroras, a kind of pulsating aurora, driven by discrete chorus elements, can be used to investigate the interaction region with a high spatial resolution. We estimated the spatial extent of wave‐particle interaction region from ground‐based observations of flash aurora at Gakona (62.39°N, 214.78°E), Alaska at subauroral latitudes, and found that the auroral expansion was predominantly to the low‐latitude side. The spatial displacement is thought to be caused by the propagation effects of chorus waves in the magnetosphere. Using ray tracing analysis to take into account chorus wave propagation, we reconstructed the spatiotemporal evolution of the volume emission rate and confirmed that the predominant expansion is toward the lower‐latitude side in the ionosphere. This study shows that chorus wave propagation in the magnetosphere gives new insight for characterizing the transverse size (across the geomagnetic field line) of wave‐particle interaction regions. The calculated spatial scale of the column auroral emission shows a correlation with the magnetic latitude of the resonance region at magnetic latitudes within 10° of the equatorial plane of the magnetosphere. Our results suggest that the spatial scale of a flash aurora is indirectly related to the chorus amplitude because the latitudinal range of the wave‐particle interaction is important for the growth of wave amplitude.
- 6 July 2021
- Published by American Geophysical Union (AGU) in Journal of Geophysical Research: Space Physics
- Vol. 126 (7)
Open AccessSpatial Evolution of Wave‐Particle Interaction Region Deduced From Flash‐Type Auroras and Chorus‐Ray Tracing
In‐situ observations of spatial variations of the wave‐particle interaction region require a large number of satellite probes. As an alternative, flash‐type auroras, a kind of pulsating aurora, driven by discrete chorus elements, can be used to investigate the interaction region with a high spatial resolution. We estimated the spatial extent of wave‐particle interaction region from ground‐based observations of flash aurora at Gakona (62.39°N, 214.78°E), Alaska at subauroral latitudes, and found that the auroral expansion was predominantly to the low‐latitude side. The spatial displacement is thought to be caused by the propagation effects of chorus waves in the magnetosphere. Using ray tracing analysis to take into account chorus wave propagation, we reconstructed the spatiotemporal evolution of the volume emission rate and confirmed that the predominant expansion is toward the lower‐latitude side in the ionosphere. This study shows that chorus wave propagation in the magnetosphere gives new insight for characterizing the transverse size (across the geomagnetic field line) of wave‐particle interaction regions. The calculated spatial scale of the column auroral emission shows a correlation with the magnetic latitude of the resonance region at magnetic latitudes within 10° of the equatorial plane of the magnetosphere. Our results suggest that the spatial scale of a flash aurora is indirectly related to the chorus amplitude because the latitudinal range of the wave‐particle interaction is important for the growth of wave amplitude.
- 1 January 2007
- Published by Institute of Image Information and Television Engineers in The Journal of The Institute of Image Information and Television Engineers
- Vol. 61 (8) , 1231-1235
Fast Matching Pursuits Using Multiple-Step Serach in Non-Low Signal Energy Regions
- 1 January 2007
- Published by Institute of Image Information and Television Engineers in The Journal of The Institute of Image Information and Television Engineers
- Vol. 61 (8) , 1231-1235
Fast Matching Pursuits Using Multiple-Step Serach in Non-Low Signal Energy Regions
- 1 January 2002
- Published by Institute of Image Information and Television Engineers in The Journal of The Institute of Image Information and Television Engineers
- Vol. 56 (9) , 1513-1517
A Fast Search Algorithm for Vector Quantization Using Hyperplane Decision Rule.
- 1 January 2002
- Published by Institute of Image Information and Television Engineers in The Journal of The Institute of Image Information and Television Engineers
- Vol. 56 (9) , 1513-1517
A Fast Search Algorithm for Vector Quantization Using Hyperplane Decision Rule.