https://scholar.dgist.ac.kr/handle/20.500.11750/11753 2026-04-04T11:17:45Z 2026-04-04T11:17:45Z Jung, Dongkyu Guezzi, Nizar Lee, Sangheon Noman, Muhammad Bae, Sua Yu, Jaesok https://scholar.dgist.ac.kr/handle/20.500.11750/59948 2026-02-08T16:10:17Z

Title: RFDNet: Robust Frequency-Based Denoising Network for 3D Ultrasound Vascular Imaging Using a Row-Column Addressed Array Author(s): Jung, Dongkyu; Guezzi, Nizar; Lee, Sangheon; Noman, Muhammad; Bae, Sua; Yu, Jaesok Abstract: Three-dimensional (3D) ultrasound vascular imaging (UVI) is essential for visualizing complex vascular structures. Row-column addressed (RCA) arrays, widely used for 3D UVI due to their hardware efficiency, suffer from point spread function (PSF) anisotropy, resulting in ramp-shaped noise that degrades image quality. Although existing denoising methods, including deep learning-based approaches, have shown promise, they are often limited by domain shift bias and the need for condition-specific data collection. Moreover, as full-volume 3D training is often impractical, many studies rely on 2D slice-wise training with 3D reconstruction, which can yield inter-slice intensity inconsistencies when slices are normalized independently. To overcome these limitations, we propose Robust Frequency-based Denoising Network (RFDNet), which integrates a Deep Frequency Filtering (DFF) module into a standard denoising model. The DFF module adaptively filters frequency components within the encoder, suppressing ramp-shaped noise while dynamically balancing spectral content to reduce sensitivity to domain shifts and inter-slice intensity inconsistencies. This adaptive filtering preserves vascular details and improves overall imaging consistency. Experiments on Doppler phantom, carotid artery, and abdominal datasets show that RFDNet significantly outperforms conventional methods in peak signal-to-noise ratio (PSNR), structural similarity (SSIM) and root mean squared error (RMSE). Further validation through 2D frequency spectrum analysis confirmed that the DFF module dynamically adjusts frequency components to maintain spectral balance. In addition, spectral KL divergence analysis demonstrated its robustness against inter-slice intensity inconsistencies introduced by slice-wise normalization. This approach improves domain generalization, reduces noise artifacts, and enhances clinical applicability by improving imaging reliability. Future work will explore 3D training and architectural refinements for better computational efficiency.

Lee, Sangheon Jung, Dongkyu Guezzi Nizar Nam, Sangwoo Yu, Jaesok https://scholar.dgist.ac.kr/handle/20.500.11750/59949 2026-02-08T16:10:18Z

Title: DAUS-Net: Toward Ultrasound Scanner-Agnostic Domain Generalized Robust and Accurate Segmentation Author(s): Lee, Sangheon; Jung, Dongkyu; Guezzi Nizar; Nam, Sangwoo; Yu, Jaesok Abstract: In medical imaging, segmentation is a critical task for analysis and diagnosis. Deep learning-based segmentation has been actively studied and has shown remarkable performance. Building high-accuracy segmentation models requires a large amount of high-quality labeled data, but the cost of collecting such data is extremely high in medical imaging. In ultrasound imaging, the differences in image features depending on the equipment are significantly greater compared to other medical imaging modalities. Consequently, models need to be trained for each specific device, which entails substantial costs and time, leading to various practical challenges. To address these challenges, we propose a robust and accurate segmentation network that can operate independently of the ultrasound equipment. We integrated the Deep Frequency Filtering (DFF) module into a U-Net-based model. The proposed model retains the U-Net's encoder-decoder structure but applies frequency filtering within the latent space of each encoder layer, enabling adaptive selection of frequency components for breast tumor detection. Moreover, batch normalization was replaced with instance normalization to remove stylish features. We evaluated the model using three public datasets acquired from different scanners, achieving superior performance on unseen testing datasets compared to existing models. Notably, when tested on the unseen BUS-BRA dataset, DAUS-Net achieved a Dice score of 0.76, compared to 0.61 by the conventional U-Net. This improvement is attributed to the synergy between the DFF module and instance normalization. Our results demonstrate that the proposed model consistently detects and segments breast tumors, highlighting its potential for generalized clinical segmentation task. The source code for implementing DAUS-Net is publicly available at https://github.com/shlee8638/DAUS-Net.

Kaja, Kushal Ruthvik Hajra, Sugato Panda, Swati Belal, Mohamed Ahmed Nam, Sangwoo Pakawanit, Phakkhananan Panigrahi, Basanta Kumar Khanbareh, Hamideh Bowen, Chris Rhys Yu, Jaesok https://scholar.dgist.ac.kr/handle/20.500.11750/59899 2026-02-03T18:01:15Z 2025-11-30T15:00:00Z

Title: Waste polyethylene-coated fabrics for dual-mode interfaces triboelectrification for self-powered sensors Author(s): Kaja, Kushal Ruthvik; Hajra, Sugato; Panda, Swati; Belal, Mohamed Ahmed; Nam, Sangwoo; Pakawanit, Phakkhananan; Panigrahi, Basanta Kumar; Khanbareh, Hamideh; Bowen, Chris Rhys; Yu, Jaesok Abstract: The reuse of waste cotton textiles and coating them with recycled polyethylene provides a route to improve environmental sustainability, reducing our landfill burden and supporting the circular economy through effective implementation of the 3Rs: Reduce, Reuse, and Recycle. This approach extends material life and minimizes resource consumption. This study presents a sustainable strategy for energy harvesting and sensor applications by repurposing worn-out cotton textiles that are coated with recycled polyethylene via a simple immersion method. The modified textiles are integrated into two triboelectric nanogenerator (TENG) configurations: a solid–solid TENG (S–S TENG) and a liquid–solid TENG (L–S TENG). The S–S TENG, paired with a polydimethylsiloxane (PDMS) elastomer, achieves a peak output of 250 V, 1.01 µA, and a power output of 83.7 µW at 2 Hz when subject to a 5 N compression. The device exhibits long-term stability and charges a 10 µF capacitor to 3.3 V, with sufficient energy to power light-emitting diodes (LEDs). For the L–S TENG, deionised water droplets interacting with the polyethylene-coated surface generate up to 45 nW for a 50 MΩ load, with a saturated charge of 1.3 nC. When used as a sensor, the device is employed in real-time motion tracking, integrated with an artificial neural network, and in milk adulteration detection. These results demonstrate a low-cost, flexible, and eco-friendly platform for multifunctional energy harvesting and self-powered sensing, advancing circular economy principles and enabling new applications in healthcare, food safety, and wearable electronics.

2025-11-30T15:00:00Z Guezzi, Nizar Lee, Sangheon Nam, Sangwoo Jung, Dongkyu Noman, Muhammad Seong, Hyojin Lee, Sanghoon Kim, Hoe Joon Yu, Jaesok https://scholar.dgist.ac.kr/handle/20.500.11750/59332 2026-01-12T12:40:12Z 2026-01-31T15:00:00Z

Title: Contrast agent-free 3D ultrasound deep-depth vascular imaging with a 2D row column addressed Array: In vivo human clinical feasibility study Author(s): Guezzi, Nizar; Lee, Sangheon; Nam, Sangwoo; Jung, Dongkyu; Noman, Muhammad; Seong, Hyojin; Lee, Sanghoon; Kim, Hoe Joon; Yu, Jaesok Abstract: Three-dimensional (3D) imaging of vascular networks is essential for accurately diagnosing deep organ diseases. However, current ultrasound imaging methods are primarily limited to visualizing 2D cross-sections, which restricts the ability to evaluate the full structure of vascular networks. Although several 3D ultrasound techniques have been proposed to overcome this limitation, most struggle to achieve deep penetration and a wide field of view due to their high resource requirements. Row-column addressed arrays (RCAs) have emerged as a promising solution, enabling 3D imaging with significantly reduced hardware complexity. Nevertheless, the limited image quality achievable with RCAs has hindered their broader application. In this study, we propose a coded plane-wave-based, contrast-free 3D imaging system using RCAs for in vivo imaging of deep human vasculature. To validate the method, we imaged the liver and spleen of two healthy adult volunteers and successfully visualized vascular structures without contrast agent injection. Flow dynamics were captured at a frame rate of 27 Hz. Additionally, we demonstrated contrast-to-noise ratio (CNR) improvements of approximately 9 dB and 10 dB in the z-y and z-x planes, respectively, compared to non-coded excitation. This approach offers strong potential for in vivo 3D visualization and assessment of complex, deeply located vascular networks.

2026-01-31T15:00:00Z