짝지어진 데이터셋을 이용한 분할-정복 U-net 기반 고화질 초음파 영상 복원
A Divide-Conquer U-Net Based High-Quality Ultrasound Image Reconstruction Using Paired Dataset
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
Commonly deep learning methods for enhancing the quality of medical images use unpaired dataset due to the impracticality of acquiring paired dataset through commercial imaging system. In this paper, we propose a super- vised learning method to enhance the quality of ultrasound images. The U-net model is designed by incorporating a di- vide-and-conquer approach that divides and processes an image into four parts to overcome data shortage and shorten the learning time. The proposed model is trained using paired dataset consisting of 828 pairs of low-quality and high-quality images with a resolution of 512x512 pixels obtained by varying the number of channels for the same subject. Out of a total of 828 pairs of images, 684 pairs are used as the training dataset, while the remaining 144 pairs served as the test data- set. In the test results, the average Mean Squared Error (MSE) was reduced from 87.6884 in the low-quality images to 45.5108 in the restored images. Additionally, the average Peak Signal-to-Noise Ratio (PSNR) was improved from 28.7550 to 31.8063, and the average Structural Similarity Index (SSIM) was increased from 0.4755 to 0.8511, demonstrating sig- nificant enhancements in image quality.
Abstract
Commonly deep learning methods for enhancing the quality of medical images use unpaired dataset due to the impracticality of acquiring paired dataset through commercial imaging system. In this paper, we propose a super- vised learning method to enhance the quality of ultrasound images. The U-net model is designed by incorporating a di- vide-and-conquer approach that divides and processes an image into four parts to overcome data shortage and shorten the learning time. The proposed model is trained using paired dataset consisting of 828 pairs of low-quality and high-quality images with a resolution of 512x512 pixels obtained by varying the number of channels for the same subject. Out of a total of 828 pairs of images, 684 pairs are used as the training dataset, while the remaining 144 pairs served as the test data- set. In the test results, the average Mean Squared Error (MSE) was reduced from 87.6884 in the low-quality images to 45.5108 in the restored images. Additionally, the average Peak Signal-to-Noise Ratio (PSNR) was improved from 28.7550 to 31.8063, and the average Structural Similarity Index (SSIM) was increased from 0.4755 to 0.8511, demonstrating sig- nificant enhancements in image quality.
참고문헌 (16)
- Stock KF, "Comparison of a pocket-size ultrasound device with a premium ultrasound machine : diagnostic value and time required in bedside ultrasound examination", Abdominal Imag, 40(7), pp.2861~2866 (2015)
- Nelson BP, "Out of hospital point of care ultrasound : Current use models and future directions", Eur J Trauma Emergency Surgery, 42(2), pp.139~150 (2016)
- Tasinkiewicz J, "3D System Aperture Imaging Method in Spectrum Domain for Low-Cost Portable Ultrasound Systems", Arch. Acoust, 48(4), pp.559~572 (2023)
- Zhou Z, "Handheld Ultrasound Video High-Quality Reconstruction Using a Low-Rank Representation Multipathway Generative Adversarial Network", IEEE Trans Neural Netw Learn Syst, 32(2), pp.575~588 (2021)
- Michailovich OV, "Despeckling of medical ultrasound images", IEEE Trans Ultrason Ferroelectr Freq Control, 53(1), pp.64~78 (2006)
- Coupé P, "Nonlocal meansbased speckle filtering for ultrasound images", IEEE Trans Image Process, 18(10), pp.2221~2229 (2009)
- Hyun D, "Beamforming and Speckle Reduction Using Neural Networks", IEEE Trans Ultrason Ferroelectr Freq Control, 66(5), pp.898~910 (2019)
- Lan Y, "Real-Time Ultrasound Image Despeckling Using Mixed-Attention Mechanism Based Residual UNet", IEEE Access, 8, pp.195327~195340 (2020)
- Moinuddin M, "Medical ultrasound image speckle reduction and resolution enhancement using texture compensated multi-resolution convolution neural network", Front Physiol, 13, pp.961571~ (2022)
- Karaoğlu O, "Removal of speckle noises from ultrasound images using five different deep learning networks", Int J Eng Sci Technol, 29, pp.101030~ (2022)
- Skandarani Y, "GANs for Medical Image Synthesis: An Empirical Study", J. Imaging, 9(3), pp.69~ (2023)
- Zagzebski JA, "Essentials of ultrasound physics", pp.45~65 (1996), Mosby
- Jang J, "Industrial Mathematics in Ultrasound Imaging", J Ksiam, 20(3), pp.175~202 (2016)
- Ronneberger O, "Medical Image Computing and Computer-Assisted Intervention (MICCAI)", pp.234~241 (2015), Springer
- Hu X, "RUNet : A robust UNet architecture for image super-resolution", Conf Proc IEEE, 2019, ~