Article
Multidisciplinary Sciences
Yuyeon Jung, Taewan Kim, Mi-Ryung Han, Sejin Kim, Geunyoung Kim, Seungchul Lee, Youn Jin Choi
Summary: In this study, a convolutional neural network model was developed to classify ovarian tumors using pre-processed and augmented ultrasound images. The performance of the model was evaluated through cross-validation and validated qualitatively using Grad-CAM. The results demonstrated the accuracy and feasibility of the model.
SCIENTIFIC REPORTS
(2022)
Article
Computer Science, Artificial Intelligence
A. Shamla Beevi, S. Ratheesha, Saidalavi Kalady, Jenu James Chackola
Summary: A deep learning-based denoising model called Convolutional-based improved despeckling autoencoder (CIDAE) is proposed in this paper for denoising transthoracic echocardiographic images. The model is trained with a dataset collected from patients with Regional Wall Motion Abnormality (RWMA). The significance of the proposed CIDAE model for denoising echo images of patients with RWMA and structurally normal hearts is demonstrated through visual and quantitative evaluation.
Article
Physics, Multidisciplinary
Shun Li, Jiandong Mao, Xin Gong, Zhiyuan Li
Summary: In this study, an improved deep belief network (DBN) denoising algorithm is proposed to reduce noise in the lidar return signal. The algorithm combines denoising autoencoder (DAE) and restricted Boltzmann machine (RBM) to form a multi-layer DBN called DADBN. Experimental results show that DADBN outperforms four other denoising methods in noise reduction.
Article
Nanoscience & Nanotechnology
Hao Shi, Yuanhe Sun, Zhaofeng Liang, Shuqi Cao, Lei Zhang, Daming Zhu, Yanqing Wu, Zeying Yao, Wenqing Chen, Zhenjiang Li, Shumin Yang, Jun Zhao, Chunpeng Wang, Renzhong Tai
Summary: Scintillation-based X-ray imaging provides convenient visual observation of absorption contrast. By integrating imaging and postprocessing into one fused optical-electronic convolutional autoencoder network, the quality of obtained images can be improved and feature-specific enhancement of incoherent images can be realized.
Article
Chemistry, Analytical
Piotr Jozwik-Wabik, Krzysztof Bernacki, Adam Popowicz
Summary: Monochromatic images are often affected by noise, which reduces the quality of the results. Deterministic algorithms like Non-Local-Means and Block-Matching-3D are commonly used to reduce noise. This article focuses on using machine learning to denoise monochromatic images, even without access to noise-free data, and demonstrates that ML-based methods can achieve high performance.
Article
Environmental Sciences
Minghuan Hu, Jiandong Mao, Juan Li, Qiang Wang, Yi Zhang
Summary: A novel denoising method based on a convolutional autoencoding neural network was proposed to reduce noise in lidar signals, achieving better results compared to other methods when tested on simulated and measured signals.
Article
Environmental Sciences
Deshan Feng, Xiangyu Wang, Xun Wang, Siyuan Ding, Hua Zhang
Summary: This paper proposed a novel network structure for convolutional denoising autoencoders to enhance the signal-noise ratio of random ground penetrating radar noise. By addressing problems like overfitting and representational bottlenecks in deep learning, this approach significantly improved noise attenuation performance. The optimized network structure, including a dropout regularization layer and a residual-connection structure, effectively attenuated various types of noise while maintaining high-fidelity data information.
Article
Physics, Multidisciplinary
Armando Adrian Miranda-Gonzalez, Alberto Jorge Rosales-Silva, Dante Mujica-Vargas, Ponciano Jorge Escamilla-Ambrosio, Francisco Javier Gallegos-Funes, Jean Marie Vianney-Kinani, Erick Velazquez-Lozada, Luis Manuel Perez-Hernandez, Lucero Veronica Lozano-Vazquez
Summary: “Noise suppression algorithms have been widely used in various tasks, and this research proposes an unsupervised neural network architecture for Gaussian denoising. The proposed method outperforms other types of neural networks in suppressing image noise, as shown by objective numerical results.”
Review
Acoustics
Wenliao Du, Lingkai Yang, Hongchao Wang, Xiaoyun Gong, Lianying Zhang, Chuan Li, Lianqing Ji
Summary: In this paper, a noise learning method based on multilevel residual convolution autoencoder network (LN-MRSCAE) is proposed to eliminate noise interference in vibration signals of mechanical equipment. The LN-MRSCAE model, consisting of a deep convolutional autoencoder network and multilevel residual structure, encodes and decodes the noise components to obtain denoised signals. Experimental results demonstrate the superiority of LN-MRSCAE model in denoising compared to the latest models and the effective suppression of noise components in the signals.
JOURNAL OF VIBRATION AND CONTROL
(2023)
Article
Biochemical Research Methods
Alexander Kensert, Gilles Collaerts, Kyriakos Efthymiadis, Peter Van Broeck, Gert Desmet, Deirdre Cabooter
Summary: In this study, a deep one-dimensional convolutional autoencoder was developed to remove baseline noise and baseline drift in chromatograms with minimal information loss, outperforming traditional methods. The results show great potential for the autoencoder in enhancing and correcting chromatograms, improving downstream data analysis.
JOURNAL OF CHROMATOGRAPHY A
(2021)
Article
Geochemistry & Geophysics
Jinsheng Jiang, Haoran Ren, Meng Zhang
Summary: Petroleum geophysical exploration based on seismic data has been influenced by deep learning technology. The improved CAE method for simultaneous reconstruction and denoising of seismic data shows effectiveness in feature extraction and noise suppression.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Geochemistry & Geophysics
Feng Qian, Wei Guo, Zhangbo Liu, Hongtao Yu, Gulan Zhang, Guangmin Hu
Summary: This article presents an unsupervised deep learning method based on a robust deep convolutional autoencoder for removing erratic-plus-Gaussian noise. The method utilizes the concept of robust image denoising by replacing the mean squared error loss with the smooth Welsch function. The proposed method demonstrates its efficacy through experiments on both synthetic and real field datasets.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Chemistry, Multidisciplinary
Yu-Syuan Jhang, Szu-Ting Wang, Ming-Hwa Sheu, Szu-Hong Wang, Shin-Chi Lai
Summary: This paper presents a denoising autoencoder design that effectively removes electrode motion artifacts in an electrocardiogram signal. The proposed design has three advantages: reduced memory usage, preservation of key features, and fewer required parameters. Experimental results demonstrate that the proposed models outperform state-of-the-art methods in denoising performance.
APPLIED SCIENCES-BASEL
(2022)
Article
Computer Science, Artificial Intelligence
Fahimeh Mohagheghian, Dong Han, Om Ghetia, Darren Chen, Andrew Peitzsch, Nishat Nishita, Eric Y. Ding, Edith Mensah Otabil, Kamran Noorishirazi, Alexander Hamel, Emily L. Dickson, Danielle Dimezza, Khanh-Van Tran, David D. Mcmanus, Ki H. Chon
Summary: This study utilized data collected from smartwatches to improve the detection performance of atrial fibrillation using a denoising autoencoder. The results showed significant improvement in detecting occult AF and increased the amount of analyzable data.
EXPERT SYSTEMS WITH APPLICATIONS
(2024)
Article
Engineering, Electrical & Electronic
Fan Wu, Haiyong Luo, Hongwei Jia, Fang Zhao, Yimin Xiao, Xile Gao
Summary: The paper proposes a noise covariance estimation algorithm for GNSS/INS-integrated navigation using multitask learning model to achieve accurate and robust localization results under various complex and dynamic environments. Extensive experiments demonstrate a significant reduction in positioning error compared to traditional KF-based integrated navigation algorithm with predefined fixed settings.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2021)