scispace - formally typeset
Search or ask a question
Author

Chenghao Wang

Bio: Chenghao Wang is an academic researcher. The author has contributed to research in topics: Autoencoder & Deep learning. The author has co-authored 1 publications.

Papers
More filters
Journal ArticleDOI
TL;DR: A multi-scale convolutional autoencoder (MCAE) to denoise GPR data and the data augmentation strategy, Wasserstein generative adversarial network (WGAN), to increase the training dataset of MCAE is designed to solve the problem of training dataset insufficiency.
Abstract: Ground-penetrating radar (GPR), as a non-invasive instrument, has been widely used in civil engineering. In GPR B-scan images, there may exist random noise due to the influence of the environment and equipment hardware, which complicates the interpretability of the useful information. Many methods have been proposed to eliminate or suppress the random noise. However, the existing methods have an unsatisfactory denoising effect when the image is severely contaminated by random noise. This paper proposes a multi-scale convolutional autoencoder (MCAE) to denoise GPR data. At the same time, to solve the problem of training dataset insufficiency, we designed the data augmentation strategy, Wasserstein generative adversarial network (WGAN), to increase the training dataset of MCAE. Experimental results conducted on both simulated, generated, and field datasets demonstrated that the proposed scheme has promising performance for image denoising. In terms of three indexes: the peak signal-to-noise ratio (PSNR), the time cost, and the structural similarity index (SSIM), the proposed scheme can achieve better performance of random noise suppression compared with the state-of-the-art competing methods (e.g., CAE, BM3D, WNNM).

7 citations

Journal ArticleDOI
TL;DR: The development history and application fields of some representative neural networks are introduced and the importance of studying deep learning technology is pointed out, as well as the reasons and advantages of using FPGA to accelerate deep learning.
Abstract: Deep learning based on neural networks has been widely used in image recognition, speech recognition, natural language processing, automatic driving, and other fields and has made breakthrough progress. FPGA stands out in the field of accelerated deep learning with its advantages such as flexible architecture and logic units, high energy efficiency ratio, strong compatibility, and low delay. In order to track the latest research results of neural network optimization technology based on FPGA in time and to keep abreast of current research hotspots and application fields, the related technologies and research contents are reviewed. This paper introduces the development history and application fields of some representative neural networks and points out the importance of studying deep learning technology, as well as the reasons and advantages of using FPGA to accelerate deep learning. Several common neural network models are introduced. Moreover, this paper reviews the current mainstream FPGA-based neural network acceleration technology, method, accelerator, and acceleration framework design and the latest research status, pointing out the current FPGA-based neural network application facing difficulties and the corresponding solutions, as well as prospecting the future research directions. We hope that this work can provide insightful research ideas for the researchers engaged in the field of neural network acceleration based on FPGA.

3 citations

Proceedings ArticleDOI
28 Oct 2022
TL;DR: In this article , a hybrid CNN-GRU model was proposed to predict the remaining useful life (RUL) of an aero-engine based on the multi-source heterogeneous sensor data.
Abstract: The prediction of remaining useful life (RUL) for an aero-engine is crucial to ensure the operation safety and reliability of an aircraft. Inspired by the data-driven method, we propose a hybrid Convolutional Neural Network (CNN) and Gated Recurrent Neural Network (GRU) model to predict the RUL based on the multi-source heterogeneous sensor data in this study. The proposed hybrid CNN-GRU model takes the advantage that CNN can effectively extract the features of multi-sensor data on spatial-temporal dimensions, and GRU can figure out the problem of long-term dependence with the superiority of less complicated model structure in the processing of time-series data. Experiments on the NASA C-MAPSS dataset are conducted by using the proposed model, and the RUL prediction results are presented. The results show that the hybrid CNN-GRU model has an improvement in prediction accuracy compared with other single-network models.

Cited by
More filters
Journal ArticleDOI
TL;DR: In this paper , a YOLOv3 model with four-scale detection layers (FDL) was proposed to detect combined B-scan and C-scan ground penetrating radar (GPR) images.

35 citations

Journal ArticleDOI
TL;DR: In this paper , a DeepAugment data augmentation strategy combined with object detection models was proposed to detect internal cracks in asphalt pavement using ground penetrating radar (GPR) images.

27 citations

Journal ArticleDOI
TL;DR: The proposed neural network has the number reduction in the trainable parameters and the increase in the similarity between the denoised or deblurred image and the original one and the similarity is increased by decreasing the main square error and increasing the structural similarity index.
Abstract: A dual autoencoder employing separable convolutional layers for image denoising and deblurring is represented. Combining two autoencoders is presented to gain higher accuracy and simultaneously reduce the complexity of neural network parameters by using separable convolutional layers. In the proposed structure of the dual autoencoder, the first autoencoder aims to denoise the image, while the second one aims to enhance the quality of the denoised image. The research includes Gaussian noise (Gaussian blur), Poisson noise, speckle noise, and random impulse noise. The advantages of the proposed neural network are the number reduction in the trainable parameters and the increase in the similarity between the denoised or deblurred image and the original one. The similarity is increased by decreasing the main square error and increasing the structural similarity index. The advantages of a dual autoencoder network with separable convolutional layers are demonstrated by a comparison of the proposed network with a convolutional autoencoder and dual convolutional autoencoder.

1 citations

Proceedings ArticleDOI
15 Feb 2022
TL;DR: In this paper , a GPR EI method based on GPR multi-frequency (MF) data and A-Unet deep learning framework is presented, which can reconstruct the shape distribution of buried objects and has become an important research direction of underground target imaging.
Abstract: Subsurface imaging technique has good application value in the field of Ground Penetrating Radar (GPR). Electromagnetic Inversion (EI) can reconstruct the shape distribution of buried objects and has become an important research direction of underground target imaging. This paper presents a GPR EI method based on GPR Multi-Frequency (MF) data and A-Unet deep learning framework. Firstly, GPR B-scan data are collected by real aperture or synthetic aperture and then pre-processed by using background removal and denoising technique. Secondly, a A-Unet deep learning network is designed to achieve underground target imaging. It’s input data is multi-scan MF amplitude and phase data extracted from pre-processed GPR B-Scan data, while it’s output is underground dielectric parameters distribution in a designated regime. This A-Unet compose of a data extraction unit and a data expansion unit. The data extraction unit is characterized by replacing the skip-connection structure of Unet with an add-structure, which improves network computing efficiency. The data expansion unit is used to improve the resolution of electrical permittivity distribution. Numerical simulation experiments have proved that this method effectively reconstructs the shape distribution of underground targets, and the training time of add-structure is shortened to 9.09% of the training time of skip-connection unit while without reducing the imaging resolution.

1 citations

Journal ArticleDOI
TL;DR: In this paper , a boosted stacking ensemble model that integrates the four base level classifiers with the extreme gradient boosting classifier as a meta-level classifier is designed and built for training the model based on the expanded data for final prediction.
Abstract: Recently, machine learning-based technologies have been developed to automate the classification of wafer map defect patterns during semiconductor manufacturing. The existing approaches used in the wafer map pattern classification include directly learning the image through a convolution neural network and applying the ensemble method after extracting image features. This study aims to classify wafer map defects more effectively and derive robust algorithms even for datasets with insufficient defect patterns. First, the number of defects during the actual process may be limited. Therefore, insufficient data are generated using convolutional auto-encoder (CAE), and the expanded data are verified using the evaluation technique of structural similarity index measure (SSIM). After extracting handcrafted features, a boosted stacking ensemble model that integrates the four base-level classifiers with the extreme gradient boosting classifier as a meta-level classifier is designed and built for training the model based on the expanded data for final prediction. Since the proposed algorithm shows better performance than those of existing ensemble classifiers even for insufficient defect patterns, the results of this study will contribute to improving the product quality and yield of the actual semiconductor manufacturing process.