检索结果-内蒙古大学图书馆

Unified Architecture for Multichannel End-to-End Speech Recognition With Neural Beamforming

IEEE JOURNAL OF SELECTED TOPICS IN SIGNAL PROCESSING 2017年第8期11卷 1274-1288页

作者： Ochiai, Tsubasa Watanabe, Shinji Hori, Takaaki Hershey, John R. Xiao, Xiong Doshisha Univ Kyotanabe 6100394 Japan Mitsubishi Elect Res Labs Cambridge MA 02139 USA Nanyang Technol Univ Singapore 639798 Singapore

This paper proposes a unified architecture for end-to-end automatic speech recognition (ASR) to encompass microphone-array signal processing such as a state-of-the-art neural beamformer within the end-to-end framework. Recently, the end-to-end ASR paradigm has attracted great research interest as an alternative to conventional hybrid paradigms with deep neural networks and hidden Markov models. Using this novel paradigm, we simplify ASR architecture by integrating such ASR components as acoustic, phonetic, and language models with a single neural network and optimize the overall components for the end-to-end ASR objective: generating a correct label sequence. Although most existing end-to-end frameworks have mainly focused on ASR in clean environments, our aim is to build more realistic end-to-end systems in noisy environments. To handle such challenging noisy ASR tasks, we study multichannel end-to-end ASR architecture, which directly converts multichannel speech signal to text through speech enhancement. This architecture allows speech enhancement and ASR components to be jointly optimized to improve the end-to-end ASR objective and leads to an end-to-end framework that works well in the presence of strong background noise. We elaborate the effectiveness of our proposed method on the multichannel ASR benchmarks in noisy environments (CHiME-4 and AMI). The experimental results show that our proposed multichannel end-to-end system obtained performance gains over the conventional end-to-end baseline with enhanced inputs from a delay-and-sum beamformer (i.e., BeamformIT) in terms of character error rate. In addition, further analysis shows that our neural beamformer, which is optimized only with the end-to-end ASR objective, successfully learned a noise suppression function.

关键词： Multichannel end-to-end ASR neural beamformer encoder-decoder network

来源：评论

学校读者我要写书评

暂无评论

Global River Monitoring Using Semantic Fusion networks

引用

WATER 2020年第8期12卷 2258页

作者： Wei, Zhihao Jia, Kebin Jia, Xiaowei Khandelwal, Ankush Kumar, Vipin Beijing Univ Technol Fac Informat Technol Dept Informat & Commun Engn Beijing 100124 Peoples R China Univ Minnesota Dept Comp Sci & Engn Minneapolis MN 55455 USA

Global river monitoring is an important mission within the remote sensing society. One of the main challenges faced by this mission is generating an accurate water mask from remote sensing images (RSI) of rivers (RSIR), especially on a global scale with various river features. Aiming at better water area classification using semantic information, this paper presents a segmentation method for global river monitoring based on semantic clustering and semantic fusion. Firstly, an encoder-decoder network (AEN)-based architecture is proposed to obtain the semantic features from RSIR. Secondly, a clustering-based semantic fusion method is proposed to divide semantic features of RSIR into groups and train convolutional neural networks (CNN) models corresponding to each group using data augmentation and semi-supervised learning. Thirdly, a semantic distance-based segmentation fusion method is proposed for fusing the CNN models result into final segmentation mask. We built a global river dataset that contains multiple river segments from each continent of the world based on Sentinel-2 satellite imagery. The result shows that the F1-score of the proposed segmentation method is 93.32%, which outperforms several state-of-the-art algorithms, and demonstrates that grouping semantic information helps better segment the RSIR in global scale.

关键词： convolution encoder-decoder network feature extraction remote sensing image of river semantic fusion semi-supervised learning

来源：评论

学校读者我要写书评

暂无评论

NeuroGAN: image reconstruction from EEG signals via an attention-based GAN

引用

NEURAL COMPUTING & APPLICATIONS 2023年第12期35卷 9181-9192页

作者： Mishra, Rahul Sharma, Krishan Jha, R. R. Bhavsar, Arnav Indian Inst Technol Mandi Sch Comp & Elect Engn MANAS Lab Mandi India Vehant Technol Pvt Ltd Noida India

In this work, we propose an approach to process electroencephalogram (EEG) for a visual perception task for synthesizing the visual stimulus that was shown during the acquisition of EEG (images of objects, digits, and characters). We demonstrate that a cross-modality-based encoder-decoder network shows good performance for image synthesis tasks on simplistic images like digits and characters but fails on complex natural object images. To address this issue, we propose a novel attention & auxiliary classifier-based GAN architecture where the generator itself is a cross-modality-based encoder-decoder network. It generates images along with producing class-specific EEG encoding as a latent representation. In addition to the traditional adversarial loss, we also propose to use perceptual loss and attention modules to generate good-quality images. The performance of the proposed network is measured using two metrics- diversity score and inception score, which quantify the relevance and quality of the reconstructed images, respectively. Experimentation results show that our approach performs better compared to the state-of-the-art for both metrics.

关键词： Attention module BCI Electroencephalogram (EEG) encoder-decoder network GAN Perceptual loss

来源：评论

学校读者我要写书评

暂无评论

Large-scale water quality prediction with integrated deep neural network

引用

INFORMATION SCIENCES 2021年 571卷 191-205页

作者： Bi, Jing Lin, Yongze Dong, Quanxi Yuan, Haitao Zhou, MengChu Beijing Univ Technol Sch Software Engn Fac Informat Technol Beijing 100124 Peoples R China Beihang Univ Sch Automat Sci & Elect Engn Beijing 100191 Peoples R China New Jersey Inst Technol Dept Elect & Comp Engn Newark NJ 07102 USA

Water environment time series prediction is important to efficient water resource manage-ment. Traditional water quality prediction is mainly based on linear models. However, owing to complex conditions of the water environment, there is a lot of noise in the water quality time series, which will seriously affect the accuracy of water quality prediction. In addition, linear models are difficult to deal with the nonlinear relations of data of time ser-ies. To address this challenge, this work proposes a hybrid model based on a long short -term memory-based encoder-decoder neural network and a Savitzky-Golay filter. Among them, the filter of Savitzky-Golay can eliminate the potential noise in the time series of water quality, and the long short-term memory can investigate nonlinear characteristics in a complicated water environment. In this way, an integrated model is proposed and effectively obtains statistical characteristics. Realistic data-based experiments prove that its prediction performance is better than its several state-of-the-art peers. (c) 2021 Elsevier Inc. All rights reserved.

关键词： Savitzky-Golay filter Prediction algorithms Water quality management Deep neural network encoder-decoder network

来源：评论

学校读者我要写书评

暂无评论

A Deep Neural network Using Double Self-Attention Mechanism for ALS Point Cloud Segmentation

引用

IEEE ACCESS 2022年 10卷 29878-29889页

作者： Yu, Lili Yu, Haiyang Yang, Shuai Henan Polytech Univ Sch Surveying & Land Informat Engn Jiaozuo 454000 Henan Peoples R China

Airborne laser scanning (ALS) point cloud segmentation is an essential procedure for 3D data understanding and applications. This task is challenging due to the unstructured, disordered, and sparse distribution of the point cloud. PointNet++ is a well known end-to-end learning network for point cloud segmentation without fully exploring the local and contextual features, which are less efficient and accurate in capturing the complexity of point clouds. On this basis, we design a novel encoder-decoder network architecture to obtain the semantic features of the ALS point cloud at different levels and achieve a better segmentation effect. The improved local feature aggregation module can merge the deep feature of the point cloud, combining local and global self-attention convolutional networks. It can adaptively explore the inherent semantics feature of points and capture more extensive context information of ALS point cloud, termed DSPNet++. Finally, the conditional random field optimization model can be used to refine the segmentation results. We evaluated the performance of our method on the Vaihingen dataset of the International Society for Photogrammetry and Remote Sensing (ISPRS) and the GML(B) 3D dataset. Experimental results show that our method fully exploits the semantic feature of the ALS point cloud and can achieve higher accuracy. A comparative study with established deep learning models also confirms that our proposed method has outstanding performance in the ALS point cloud segmentation task.

关键词： Point cloud compression Feature extraction Laser radar Semantics Three-dimensional displays Convolutional neural networks Neural networks ALS point cloud segmentation PointNet plus plus network encoder-decoder network local and global self-attention convolutional network

来源：评论

学校读者我要写书评

暂无评论

A multi-scale attentive recurrent network for image dehazing

引用

MULTIMEDIA TOOLS AND APPLICATIONS 2021年第21-23期80卷 32539-32565页

作者： Wang, Yibin Yin, Shibai Basu, Anup Sichuan Normal Univ Dept Engn Chengdu 610066 Peoples R China Southwestern Univ Finance & Econ Dept Econ Informat Engn Chengdu 611130 Peoples R China Univ Alberta Dept Comp Sci Edmonton AB T6G 2E8 Canada

Image dehazing is a pre-processing step in computer vision tasks, that has attracted considerable attention from the research community. Existing CNN-based methods ignore haze-related priors and rarely use a coarse-to-fine scheme in a feed-forward architecture to remove haze due to increasing network depth and parameters. This results in sub-optimal dehazing results. To address these problems, a multi-scale attentive recurrent network is proposed for image dehazing, which consists of a haze attention map predicted network and a recurrent encoder-decoder network. First, by assuming that haze in an image is formed by multiple layers with different depths, the haze attention map predicted network is designed for generating the map with multiple stages via a multi-scale recurrent framework. Second, the haze attention map is viewed as the haze-related prior and guides the subsequent recurrent encoder-decoder network to be aware of haze concentration information. Finally, for leveraging the intermediate information and optimizing the dehazing result with less parameters and more robust features, the recurrent residual operations which pass the features of selected layers at the current time step to the corresponding layers at the next time step are applied in the recurrent encoder-decoder network for removing haze following a coarse-to-fine strategy. Experiments on synthetic and real images demonstrate that our method outperforms state-of-the-art methods considering both visual and quantitative evaluations. In addition, our method is also suitable for real-time processing.

关键词： encoder-decoder network Image dehazing Multi-scale fusion Recurrent residual operation

来源：评论

学校读者我要写书评

暂无评论

Survey of Mathematical Expression Recognition for Printed and Handwritten Documents

引用

IETE TECHNICAL REVIEW 2022年第6期39卷 1245-1253页

作者： Aggarwal, Ridhi Pandey, Shilpa Tiwari, Anil Kumar Harit, Gaurav IIT Jodhpur Dept Comp Sci & Engn Jodhpur 342011 Rajasthan India Adani Inst Infrastruct Engn Ahmedabad Dept ICT Ahmadabad 382421 Gujarat India IIT Jodhpur Dept Elect Engn Jodhpur 342011 Rajasthan India

In this paper, we provide a review of Mathematical Expression Recognition (MER) for both printed and handwritten domains. We describe the past work in a manner that clearly highlights the common and distinguishing aspects. The review reveals the advances in the techniques used for the detection and recognition of MEs for online and offline modalities. Systems have evolved from the use of domain-specific crafted modules to architectures learned in a completely data-driven manner. Learning tasks and datasets have been standardized, thus allowing performance comparison of the systems. With automated learning of improved feature representations and guided by the related fields such as machine translation and captioning, the systems have progressed to demonstrate improved robustness to the variability in the inputs.

关键词： encoder-decoder network Mathematical expression detection Mathematical expression recognition Symbol recognition Visual parsing

来源：评论

学校读者我要写书评

暂无评论

ColonFormer: An Efficient Transformer Based Method for Colon Polyp Segmentation

引用

IEEE ACCESS 2022年 10卷 80575-80586页

作者： Nguyen Thanh Duc Nguyen Thi Oanh Nguyen Thi Thuy Tran Minh Triet Dinh Viet Sang Hanoi Univ Sci & Technol Sch Informat & Commun Technol Hanoi 10000 Vietnam Univ Sci VNU HCM Fac Informat Technol Ho Chi Minh City 70000 Vietnam Univ Sci VNU HCM Software Engn Lab Ho Chi Minh City 70000 Vietnam Univ Sci VNU HCM Ho Chi Minh City 70000 Vietnam Viet Nam Natl Univ Ho Chi Minh City John Von Neumann Inst Ho Chi Minh City 70000 Vietnam

Identifying polyps is challenging for automatic analysis of endoscopic images in computer-aided clinical support systems. Models based on convolutional networks (CNN), transformers, and their combinations have been proposed to segment polyps with promising results. However, those approaches have limitations either in modeling the local appearance of the polyps only or lack of multi-level feature representation for spatial dependency in the decoding process. This paper proposes a novel network, namely ColonFormer, to address these limitations. ColonFormer is an encoder-decoder architecture capable of modeling long-range semantic information at both encoder and decoder branches. The encoder is a lightweight architecture based on transformers for modeling global semantic relations at multi scales. The decoder is a hierarchical network structure designed for learning multi-level features to enrich feature representation. Besides, a refinement module is added with a new skip connection technique to refine the boundary of polyp objects in the global map for accurate segmentation. Extensive experiments have been conducted on five popular benchmark datasets for polyp segmentation, including Kvasir, CVC-Clinic DB, CVC-ColonDB, CVC-T, and ETIS-Larib. Experimental results show that our ColonFormer outperforms other state-of-the-art methods on all benchmark datasets. Our code is available at: https://***/ducnt9907/ColonFormer.

关键词： Transformers Image segmentation Computer architecture Decoding Feature extraction Convolutional neural networks Computational modeling Polyp segmentation deep learning encoder-decoder network hierarchical multi-scale CNN computer-aided diagnosis

来源：评论

学校读者我要写书评

暂无评论

sEnDec: An Improved Image to Image CNN for Foreground Localization

引用

IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS 2020年第10期21卷 4435-4443页

作者： Akilan, Thangarajah Wu, Q. M. Jonathan Lakehead Univ Dept Comp Sci Thunder Bay ON P7B 5E1 Canada Univ Windsor Ctr Comp Vis & Deep Learning Windsor ON N9B 2P4 Canada Univ Windsor Dept Elect & Comp Engn Windsor ON N9B 2P4 Canada

Although it is not immediately intuitive that Deep Convolutional Neural networks (DCNNs) can yield adequate feature representation for a Foreground Localization (FGL) task, recent architectural and algorithmic advancements in Deep Learning (DL) have shown that the DCNNs have become forefront methodology for this pixel-level classification problem. In FGL, the DCNNs face an inherent trade-off between moving objects, i.e., the foreground (FG) and the non-static background (BG) scenes, through learning from local- and global-level features. Driven by the latest success of the innovative structures for image classification and semantic segmentation, this work introduces a novel architecture, called Slow encoder-decoder (sEnDec) that aims to improve the learning capacity of a traditional image-to-image DCNN. The proposed model subsumes two subnets for contraction (encoding) and expansion (decoding), wherein both phases, it employs an intermediate feature map up-sampling and residual connections. In this way, the lost structural details due to spatial subsampling are recovered. It helps to get a more delineated FG region. The experimental study is carried out with two variants of the proposed model: one with strided convolution (conv) and the other with max pooling for spatial subsampling. A comparative analysis on sixteen benchmark video sequences, including baseline, dynamic background, camera jitter, shadow effects, intermittent object motion, night videos, and bad weather show that the proposed sEnDec model performs very competitively against the prior- and state-of-the-art approaches.

关键词： Computational modeling Computer architecture Encoding Image segmentation Image color analysis Decoding Visualization Foreground localization DCNN encoder-decoder network

来源：评论

学校读者我要写书评

暂无评论

A multi-step water quality prediction model based on the Savitzky-Golay filter and Transformer optimized network

引用

ENVIRONMENTAL SCIENCE AND POLLUTION RESEARCH 2023年第50期30卷 109299-109314页

作者： Wang, Ruiqi Qi, Ying Zhang, Qiang Wen, Fei Northwest Normal Univ Dept Comp Sci & Engn Lanzhou Gansu Peoples R China Gansu Acad Ecoenvironm Sci Lanzhou Gansu Peoples R China

Effective water quality prediction techniques are essential for the sustainable development of water resources and implementation of emergency response mechanisms. However, the water environment conditions are complex, and the presence of a large amount of noise in the water quality data makes it difficult to reveal the long-term trends or cycles of the data, affecting the acquisition of serial correlation in the data. In addition, the loss function based on the vertical Euclidean distance will produce a prediction lag problem, and it is difficult to make an accurate multi-step prediction of water quality series. This paper presents a multi-step water quality prediction model for watersheds that combines Savitzky-Golay (SG) filter with Transformer optimized networks. Among them, the SG filter highlights data trend change and improves sequence correlation by smoothing the potential noise of original data. The transformer network adopts a sequence-to-sequence framework, which contains a position encoding module and a self-attentive mechanism to perform multi-step prediction while effectively obtaining the sequence correlation. Moreover, the DIstortion Loss including shApe and TimE (DILATE) loss function is introduced into the model to solve the problem of prediction lag from two aspects of shape error and time error to improve the model's generalization ability. An example validates the model with the benchmark model at four monitoring stations in the Lanzhou section of the Yellow River basin in China. The results show that the predictions of the proposed model have the correct shape, temporal positioning, and the best accuracy in a multi-step prediction task for four sites. It can provide a decision-making basis for comprehensive water quality control and pollutant control in the basin.

关键词： Savitzky-Golay Filter Transformer network Attention mechanism DILATE Water quality prediction encoder-decoder network

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：