2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP)最新文献_第2页

Human fall detection using segment-level cnn features and sparse dictionary learning 基于片段级cnn特征和稀疏字典学习的人体跌倒检测

2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP) Pub Date : 2017-09-01 DOI: 10.1109/MLSP.2017.8168185

C. Ge, I. Gu, Jie Yang

引用次数: 10

Blind channel equalization of encoded data over galois fields 伽罗瓦域上编码数据的盲信道均衡

2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP) Pub Date : 2017-09-01 DOI: 10.1109/MLSP.2017.8168135

D. Fantinato, A. Neves, D. G. Silva, R. Attux

引用次数: 0

The time series cluster kernel 时间序列簇核

2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP) Pub Date : 2017-09-01 DOI: 10.1109/MLSP.2017.8168187

Karl Øyvind Mikalsen, F. Bianchi, C. Soguero-Ruíz, R. Jenssen

引用次数: 1

A deep neural network witharestricted noisy channel for identification of functional introns 基于受限噪声通道的深度神经网络识别功能内含子

2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP) Pub Date : 2017-09-01 DOI: 10.1109/MLSP.2017.8168186

Alan Joseph Bekker, M. Chorev, L. Carmel, J. Goldberger

引用次数: 1

Noise reduction in low-dose ct using a 3D multiscale sparse denoising autoencoder 基于三维多尺度稀疏去噪自编码器的低剂量ct降噪研究

2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP) Pub Date : 2017-09-01 DOI: 10.1109/MLSP.2017.8168176

K. Mentl, B. Mailhé, Florin C. Ghesu, Frank Schebesch, T. Haderlein, A. Maier, M. Nadar

{"title":"Noise reduction in low-dose ct using a 3D multiscale sparse denoising autoencoder","authors":"K. Mentl, B. Mailhé, Florin C. Ghesu, Frank Schebesch, T. Haderlein, A. Maier, M. Nadar","doi":"10.1109/MLSP.2017.8168176","DOIUrl":"https://doi.org/10.1109/MLSP.2017.8168176","url":null,"abstract":"This article presents a novel neural network-based approach for enhancement of 3D medical image data. The proposed networks learn a sparse representation basis by mapping the corrupted input data to corresponding optimal targets. To reinforce the adjustment of the network to the given data, the threshold values are also adaptively learned. In order to capture important image features on various scales and be able to process large computed tomography (CT) volumes in a reasonable time, a multiscale approach is applied. Recursively downsampled versions of the input are used and denoising operator of constant size are learnt at each scale. The networks are trained end-to-end from a database of real highdose acquisitions with synthetic additional noise to simulate the corresponding low-dose scans. Both 2D and 3D networks are evaluated on CT volumes and compared to the block-matching and 3D filtering (BM3D) algorithm. The presented methods achieve an increase of 4% to 11% in the SSIM and of 2.4 to 2.8 dB in the PSNR with respect to the ground truth, outperform BM3D in quantitative comparisions and present no visible texture artifacts. By exploiting volumetric information, 3D networks achieve superior results over 2D networks.","PeriodicalId":6542,"journal":{"name":"2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP)","volume":"6 1","pages":"1-6"},"PeriodicalIF":0.0,"publicationDate":"2017-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87680203","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 4

Upper bound performance of semi-definite programming for localisation in inhomogeneous media 非均匀介质中半确定规划的上界性能

2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP) Pub Date : 2017-09-01 DOI: 10.1109/MLSP.2017.8168145

E. Nadimi, V. Blanes-Vidal

引用次数: 0

A two-stage training deep neural network for small pedestrian detection 一种用于小型行人检测的两阶段训练深度神经网络

2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP) Pub Date : 2017-09-01 DOI: 10.1109/MLSP.2017.8168106

Tran Duy Linh, Masayuki Arai

引用次数: 1

A layer-block-wise pipeline for memory and bandwidth reduction in distributed deep learning 分布式深度学习中减少内存和带宽的分层块管道

2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP) Pub Date : 2017-09-01 DOI: 10.1109/MLSP.2017.8168127

Haruki Mori, Tetsuya Youkawa, S. Izumi, M. Yoshimoto, H. Kawaguchi, Atsuki Inoue

{"title":"A layer-block-wise pipeline for memory and bandwidth reduction in distributed deep learning","authors":"Haruki Mori, Tetsuya Youkawa, S. Izumi, M. Yoshimoto, H. Kawaguchi, Atsuki Inoue","doi":"10.1109/MLSP.2017.8168127","DOIUrl":"https://doi.org/10.1109/MLSP.2017.8168127","url":null,"abstract":"This paper describes a pipelined stochastic gradient descent (SGD) algorithm and its hardware architecture with a memory distributed structure. In the proposed architecture, a pipeline stage takes charge of multiple layers: a “layer block.” The layer-block-wise pipeline has much less weight parameters for network training than conventional multithreading because weight memory is distributed to workers assigned to pipeline stages. The memory capacity of 2.25 GB for the four-stage proposed pipeline is about half of the 3.82 GB for multithreading when a batch size is 32 in VGG-F. Unlike multithreaded data parallelism, no parameter server for weight update or shared I/O data bus is necessary. Therefore, the memory bandwidth is drastically reduced. The proposed four-stage pipeline only needs memory bandwidths of 36.3 MB and 17.0 MB per batch, respectively, for forward propagation and backpropagation processes, whereas four-thread multithreading requires a bandwidth of 974 MB overall for send and receive processes to unify its weight parameters. At the parallelization degree of four, the proposed pipeline maintains training convergence by a factor of 1.12, compared with the conventional multithreaded architecture although the memory capacity and the memory bandwidth are decreased.","PeriodicalId":6542,"journal":{"name":"2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP)","volume":"1 1","pages":"1-6"},"PeriodicalIF":0.0,"publicationDate":"2017-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74535184","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 3

Mutual singular spectrum analysis for bioacoustics classification 生物声学分类的互奇异谱分析

2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP) Pub Date : 2017-09-01 DOI: 10.1109/MLSP.2017.8168113

B. Gatto, J. Colonna, E. M. Santos, E. Nakamura

引用次数: 18

A regularized sequential dictionary learning algorithm for fmri data analysis 一种用于fmri数据分析的正则顺序字典学习算法

2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP) Pub Date : 2017-09-01 DOI: 10.1109/MLSP.2017.8168146

A. Seghouane, Asif Iqbal

引用次数: 8