IEEE Signal Processing - July 2018 - 10

bandwidth of 1 GHz. Extensive experiment results demonstrate a perfect 5-cm
precision in a 20 cm × 70 cm area in a
nonline-of-sight office environment with
one link measurement.
2016

Automatic Speech Emotion
Recognition Using Recurrent Neural
Networks with Local Attention
Mirsamadi, S.; Barsoum, E.; Zhang, C.
In this work, the authors use a deep recurrent neural network to learn both the
short-time frame-level acoustic features
that are emotionally relevant as well as
an appropriate temporal aggregation of
those features into a compact utterancelevel representation.
2017

Deep Convolutional Activation
Features for Large-Scale Brain
Tumor Histopathology image
Classification and Segmentation
Xu, Y.; Jia, Z.; Ai, Y.; Zhang, F.;
Lai, M.; Chang, E. I.-C.
The authors propose a simple, efficient,
and effective method using deep convolutional activation features (CNNs)
to achieve state-of-the-art classification and segmentation for the MICCAI
2014 Brain Tumor Digital Pathology
Challenge. They use CNN activations trained by ImageNet to extract

IMAGE LICENSED BY GRAPHIC STOCK

10

features and transfer the features to the
medical image.
2015

Neural Network Based Spectral
Mask Estimation for Acoustic
Beamforming
Heymann J.; Drude, L.; Haeb-Umbach, R.
A neural network-based approach to acoustic beamforming is presented in this paper.
The network is used to estimate spectral
masks from which the cross-power spectral density matrices of speech and noise
are estimated, which, in turn, are used to
compute the beamformer coefficients.
2016

informed Source Separation via
Compressive Graph Signal Sampling
Puy, G.; Ozerov, A.; Duong,
N.Q.K.; Pérez, P.
The authors propose a new informed
source separation method for audio object
coding based on a recent sampling theory
for smooth signals on graphs. They show
that the proposed approach performs better
than state-of-the-art methods at low bit rate.
2017

ECG-Based Biometrics Using
Recurrent Neural Networks
Salloum, R.; Kuo, C.-C. Jay
In this paper, the authors propose the use
of recurrent neural networks to develop

We want
to hear
from you!

an effective solution to two problems in
electrocardiogram-based biometrics:
identification/classification and authentication. The method was evaluated using
two publicly available data sets.
2017

Lognet: Energy-Efficient Neural
Networks Using Logarithmic
Computation
Lee, E.H.; Miyashita, D.; Chai, E;
Murmann, B.; Wong, S.S.
The authors present the concept of logarithmic computation for neural networks
and explore how logarithmic encoding
of nonuniformly distributed weights
and activations is preferred over linear encoding at resolutions of 4 bits and fewer.
2017

Digital Predistortion for Hybrid
Precoding Architecture in MillimeterWave Massive MiMo Systems
Yan, H.; Cabric, D.
In this work, the authors use memory
polynomial model for modeling millimeter-wave power amplifier and study
the interbeam and intersymbol interference in hybrid precoding architecture.
They also propose a digital predistortion algorithm to mitigate interference.
2017

Do you like what you're reading?
Your feedback is important.
Let us know-send the editor-in-chief an e-mail!

IEEE Signal Processing Magazine

|

July 2018

|

sP
Table of Contents for the Digital Edition of IEEE Signal Processing - July 2018
