Signal Processing - November 2017 - 22

derived. A constrained optimization
problem that considers group sparsity is
formulated to recover the 3-D motion
field from the 2-D motion. The recovered
3-D motion field is then clustered to provide the segmentation of moving objects.
2016

ORB-SLAM Map Initialization
Improvement Using Depth

GRAPHICSTOCK

A Deep Neural Network
for Image Quality Assessment
Bosse, S.; Maniry, D.; Wiegand, T.;
Samek, W.
This paper presents a no-reference
image quality assessment method
based on a deep convolutional neural
network (CNN). The CNN takes un--
preprocessed image patches as an
input and estimates the quality without
employing any domain knowledge. By
that, features and natural scene statistics are learned purely data driven and
combined with pooling and regression
in one framework.
2016

Fast Multidimensional Image
Processing with OpenCL
Oliveira Dantas, D.; Danilo Passos Leal, H.;
Oliveira Barros Sousa, D.
VisionGL is an open-source library
that provides a set of image processing
functions and can help the programmer by automatically generating code.
The objective of this work is to augment VisionGL by adding multidimensional image processing support with
OpenCL for high performance through
use of graphic processing units.

2016

Dimensionality Reduction
of Brain Imaging Data Using Graph
Signal Processing
Rui, L.; Nejati, H.; Cheung, N.-M.
This paper presents a new dimensionality reduction method based on the
recent graph signal processing theory
for the task of classifying the brain
imaging signals recording the cortical
activities in response to visual stimuli.
Authors propose using the resting-state
measurements (i.e., before onset of the
stimulus) of the subjects to build a
connectivity graph. The graph Laplacian and graph-based filtering are then
applied to learn the low-dimensional
linear subspace for the task-state
-m easurements (i.e., after onset of
the stimulus).
2016

Moving Object Segmentation
Using Depth and Optical Flow
in Car Driving Sequences
Kao, J.-Y.; Tian, D.; Mansour, H.;
Vetro, A.; Ortega, A.
In this paper, based on an analysis of
motion vanishing points of the scene and
estimated depth, a geometric model that
relates extracted two-dimensional (2-D)
motion to a three-dimensional (3-D)
motion field relative to the camera is
IEEE SIGNAL PROCESSING MAGAZINE

November 2017

Fujimoto, S.; Hu, Z.; Chapuis, R.;
Aufrère, R.
Map initialization and scale ambiguity
are well-known challenging problems
for visual simultaneous localization and mapping. In this paper, a triangulation is used on red, green, and blue
feature points for getting three-dimensional points from out of the limited
area in depth. The authors combined
both advantages of triangulation and
depth to improve the performance of
robustness to initialization and t-racking.

2016

Deep Learning Network for Blind
Image Quality Assessment
Gu, K.; Zhai, G.; Yang, X.; Zhang, W.
The authors in this paper introduce a
new deep-learning-based image quality
index (DIQI) for blind quality assessment. Extensive studies are conducted
on the new TID2013 database and confirm the effectiveness of their DIQI
relative to classical full-reference and
state-of-the-art reduced- and no-reference IQA approaches.
2014

Depth Augmented Stereo
Panorama for Cinematic Virtual
Reality with Focus Cues
Thatte, J.; Boin, J.-B.; Lakshman, H.;
Wetzstein, G.; Girod, B.
Cinematic virtual reality aims to provide
immersive visual experiences of realworld scenes on head-mounted displays.
The authors propose a new content re--
presentation, depth augmented stereo
panorama, which permits generating
light fields across the observer's pupils,
achieving an order of magnitude reduction in data requirements compared to the
existing techniques.
2016

Table of Contents for the Digital Edition of Signal Processing - November 2017