Sorry, you need to enable JavaScript to visit this website.

facebooktwittermailshare

Learning Dynamic Stream Weights for Linear Dynamical Systems using Natural Evolution Strategies

Abstract: 

Multimodal data fusion is an important aspect of many object localization and tracking frameworks that rely on sensory observations from different sources. A prominent example is audiovisual speaker localization, where the incorporation of visual information has shown to benefit overall performance, especially in adverse acoustic conditions. Recently, the notion of dynamic stream weights as an efficient data fusion technique has been introduced into this field. Originally proposed in the context of audiovisual automatic speech recognition, dynamic stream weights allow for effective sensory-level data fusion on a per-frame basis, if reliability measures for the individual sensory streams are available. This study proposes a learning framework for dynamic stream weights based on natural evolution strategies, which does not require the explicit computation of oracle information. An experimental evaluation based on recorded audiovisual sequences shows that the proposed approach outperforms conventional methods based on supervised training in terms of localization performance.

up
0 users have voted:

Paper Details

Authors:
Submitted On:
10 May 2019 - 3:53am
Short Link:
Type:
Presentation Slides
Event:
Presenter's Name:
Christopher Schymura
Paper Code:
2227
Document Year:
2019
Cite

Document Files

icassp2019_schymura.pdf

(77)

Subscribe

[1] , "Learning Dynamic Stream Weights for Linear Dynamical Systems using Natural Evolution Strategies", IEEE SigPort, 2019. [Online]. Available: http://sigport.org/4283. Accessed: Aug. 03, 2020.
@article{4283-19,
url = {http://sigport.org/4283},
author = { },
publisher = {IEEE SigPort},
title = {Learning Dynamic Stream Weights for Linear Dynamical Systems using Natural Evolution Strategies},
year = {2019} }
TY - EJOUR
T1 - Learning Dynamic Stream Weights for Linear Dynamical Systems using Natural Evolution Strategies
AU -
PY - 2019
PB - IEEE SigPort
UR - http://sigport.org/4283
ER -
. (2019). Learning Dynamic Stream Weights for Linear Dynamical Systems using Natural Evolution Strategies. IEEE SigPort. http://sigport.org/4283
, 2019. Learning Dynamic Stream Weights for Linear Dynamical Systems using Natural Evolution Strategies. Available at: http://sigport.org/4283.
. (2019). "Learning Dynamic Stream Weights for Linear Dynamical Systems using Natural Evolution Strategies." Web.
1. . Learning Dynamic Stream Weights for Linear Dynamical Systems using Natural Evolution Strategies [Internet]. IEEE SigPort; 2019. Available from : http://sigport.org/4283