GenRadar: Self-Supervised Probabilistic Camera Synthesis Based on Radar Frequencies

Autonomous systems require a continuous and dependable environment perception for navigation and decision-making, which is best achieved by combining different sensor types. Radar continues to function robustly in compromised circumstances in which cameras become impaired, guaranteeing a steady infl...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Carsten Ditzel, Klaus Dietmayer
Formato:	article
Lenguaje:	EN
Publicado:	IEEE 2021
Materias:	Radar signal processing computer vision sensor fusion deep learning machine learning variational autoencoder Electrical engineering. Electronics. Nuclear engineering TK1-9971
Acceso en línea:	https://doaj.org/article/dfdceb8ac3b54d1bbc8b5e82b6e69b33
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:dfdceb8ac3b54d1bbc8b5e82b6e69b33
record_format	dspace
spelling	oai:doaj.org-article:dfdceb8ac3b54d1bbc8b5e82b6e69b332021-11-18T00:08:07ZGenRadar: Self-Supervised Probabilistic Camera Synthesis Based on Radar Frequencies2169-353610.1109/ACCESS.2021.3120202https://doaj.org/article/dfdceb8ac3b54d1bbc8b5e82b6e69b332021-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/9570339/https://doaj.org/toc/2169-3536Autonomous systems require a continuous and dependable environment perception for navigation and decision-making, which is best achieved by combining different sensor types. Radar continues to function robustly in compromised circumstances in which cameras become impaired, guaranteeing a steady inflow of information. Yet, camera images provide a more intuitive and readily applicable impression of the world. This work combines the complementary strengths of both sensor types in a unique self-learning fusion approach for a probabilistic scene reconstruction in adverse surrounding conditions. After reducing the memory requirements of both high-dimensional measurements through a decoupled stochastic self-supervised compression technique, the proposed algorithm exploits similarities and establishes correspondences between both domains at different feature levels during training. Then, at inference time, relying exclusively on radio frequencies, the model successively predicts camera constituents in an autoregressive and self-contained process. These discrete tokens are finally transformed back into an instructive view of the respective surrounding, allowing to visually perceive potential dangers for important tasks downstream.Carsten DitzelKlaus DietmayerIEEEarticleRadar signal processingcomputer visionsensor fusiondeep learningmachine learningvariational autoencoderElectrical engineering. Electronics. Nuclear engineeringTK1-9971ENIEEE Access, Vol 9, Pp 148994-149042 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Radar signal processing computer vision sensor fusion deep learning machine learning variational autoencoder Electrical engineering. Electronics. Nuclear engineering TK1-9971
spellingShingle	Radar signal processing computer vision sensor fusion deep learning machine learning variational autoencoder Electrical engineering. Electronics. Nuclear engineering TK1-9971 Carsten Ditzel Klaus Dietmayer GenRadar: Self-Supervised Probabilistic Camera Synthesis Based on Radar Frequencies
description	Autonomous systems require a continuous and dependable environment perception for navigation and decision-making, which is best achieved by combining different sensor types. Radar continues to function robustly in compromised circumstances in which cameras become impaired, guaranteeing a steady inflow of information. Yet, camera images provide a more intuitive and readily applicable impression of the world. This work combines the complementary strengths of both sensor types in a unique self-learning fusion approach for a probabilistic scene reconstruction in adverse surrounding conditions. After reducing the memory requirements of both high-dimensional measurements through a decoupled stochastic self-supervised compression technique, the proposed algorithm exploits similarities and establishes correspondences between both domains at different feature levels during training. Then, at inference time, relying exclusively on radio frequencies, the model successively predicts camera constituents in an autoregressive and self-contained process. These discrete tokens are finally transformed back into an instructive view of the respective surrounding, allowing to visually perceive potential dangers for important tasks downstream.
format	article
author	Carsten Ditzel Klaus Dietmayer
author_facet	Carsten Ditzel Klaus Dietmayer
author_sort	Carsten Ditzel
title	GenRadar: Self-Supervised Probabilistic Camera Synthesis Based on Radar Frequencies
title_short	GenRadar: Self-Supervised Probabilistic Camera Synthesis Based on Radar Frequencies
title_full	GenRadar: Self-Supervised Probabilistic Camera Synthesis Based on Radar Frequencies
title_fullStr	GenRadar: Self-Supervised Probabilistic Camera Synthesis Based on Radar Frequencies
title_full_unstemmed	GenRadar: Self-Supervised Probabilistic Camera Synthesis Based on Radar Frequencies
title_sort	genradar: self-supervised probabilistic camera synthesis based on radar frequencies
publisher	IEEE
publishDate	2021
url	https://doaj.org/article/dfdceb8ac3b54d1bbc8b5e82b6e69b33
work_keys_str_mv	AT carstenditzel genradarselfsupervisedprobabilisticcamerasynthesisbasedonradarfrequencies AT klausdietmayer genradarselfsupervisedprobabilisticcamerasynthesisbasedonradarfrequencies
_version_	1718425255452606464

GenRadar: Self-Supervised Probabilistic Camera Synthesis Based on Radar Frequencies

Ejemplares similares