# BPG-Based Automatic Lossy Compression of Noisy Remote Sensing Images _by Vozel Benoit (IETR - MULTIP) - 2022.11.10_ ###### tags: `VAADER` `Seminar` ## Video {%youtube HRfsNOVGWX4%} ## Abstract _Work by Vladimir Lukin, Bogdan Kovalenko, Sergii Kryvenko, Victoriya Naumenko and Benoit Vozel_ We investigate lossy compression of noisy images. In this case, there are some peculiarities relating to noise filtering effect and analysis of compression technique performance. As a result of noise filtering due to lossy compression, the so-called optimal operation point (OOP) may exist. OOP is related to such parameters of a coder that quality of a compressed image is closer to the corresponding noise-free (true) image compared to quality of uncompressed (original, noisy) image where quality is characterized by a certain metric, conventional (PSNR) or visual one (PSNR-HVS-M, PSNR-HA, MDSI). If OOP for a given noisy image exists, it occurs expedient to automatically compress an image under interest in OOP or, at least, its neighbourhood. We show that it is, in general, possible to predict OOP existence (according to different metrics) before image compression. It is also possible to reasonably set the coder parameters where better portable graphics (BPG) coder is paid here the main attention. The extensions of the proposed approach to joint compression of three-component images are discussed. Image subsampling in color components is studied. The possibility of OOP existence for visual quality metrics PSNR-HA and MDSI is illustrated.
Last changed by  
 
Karol Desnos
96

Read more

[<span style="font-variant: small-caps;">Vaader</span> Seminars](https://hackmd.io/s/SJWHYpzPG#)

Organization contact [name= K. Desnos (kdesnos at insa-rennes.fr)] Mailing list: You can subscribe to a dedicated mailing-list, that will exclusively be used to send seminar and Ph.D. defense announcements. To subscribe, send an e-mail to sympa@insa-rennes.fr with the following subject line (with YOUR first and last names): subscribe vaader-seminars FirstName LastName Next Seminars 14.02.2023 - 10:15 INSA Build. 10, Room 229 & Zoom

2/6/2023
Le Véhicule Autonome et L’Intelligence Embarquée

by Yassin Kortli (IETR SIGNAL) - 2023.03.07 Abstract Le véhicule autonome est perçu comme un moyen de réduire le stress du conducteur et des passagers dans certaines situations de conduite et de lui libérer du temps. Le véhicule autonome, ce n’est pas une nouveauté. Aujourd’hui déjà, nos véhicules sont connectés, dotés de régulateurs de vitesse, ou de capteurs pour nous aider à nous garer, et même de systèmes d'aide à la conduite. L’intelligence artificielle (IA) permet d’améliorer plusieurs fonctionnalités des systèmes ADAS et de révolutionner les tâches de perception automatique et, dans notre cas, la capacité à identifier les lignes de la route, des panneaux routiers, des piétons, des cyclistes ou d’autres véhicules. Des caméras embarquant cette technologie sont déjà largement utilisées dans les véhicules actuellement commercialisés. L’utilisation de plusieurs capteurs et la fusion de données issues de ces capteurs demandent souvent des temps de calculs assez longs. Par conséquent, pour que le véhicule puisse évoluer à une vitesse suffisante, il serait nécessaire d’envisager une solution embarquée des systèmes développés.

2/6/2023
SCAPE: HW-Aware Clustering of Dataflow Actors for Tunable Scheduling Complexity.

by Ophélie Renaud (IETR - VAADER) - 2023.03.01 Abstract This paper introduces a fast method to generate high performance parallelized code from a dataflow specification of an application. Dataflow Models of Computation (MoCs) are efficient programming paradigms for expressing the parallelism of an application. Traditionally, mapping and scheduling methods for dataflow MoCs rely on complex graph's transformations to explicit their parallelism which can result in complex graph for embarrassingly parallel applications. For such applications, state-of-the-art mapping and scheduling techniques are prohibitively complex, while the exposed parallelism often exceeds the parallel processing capabilities of the target architecture. We propose SCAPE, an automated method to control the complexity of the pre-scheduling graph transformation by using information from the architecture and application models. By decreasing the complexity of the graph, the mapping scheduling task is accelerated at the potential expense of the produced schedule. Our method offers a limited and controlled decrease of the schedule quality while enabling mapping and scheduling execution time between 1 and 2 orders of magnitude faster than state-of-the-art techniques.

1/30/2023
Hierarchical Learning and Dummy Triplet Loss for Efficient Deepfake Detection

by Beuve Nicolas (IETR VAADER) - 2023.01.10 Video

1/27/2023
Read more from Karol Desnos
Published on HackMD