Reconstructing Objects from Noisy Images at Low Resolution
Publikation: Bidrag til bog/antologi/rapport › Konferencebidrag i proceedings › Forskning › fagfællebedømt
We study the problem of reconstructing small objects from their low-resolution images, by modelling them as r-regular objects. Previous work shows how the boundary constraints imposed by r-regularity allows bounds on estimation error for noise-free images. In order to utilize this for noisy images, this paper presents a graph-based framework for reconstructing noise-free images from noisy ones. We provide an optimal, but potentially computationally demanding algorithm, as well as a greedy heuristic for reconstructing noise-free images of r-regular objects from images with noise.
Originalsprog | Engelsk |
---|---|
Titel | Graph-Based Representations in Pattern Recognition - 12th IAPR-TC-15 International Workshop, GbRPR 2019, Proceedings |
Redaktører | Donatello Conte, Jean-Yves Ramel, Pasquale Foggia |
Forlag | Springer |
Publikationsdato | 2019 |
Sider | 204-214 |
ISBN (Trykt) | 9783030200800 |
DOI | |
Status | Udgivet - 2019 |
Begivenhed | 12th IAPR-TC15 Workshop on Graph-Based Representations in Pattern Recognition, GbRPR 2019 - Tours, Frankrig Varighed: 19 jun. 2019 → 21 jun. 2019 |
Konference
Konference | 12th IAPR-TC15 Workshop on Graph-Based Representations in Pattern Recognition, GbRPR 2019 |
---|---|
Land | Frankrig |
By | Tours |
Periode | 19/06/2019 → 21/06/2019 |
Navn | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
---|---|
Vol/bind | 11510 LNCS |
ISSN | 0302-9743 |
ID: 227334305