Title: | Investigation on Encoder-Decoder Networks for Segmentation of Very Degraded X-Ray CT Tomograms |
Authors: | Dulau, Idris Beurton-Aimar, Marie Hwu, Yeykuang Recur, Benoit |
Citation: | WSCG 2023: full papers proceedings: 1. International Conference in Central Europe on Computer Graphics, Visualization and Computer Vision, p. 11-19. |
Issue Date: | 2023 |
Publisher: | Václav Skala - UNION Agency |
Document type: | konferenční příspěvek conferenceObject |
URI: | http://hdl.handle.net/11025/54394 |
ISBN: | 978-80-86943-32-9 |
ISSN: | 2464–4617 (print) 2464–4625 (CD/DVD) |
Keywords: | rentgenová nanotomografie;segmentace;hluboké učení;zobrazování mozku |
Keywords in different language: | X-Ray nano-tomography;segmentation;deep-learning;brain imaging |
Abstract in different language: | Field of View (FOV) Nano-CT X-Ray synchrotron imaging is used for acquiring brain neuronal features from Golgi-stained bio-samples. It theoretically requires a large number of acquired data for compensating CT recon struction noise and artefacts (both reinforced by the sparsity of brain features). However reducing the number of radiographs is essential in routine applications but it results to degraded tomograms. In such a case, traditional segmentation techniques are no longer able to distinguish neuronal structures from surrounding noise. Thus, we investigate several deep-learning networks to segment brain features from very degraded tomograms. We focus on encoder-decoder networks and define new ones addressing specifically our application. We demonstrate that some networks wildly outperform traditional segmentation and discuss the superiority of the proposed networks. |
Rights: | © Václav Skala - UNION Agency |
Appears in Collections: | WSCG 2023: Full Papers Proceedings |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
D73-full.pdf | Plný text | 2,81 MB | Adobe PDF | View/Open |
Please use this identifier to cite or link to this item:
http://hdl.handle.net/11025/54394
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.