Details

Automatic segmentation of intracellular compartments in volumetric electron microscopy data
ID Žerovnik Mekuč, Manca (Author), ID Marolt, Matija (Mentor) More about this mentor... This link opens in a new window

.pdfPDF - Presentation file, Download (37,93 MB)
MD5: D5505E7CD58963F26D53DE745AF70F4F

Abstract
In recent years, electron microscopy has enabled the acquisition of volumetric data with resolving power to directly observe the ultrastructure of intracellular compartments. New insights and knowledge about cellular processes offered by such data require comprehensive analysis, which is limited by the time-consuming manual segmentation and reconstruction methods. We present methods for automatic segmentation, reconstruction and analysis of intracellular compartments from volumetric data obtained by dual-beam electron microscopy which combines a focused ion beam and a scanning electron microscope (FIB-SEM). In particular, we focus on the segmentation of mitochondria, endolysosomes, fusiform vesicles and the Golgi apparatus, the reconstruction of mitochondria and fusiform vesicles and the morphological analysis of the reconstructed mitochondria. The segmentation methods are based on supervised deep learning and include mechanisms that reduce the impact of dependencies in the input data, artifacts and annotation errors, while the reconstruction methods are based on more traditional image processing techniques that incorporate knowledge about the morphology of the structures. We present UroCell, a new publicly available volumetric electron microscopy (EM) dataset on which we evaluated our methods. The evaluation revealed that the proposed methods for segmentation, reconstruction and analysis achieve higher accuracy compared to existing state-of-the-art methods.

Language:English
Keywords:segmentation, reconstruction, deep learning
Work type:Doctoral dissertation
Typology:2.08 - Doctoral Dissertation
Organization:FRI - Faculty of Computer and Information Science
Year:2025
PID:20.500.12556/RUL-167355 This link opens in a new window
COBISS.SI-ID:227610883 This link opens in a new window
Publication date in RUL:17.02.2025
Views:1101
Downloads:332
Metadata:XML DC-XML DC-RDF
:
Copy citation
Share:Bookmark and Share

Secondary language

Language:Slovenian
Title:Avtomatska segmentacija celičnih predelkov v volumetričnih podatkih pridobljenih z elektronskim mikroskopom
Abstract:
V zadnjih letih je napredek na področju elektronske mikroskopije omogočil pridobivanje volumetričnih podatkov s tako dobro ločljivostjo, da je možno neposredno opazovanje struktur celičnih predelkov. S tem se je odprla možnost novih raziskav in odkritij, vendar ta zahteva celostno analizo podatkov, ki pa je zaradi zahtevnosti in počasnosti ročne segmentacije in rekonstrukcije struktur v podatkih zelo omejena. Da bi zamudno ročno delo pohitrili, v delu predstavljamo metode za avtomatsko segmentacijo in rekonstrukcijo medceličnih predelkov, pridobljenih z metodo fokusiranega ionskega snopa in vrstične elektronske mikroskopije (FIB-SEM). Predstavljamo metode za avtomatsko segmentacijo mitohondrijev, endolizosomov, fuziformnih veziklov in Golgijevega aparata, rekonstrukcijo mitohondrijev in fuziformnih veziklov ter morfološko analizo rekonstruiranih mitohondrijev. Metode segmentacije temeljijo na nadzorovanem globokem učenju in vključujejo mehanizme, ki zmanjšujejo vpliv odvisnosti v vhodnih podatkih, artefaktov in napak pri anotaciji, medtem ko so metode rekonstrukcije osnovane na bolj tradicionalnih tehnikah obdelave slik in vključujejo znanje o morfologiji struktur. V delu predstavljamo tudi novo javno dostopno podatkovno zbirko, ki se imenuje UroCell in na kateri ovrednotimo naše pristope ter jih primerjamo s podobnimi pristopi s področja. Na koncu pokažemo, da predstavljene metode za segmentacijo, rekonstrukcijo in analizo dosegajo natančnost, ki je boljša od nam znanih obstoječih metod, ki naslavljajo podobne probleme.

Keywords:segmentacija, rekonstrukcija, globoko učenje

Similar documents

Similar works from RUL:
Similar works from other Slovenian collections:

Back