Conditional Random Fields for tubulin-microtubule segmentation in cryo-electron tomography

Abstract : Cryo-electron tomography allows 3D observation of biological specimens in their native and hydrated state at high spatial resolution (4-5 nanometers). Traditionally cryo-tomograms have very low signal-to-noise ratios and conventional image segmentation methods are limited yet. In this paper, we formulate the segmentation problem of both small tubulin aggregates and microtubules against the background as a two class labeling problem in the Conditional Random Field framework. In our approach, we exploit image patches to take into account spatial contexts and to improve robustness to noise. Because of the contrast anisotropy in the specimen thickness direction, each 2D section of the 3D tomogram is segmented separately with an optional update of reference patches. This method is evaluated on synthetic data and on cryo-electron tomograms of in vitro microtubules.
Document type :
Conference papers
Complete list of metadatas

Cited literature [22 references]  Display  Hide  Download

https://hal.inria.fr/hal-01103330
Contributor : Charles Kervrann <>
Submitted on : Friday, January 23, 2015 - 6:27:52 PM
Last modification on : Wednesday, May 16, 2018 - 11:23:52 AM
Long-term archiving on: Friday, April 24, 2015 - 10:55:57 AM

File

2014-icip-cryoseg.pdf
Files produced by the author(s)

Identifiers

Citation

Charles Kervrann, Sophie Blestel, Denis Chrétien. Conditional Random Fields for tubulin-microtubule segmentation in cryo-electron tomography. IEEE International Conference on Image Processing (top 10%” papers in ICIP 2014), Oct 2014, Paris, France. pp.2080 - 2084, ⟨10.1109/ICIP.2014.7025417⟩. ⟨hal-01103330⟩

Share

Metrics

Record views

297

Files downloads

293