Fast virtual histology using X-ray in-line phase tomography: application to the 3D anatomy of maize developing seeds
- David Rousseau†1Email author,
- Thomas Widiez†2,
- Sylvaine Di Tommaso1,
- Hugo Rositi1,
- Jerome Adrien3,
- Eric Maire3,
- Max Langer1,
- Cécile Olivier1,
- Françoise Peyrin1 and
- Peter Rogowsky2
© Rousseau et al. 2015
Received: 30 August 2015
Accepted: 26 November 2015
Published: 18 December 2015
Despite increasing demand, imaging the internal structure of plant organs or tissues without the use of transgenic lines expressing fluorescent proteins remains a challenge. Techniques such as magnetic resonance imaging, optical projection tomography or X-ray absorption tomography have been used with various success, depending on the size and physical properties of the biological material.
X-ray in-line phase tomography was applied for the imaging of internal structures of maize seeds at early stages of development, when the cells are metabolically fully active and water is the main cell content. This 3D imaging technique with histology-like spatial resolution is demonstrated to reveal the anatomy of seed compartments with unequalled contrast by comparison with X-ray absorption tomography. An associated image processing pipeline allowed to quantitatively segment in 3D the four compartments of the seed (embryo, endosperm, nucellus and pericarp) from 7 to 21 days after pollination.
This work constitutes an innovative quantitative use of X-ray in-line phase tomography as a non-destructive fast method to perform virtual histology and extends the developmental stages accessible by this technique which had previously been applied in seed biology to more mature samples.
KeywordsX ray in-line phase tomography Image segmentation Virtual histology plant development maize seeds
Plant developmental biology and plant functional genetics have been strongly impacted by novel imaging techniques allowing to monitor in a non-destructive manner developmental processes or to compare wild-type and mutant structures. The global analysis of morphological, genetic and biochemical information not only deepens our understanding of developmental processes, it is also the basis for modelling (systems biology) and ultimately plant breeding. Over the past 10 years confocal microscopy has become the tool of choice to obtain 3D and 4D information on the development of organ shape and the dynamics of gene expression patterns . However, observations by confocal microscopy necessitate the presence of fluorescent dyes or proteins, generally achieved by plant transformation. It also requires a depth of the structure smaller than 50 µm for 3D reconstruction. The production of transgenic reporter lines by plant transformation remains very cumbersome or is simply not feasible in the majority of plant species and cannot be applied to hundreds of accessions that need to be phenotyped for quantitative genetics studies. Consequently the interest in alternative methods applicable to any type of non-transformed plant organ or tissue is ever increasing. Techniques such as magnetic resonance imaging, positon emission tomography, optical coherence tomography or X-ray absorption tomography have been used with various success depending on the size and physical properties of the biological material [2–7].
In this framework, X-ray absorption radiography or tomography is a widely used non-destructive method specially on dry seeds (see for instance [8–11]). The contrast in atomic number, at the physical origin of contrast in X-ray imaging, is strong in dry tissues because of the presence of air networks. However, in metabolically active “wet” tissues, in which the cell content consists essentially of water, the contrast in atomic number dramatically drops and consequently limits the domain of applicability of X-ray imaging for the characterization of fresh tissues. A possible way to enhance the contrast in X-ray imaging is to use contrast agents. This approach, very common in the biomedical domain, has only recently been introduced in plant sciences [12–14] and the selection of contrast agents and their penetration into wet plant tissues remains an open problem. An alternative way to enhance the contrast in X-ray imaging is to move from absorption tomography to in-line phase contrast tomography with coherent X-ray sources accessible via synchrotron radiation.
In X-ray in-line phase contrast imaging, the refraction of a partially coherent X-ray beam by an object of interest slightly modifies the original wave front profile. These variations result in changes in the locally transmitted intensity of the wave which contains quantitative information on the phase shift induced by the object. Due to its enhanced contrast over standard attenuation imaging, X-ray in-line phase contrast imaging is receiving more and more attention in medicine and biology (see for a recent review in medicine  and plant sciences [14, 16]). The enhanced contrast of X-ray in-line phase contrast tomography has so far been applied to various plant organs (see for the most recent review ), including the characterization of dry seeds [17–19] and developping wet tissues [20–22].
In previous work using X-ray phase contrast for imaging of developing wet plant tissues, the main focus was on the specific contrast due to local void spaces inside the plant tissues (seed [19, 22], fruit , leaf ) and their possible functional role on the physiology of these tissues or during ulterior imbibition. Distinctly, in this manuscript, the objective of phase contrast imaging is to exploit contrast between adjacent tissues to perform the global anatomical 3D segmentation of the compartments of a developing organism at various stages of development. The developmental stages of the seed considered in this manuscript precede the single mid-filling stage considered in the first publication which demonstrated the presence of the embryo in wet tissue of developing seeds . We specifically focus on the early stages of seed development because they are characterized by major developmental events, during which some seed compartements grow while other disappear. These phenomenona are displayed in 3D after a detailed quantitative analysis of the recorded contrasts based on a segmentation process, whereas the previous demonstration of X-ray phase contrast imaging on seeds were illustrated with qualitative maximum intensity projections, or single thresholded 3D views colored for the eye with lookup tables. In this work, we demonstrate for the first time to our knowledge the value of X-ray in-line phase contrast imaging to developing seeds with a quantitative approach including the assessment of the observed phase contrasts and their use for segmentation purposes.
X-ray absorption tomography
X-ray in-line phase contrast image acquisition and reconstruction
To visualise the overall morphology of embryo, endosperm, nucellus and pericarp by X-ray in-line phase contrast image different maize seeds were harvested at 7, 9, 12 and 21 days after pollination (DAP) and imaged individually. These stages were chosen because they cover the transition from morphogenesis (1 to 15 DAP) to seed filling (15 to 45 DAP) and include the nonmonotonic growth and disappearance of the nucellus together with the growth of the embryo and the endosperm from 0.5 mm/1.2 mm at 7 DAP up to 4/12 mm at 21 DAP (for a review see ).
All the seeds were imaged on the synchrotron radiation beamline ID19 of the European Synchrotron Radiation Facility (ESRF) with a beam of energy 17.6 keV and a detector positioned at 1 m from the sample in order to enhance propagation phase contrast. The 2D projections of a stack of 1000 virtual slices were acquired in 20 min. The pixel size was set to 5 µm by the optics of the detector. After applying Paganin’s algorithm for phase retrieval, the phase maps were used as input to a tomographic reconstruction algorithm and ring artefacts were removed.
Conventional histology was performed on different maize seeds taken at the same dates after pollination to correlate the size and shape of kernel compartments with zones of different contrast in X-ray in-line phase tomography (Fig. 1b). Despite cautious care, part of the tissues was retracted and/or damaged during the fixation, inclusion, sectioning and staining steps, as this is very often the case. This is why a 3D reconstruction with non-destructive 3D microscopy such as X-ray in-line tomography is expected to bring a specific added value by comparison with conventional histology.
Phase contrast assessment
Contrast measured by the Fisher ratio \(F_r\) between compartment of the maize seed with X-ray in-line phase tomography at 7, 9, 12, and 21 DAP
\(F_r\) at 7 DAP
\(F_r\) at 9 DAP
\(F_r\) at 12 DAP
\(F_r\) at 21 DAP
Comparison of global measures of length between X-ray in-line phase tomography and conventional histology of different maize seed compartment at 7 DAP
X-ray in-line phase tomography (%)
Conventional histology (%)
In this article we demonstrate the feasibility of 3D imaging of maize embryo and endosperm morphology inside the seed coat with X-ray imaging during early developmental stages when both are metabolically highly active and the cells essentially filled with water. The contrasts are too poor for segmentation in conventional absorption X-ray but large enough with X-ray in-line phase contrast imaging. The coherent beam of X-ray allows to visualize and segment the four main compartments of the seed during morphogenesis without having to resort to the use of contrast agents. The cumulated data from four different developmental stages establish for the first time a 3D dynamic picture of the growth of embryo and endosperm and the concomitant reduction and nutritional recycling of the seed coat (nucellus and pericarp). The sizes and shapes of the compartments obtained by non-destructive 3D virtual histology were compared with success to knowledge from the literature  and data obtained in parallel by conventional destructive histology. The acquisition time was 20 min for a single seed with X-ray in-line phase contrast imaging. This is much shorter than the time requested for the realization of fixation, inclusion, sectioning and scanning of a single seed with destructive histology. X-ray in-line phase contrast imaging can thus be considered as fast by comparison with destructive histology which is the most current reference for anatomical atlas in plant biology. Comparison between histology and X-ray phase contrast imaging was also proposed in . This was at the cellular scale to demonstrate that the phase contrast was correlated with the presence of gaz in the tissue. In this work, thanks to the use of automated histological scanning system, we performed a global comparison at the scale of the entire seed which shows good agreement on the size of the measured compartments in the seed by the destructive histological technique and non destructive 3D imaging by X-ray phase contrast tomography. For this successful outcome a complete image processing pipeline had to be developed, which described how to take benefit of the contrast between the various subparts of the maize seed. We identified and quantitatively assessed the level of difficulty for the segmentation of the different compartments of the seed, the most challenging being the segmentation of the embryo before 12 DAP. Such challenges are now open for studies with more advanced image processing tools.
The use of X-ray in-line phase tomography, not only promises to outperform conventional histology in terms of time of acquisition and precision for 3D measurement on native samples. It also enables, thanks to its high sensitivity to interfaces, to highlight specific areas that did not attract particular attention in conventional histology and are not referenced in atlases of developmental biology. This is for instance the case, with the very textured part located after the outer part of the nucellus at DAP 21 visible in Fig. 3. Thanks to the quantitative comparison with conventional histology, we are in position to state that this textured region is part of the pericarp. The texture can physically originate from multiscale interfaces, which would be compatible with a porous air-tissues interface. This conjecture is now open for further investigations.
This proof of concept for quantitative use of X-ray in-line phase contrast for developing seeds also pinpointed several limitations of this method. The most obvious one is the lack of cellular resolution when compared to conventional histology. This could possibly be overcome by higher spatial resolution imaging a small part of the sample in a local tomography mode inside the same type of seeds. In addition, the obtained results were established in non-natural conditions since for room constrains, the grains were extracted from their ear during the 20 min acquisition. Also, because of the limitation of the temporal access to the synchrotron beam, it was not possible to realize a time lapse experiment monitoring the morphogenesis of a single seed throughout different developmental stages. Consequently, the knowledge of the possible impact of the X-ray dose on maize seed development was not accessible. New development in coherent X-ray sources accessible to standard lab environments may open access to such complementary experiments. Desktop phase contrast micro-CT devices are beginning to develop. The usual solution is to use interferometry which requires several scans of the sample with adapted gratings yielding to long scan times, and is generally limited to a spatial resolution of a few micrometers, thus much higher than what was used here. The other possibility is to exploit edge enhancement effect with microfocus sources but such systems are known to offer less monochromaticity and spatial coherence than the synchrotron facility. Consequently, the assumption required for the validity of the phase reconstruction algorithm together with the assumption required for the tomographic reconstruction are not satisfied. This contrast is thus expected to be of lower quality with a desktop phase contrast micro-CT than with a synchrotron facility. To our knowledge phase CT on a desktop system has so far never been demonstrated. Although not yet accessible in routine, synchrotron radiations are now available worldwide as national or transational public facilities with an acquisition time of 20 min enabling to realize the 3D segmentation of some 72 seeds during a typical 24H run. Such acquisition campaigns could now be undertaken thanks to the methodological details on image acquisition and post-processing provided in our article for fast 3D virtual histology compared to the conventional destructive 2D histology.
The domain of 3D imaging techniques with cellular resolution and field of view compatible with the smallest entire plant organisms, i.e. seed or seedling, is progressing very rapidly . It would be interesting to develop a similar approach to the one developed in this article for developmental imaging with other imaging techniques. Optical coherence tomography, another 3D virtual histology, was for instance unfruitful for maize seed screening because of too large thickness of the tissues while it is very adapted to the screening of the development of seedlings of Arabidopsis as demonstrated in . This illustrates that the knowledge of which 3D microscopy technique is suited for which biological question is open for further investigations such as the one presented in this article for X-ray in-line phase contrast imaging applied to maize seed morphogenesis.
Maize plants of genotype A188 were grown in a greenhouse with a 16 h illumination period (100 \(\rm W/m^2\)) at 24/19 °C (day/night) and without control of the relative humidity. At maturity the plants underwent controlled self pollination. The substrate, watering and fertilizing were as described previously .
Maize seed sides were cut for better penetration of the fixative and seeds fixed with paraformaldehyde 48 h at 4 °C, dehydrated in an ethanol series and included in paraplast with a Leica TP1020 benchtop tissue processor. 10 µm thick serial sections were obtained with a Leica RM2235 rotary manual microtome. Histoclear was used to remove the wax. After rehydration in a decreasing ethanol series the tissues were stained by periodic acid and Schiff’s reagent (Sigma). Automated acquisitions were realized on a slide scan Axioscan Z1 (Zeiss). It took approximately 8 h to acquire the all set of serial sections representing an entire seed. The slices were automatically positioned in their respective order of apparition to constitute a stack of the seed.
X-ray absorption tomography
Self pollinated ears were separated from the plant and soaked up to 48 h in either Gadoteric acid dye (Dotarem) at 0.5 nmol/ml or Iobitribol dye (Xenetix) at 300 nmol/ml. Ear sections containing 5 rows of kernels were imaged on a GE Phoenix v—tome—x equiped with a nanofocus X-ray tube and a Varian paxscan detector. The tomograph is described in more details in . It was operated with a 80kV acceleration voltage using a tungsten transmission target with a 280 µm current. The number of projections was 900, each radiograph was an average of 3 exposures of 333 ms each to reduce the noise. The voxel size was set at 30 µm. The acquisition time was 20 min.
X-ray in-line phase contrast image acquisition and reconstruction
For X-ray in-line phase tomography, raw projection of the seeds were acquired on the synchrotron radiation beamline ID19 of the European Synchrotron Radiation Facility (ESRF) in Grenoble, France, with an experimental image acquisition characterized by a beam of energy 17.6 keV, selected from undulator radiation using Al filters. The X-ray beam transmitted through the specimen is acquired on a detector using a LuAg scintillator screen, visible light optics, and a 2048 × 2048 CCD detector with 32 bit resolution. The detector was positioned at 1 m from the sample for in-line phase-contrast imaging. The acquisition duration was 20 min per seed. Phase retrieval was performed from a single phase-contrast image at each projection angle, using Paganin’s method  where a low-pass filter followed by a logarithmic operator is applied. The term ‘in-line’ in this context of X-ray phase contrast tomography is a term used in physics because the method used to capture the phase information requires only a single projection by angle while other physical methods based on gratings, or interference demand more acquisitions. The so-called \(\delta /\beta\) coefficient which governs the cut-off frequency of this filter was chosen at \(\delta /\beta = 300\) following the approach given in . After phase retrieval, the phase maps were used as input to a 3D parallel-beam tomographic reconstruction algorithm based on the filtered back projection algorithm described in . The resulting 3D volume is a stack of 1000 slices of 2048 × 2048 voxels and slice thickness equal to pixel size (isotropic voxels) 5 µm. For each scan, this procedure provided a reconstructed 3D volume which suffers from spatially circular noise typical of reconstruction artefacts in tomography caused by differences in the individual pixel responses of the detector. These ring artefacts were removed with the nonlinear median-based filter algorithm of .
DR conceived and designed this study, carried out the acquisition and reconstruction of phase-contrast tomography, conceived and supervised the realization of the image processing pipeline, wrote and revised the manuscript. TW contributed in the design of the study, took care of maize ears’s fertilization, realized the treatments with contrast agents, helped in the acquisition of tomographies, realized histology experiments and contributed in the revision of the manuscript. SDT realized the image processing pipeline for phase contrast tomography and helped in the realization of histology experiments and contributed in the production of figures of the manuscript. HR, JA, EM, ML, CO and FP helped and advised in the acquisition of X-Ray tomographies. PR conceived and designed this study, helped in the acquisition of phase contrast tomography and contributed in the writing and revision of the manuscript. All authors read and approved the final manuscript.
This work was supported by the European Synchrotron Research Facility (ESRF, project LS-2290) through the allocation of beam time and by the European Phenotyping Platform Network program (project ID2MDS). Authors thank Dr Craig Sturrock at the Hounsfield Facility, University of Nottingham (UK) for useful preliminary microCT experiments and discussions. The Hounsfield Facility received funding from the European Research Council (FutureROOTS), BBSRC and The Wolfson Foundation. Authors thank Ghislaine Gendrot at RDP, INRA for unvaluable technical support in the preparation of the biological samples and the Lyon Bio Image platform for assistance in the use of the automated scanner of histological slices.
The authors declare that they have no competing interests.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Fischer RS, Wu Y, Kanchanawong P, Shroff H, Waterman CM. Microscopy in 3D: a biologist’s toolbox. Trends Cell Biol. 2011;21(12):682–91.PubMedPubMed CentralView ArticleGoogle Scholar
- Dean RJ, Stait-Gardner T, Clarke SJ, Rogiers SY, Bobek G, Price WS. Use of diffusion magnetic resonance imaging to correlate the developmental changes in grape berry tissue structure with water diffusion patterns. Plant methods. 2014;10(1):35.PubMedPubMed CentralView ArticleGoogle Scholar
- Metzner R, Eggert A, van Dusschoten D, Pflugfelder D, Gerth S, Schurr U, Uhlmann N, Jahnke S. Direct comparison of mri and X-ray CT technologies for 3D imaging of root systems in soil: potential and challenges for root trait quantification. Plant Methods. 2015;11(1):17.PubMedPubMed CentralView ArticleGoogle Scholar
- Jahnke S, Menzel MI, Van Dusschoten D, Roeb GW, Bühler J, Minwuyelet S, Blümler P, Temperton VM, Hombach T, Streun M. Combined mri-pet dissects dynamic changes in plant structures and functions. Plant J. 2009;59(4):634–44.PubMedView ArticleGoogle Scholar
- Suuronen J-P, Peura M, Fagerstedt K, Serimaa R. Visualizing water-filled versus embolized status of xylem conduits by desktop X-ray microtomography. Plant Methods. 2013;9(1):11.PubMedPubMed CentralView ArticleGoogle Scholar
- Dhondt S, Vanhaeren H, Van Loo D, Cnudde V, Inzé D. Plant structure visualization by high-resolution X-ray computed tomography. Trends Plant Sci. 2010;15(8):419–22.PubMedView ArticleGoogle Scholar
- Stuppy WH, Maisano JA, Colbert MW, Rudall PJ, Rowe TB. Three-dimensional analysis of plant structure using high-resolution X-ray computed tomography. Trends Plant Sci. 2003;8(1):2–6.PubMedView ArticleGoogle Scholar
- Rousseau D, Chéné Y, Belin E, Semaan G, Trigui G, Boudehri K, Franconi F, Chapeau-Blondeau F. Multiscale imaging of plants: current approaches and challenges. Plant Methods. 2015;11(1):6.PubMedPubMed CentralView ArticleGoogle Scholar
- Belin E, Rousseau D, Léchappé J, Langlois-Meurinne M, Dürr C. Rate-distortion tradeoff to optimize high-throughput phenotyping systems. application to X-ray images of seeds. Comp Elect Agricul. 2011;77(2):188–94.View ArticleGoogle Scholar
- Dell’Aquila A. Development of novel techniques in conditioning, testing and sorting seed physiological quality. Seed Sci Technol. 2009;37(3):608–24.View ArticleGoogle Scholar
- Bino R, Aartse J, Van Der Burg W. Non-destructive x-ray analysis of arabidopsis embryo mutants. Seed Sci Res. 1993;3(03):167–70.View ArticleGoogle Scholar
- Blonder B, Carlo F, Moore J, Rivers M, Enquist BJ. X-ray imaging of leaf venation networks. New Phytol. 2012;196(4):1274–82.PubMedView ArticleGoogle Scholar
- Staedler YM, Masson D, Schönenberger J. Plant tissues in 3d via x-ray tomography: simple contrasting methods allow high resolution imaging. PloS one. 2013; 8(9)Google Scholar
- Karunakaran C, Lahlali R, Zhu N, Webb AM, Schmidt M, Fransishyn K, Belev G, Wysokinski T, Olson J, Cooper DM et al. Factors influencing real time internal structural visualization and dynamic process monitoring in plants using synchrotron-based phase contrast x-ray imaging. Sci Reports. 2015; 5.Google Scholar
- Bravin A, Coan P, Suortti P. X-ray phase-contrast imaging: from pre-clinical applications towards clinics. Phys Med Biol. 2013;58(1):1.View ArticleGoogle Scholar
- Lombi E, Susini J. Synchrotron-based techniques for plant and soil science: opportunities, challenges and future perspectives. Plant Soil. 2009;320(1–2):1–35.View ArticleGoogle Scholar
- Friis EM, Marone F, Pedersen KR, Crane PR, Stampanoni M. Three-dimensional visualization of fossil flowers, fruits, seeds, and other plant remains using synchrotron radiation x-ray tomographic microscopy (srxtm): new insights into cretaceous plant diversity. J Paleontol. 2014;88(4):684–701.View ArticleGoogle Scholar
- Moreau J-D, Cloetens P, Gomez B, Daviero-Gomez V, Néraudeau D, Lafford TA, Tafforeau P. Multiscale 3d virtual dissections of 100-million-year-old flowers using x-ray synchrotron micro-and nanotomography. Microscopy Microanal. 2014;20(01):305–12.View ArticleGoogle Scholar
- Cloetens P, Mache R, Schlenker M, Lerbs-Mache S. Quantitative phase tomography of arabidopsis seeds reveals intercellular void network. Proc Natl Acad Sci USA. 2006;103(39):14626–30.PubMedPubMed CentralView ArticleGoogle Scholar
- Lauridsen T, Glavina K, Colmer TD, Winkel A, Irvine S, Lefmann K, Feidenhans R, Pedersen O. Visualisation by high resolution synchrotron X-ray phase contrast micro-tomography of gas films on submerged superhydrophobic leaves. J Struct Biol. 2014;188(1):61–70.PubMedView ArticleGoogle Scholar
- Verboven P, Kerckhofs G, Mebatsion HK, Ho QT, Temst K, Wevers M, Cloetens P, Nicolaï BM. Three-dimensional gas exchange pathways in pome fruit characterized by synchrotron x-ray computed tomography. Plant Physiol. 2008;147(2):518–27.PubMedPubMed CentralView ArticleGoogle Scholar
- Verboven P, Herremans E, Borisjuk L, Helfen L, Ho QT, Tschiersch H, Fuchs J, Nicolaï BM, Rolletschek H. Void space inside the developing seed of brassica napus and the modelling of its function. New Phytol. 2013;199(4):936–47.PubMedPubMed CentralView ArticleGoogle Scholar
- Sosso D, Javelle M, Rogowsky P. Embryogenesis and early grain development. Adv Maize. 2011; 3.Google Scholar
- Pouvreau B, Baud S, Vernoud V, Morin V, Py C, Gendrot G, Pichon J-P, Rouster J, Paul W, Rogowsky PM. Duplicate maize wrinkled1 transcription factors activate target genes involved in seed oil biosynthesis. Plant Physiol. 2011;156(2):674–86.PubMedPubMed CentralView ArticleGoogle Scholar
- Buffiere J-Y, Maire E, Adrien J, Masse J-P, Boller E. In situ experiments with X ray tomography: an attractive tool for experimental mechanics. Exp Mechan. 2010;50(3):289–305.View ArticleGoogle Scholar
- Paganin D, Mayo S, Gureyev TE, Miller PR, Wilkins SW. Simultaneous phase and amplitude extraction from a single defocused image of a homogeneous object. J Microscopy. 2002;206(1):33–40.View ArticleGoogle Scholar
- Rositi H, Frindel C, Wiart M, Langer M, Olivier C, Peyrin F, Rousseau D. Computer vision tools to optimize reconstruction parameters in X-ray in-line phase tomography. Phys Med Biol. 2014;59(24):7767–75.PubMedView ArticleGoogle Scholar
- Salomé M, Peyrin F, Cloetens P, Odet C, Laval-Jeantet A-M, Baruchel J, Spanne P. A synchrotron radiation microtomography system for the analysis of trabecular bone samples. Med Phys. 1999;26(10):2194–204.PubMedView ArticleGoogle Scholar
- Boin M, Haibel A. Compensation of ring artefacts in synchrotron tomographic images. Optics Express. 2006;14(25):12071–5.PubMedView ArticleGoogle Scholar
- Caselles V, Kimmel R, Sapiro G. Geodesic active contours. Int J Comp Vision. 1997;22(1):61–79.View ArticleGoogle Scholar
- Dietenbeck T, Alessandrini M, Friboulet D, Bernard O. Creaseg: a free software for the evaluation of image segmentation algorithms based on level-set. In: 17th IEEE International Conference On Image Processing (ICIP), 2010. p. 665–668.Google Scholar