Technical note: A new automated radiolarian image acquisition, stacking, processing, segmentation and identification workflow
<p>Identification of microfossils is usually done by expert taxonomists and requires time and a significant amount of systematic knowledge developed over many years. These studies require manual identification of numerous specimens in many samples under a microscope, which is very tedious and...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Copernicus Publications
2020-12-01
|
| Series: | Climate of the Past |
| Online Access: | https://cp.copernicus.org/articles/16/2415/2020/cp-16-2415-2020.pdf |
| _version_ | 1818566112601702400 |
|---|---|
| author | M. Tetard R. Marchant R. Marchant G. Cortese Y. Gally T. de Garidel-Thoron L. Beaufort |
| author_facet | M. Tetard R. Marchant R. Marchant G. Cortese Y. Gally T. de Garidel-Thoron L. Beaufort |
| author_sort | M. Tetard |
| collection | DOAJ |
| description | <p>Identification of microfossils is usually done by expert taxonomists and requires time and a significant amount of systematic knowledge developed over many years. These studies require manual identification of numerous specimens in many samples under a microscope, which is very tedious and time-consuming. Furthermore, identification may differ between operators, biasing reproducibility. Recent technological advances in image acquisition, processing and recognition now enable automated procedures for this process, from microscope image acquisition to taxonomic identification.</p>
<p>A new workflow has been developed for automated radiolarian image acquisition, stacking, processing, segmentation and identification. The protocol includes a newly proposed methodology for preparing radiolarian microscopic slides. We mount eight samples per slide, using a recently developed 3D-printed decanter that enables the random and uniform settling of particles and minimizes the loss of material. Once ready, slides are automatically imaged using a transmitted light microscope. About 4000 specimens per slide (500 per sample) are captured in digital images that include stacking techniques to improve their focus and sharpness. Automated image processing and segmentation is then performed using a custom plug-in developed for the ImageJ software. Each individual radiolarian image is automatically classified by a convolutional neural network (CNN) trained on a Neogene to Quaternary radiolarian database (currently 21 746 images, corresponding to 132 classes) using the ParticleTrieur software.</p>
<p>The trained CNN has an overall accuracy of about 90 %. The whole procedure, including the image acquisition, stacking, processing, segmentation and recognition, is entirely automated via a LabVIEW interface, and it takes approximately 1 h per sample. Census data count and classified radiolarian images are then automatically exported and saved. This new workflow paves the way for the analysis of long-term, radiolarian-based palaeoclimatic records from siliceous-remnant-bearing samples.</p> |
| first_indexed | 2024-12-14T01:49:35Z |
| format | Article |
| id | doaj.art-b4d4596c35384ce09c938f7c92e3be1f |
| institution | Directory Open Access Journal |
| issn | 1814-9324 1814-9332 |
| language | English |
| last_indexed | 2024-12-14T01:49:35Z |
| publishDate | 2020-12-01 |
| publisher | Copernicus Publications |
| record_format | Article |
| series | Climate of the Past |
| spelling | doaj.art-b4d4596c35384ce09c938f7c92e3be1f2022-12-21T23:21:25ZengCopernicus PublicationsClimate of the Past1814-93241814-93322020-12-01162415242910.5194/cp-16-2415-2020Technical note: A new automated radiolarian image acquisition, stacking, processing, segmentation and identification workflowM. Tetard0R. Marchant1R. Marchant2G. Cortese3Y. Gally4T. de Garidel-Thoron5L. Beaufort6Aix Marseille Univ, CNRS, IRD, Coll France, INRAE, CEREGE, Aix-en-Provence, FranceAix Marseille Univ, CNRS, IRD, Coll France, INRAE, CEREGE, Aix-en-Provence, Francepresent address: School of Electrical Engineering and Robotics, Queensland University of Technology, Brisbane, AustraliaGNS Science, Lower Hutt, New ZealandAix Marseille Univ, CNRS, IRD, Coll France, INRAE, CEREGE, Aix-en-Provence, FranceAix Marseille Univ, CNRS, IRD, Coll France, INRAE, CEREGE, Aix-en-Provence, FranceAix Marseille Univ, CNRS, IRD, Coll France, INRAE, CEREGE, Aix-en-Provence, France<p>Identification of microfossils is usually done by expert taxonomists and requires time and a significant amount of systematic knowledge developed over many years. These studies require manual identification of numerous specimens in many samples under a microscope, which is very tedious and time-consuming. Furthermore, identification may differ between operators, biasing reproducibility. Recent technological advances in image acquisition, processing and recognition now enable automated procedures for this process, from microscope image acquisition to taxonomic identification.</p> <p>A new workflow has been developed for automated radiolarian image acquisition, stacking, processing, segmentation and identification. The protocol includes a newly proposed methodology for preparing radiolarian microscopic slides. We mount eight samples per slide, using a recently developed 3D-printed decanter that enables the random and uniform settling of particles and minimizes the loss of material. Once ready, slides are automatically imaged using a transmitted light microscope. About 4000 specimens per slide (500 per sample) are captured in digital images that include stacking techniques to improve their focus and sharpness. Automated image processing and segmentation is then performed using a custom plug-in developed for the ImageJ software. Each individual radiolarian image is automatically classified by a convolutional neural network (CNN) trained on a Neogene to Quaternary radiolarian database (currently 21 746 images, corresponding to 132 classes) using the ParticleTrieur software.</p> <p>The trained CNN has an overall accuracy of about 90 %. The whole procedure, including the image acquisition, stacking, processing, segmentation and recognition, is entirely automated via a LabVIEW interface, and it takes approximately 1 h per sample. Census data count and classified radiolarian images are then automatically exported and saved. This new workflow paves the way for the analysis of long-term, radiolarian-based palaeoclimatic records from siliceous-remnant-bearing samples.</p>https://cp.copernicus.org/articles/16/2415/2020/cp-16-2415-2020.pdf |
| spellingShingle | M. Tetard R. Marchant R. Marchant G. Cortese Y. Gally T. de Garidel-Thoron L. Beaufort Technical note: A new automated radiolarian image acquisition, stacking, processing, segmentation and identification workflow Climate of the Past |
| title | Technical note: A new automated radiolarian image acquisition, stacking, processing, segmentation and identification workflow |
| title_full | Technical note: A new automated radiolarian image acquisition, stacking, processing, segmentation and identification workflow |
| title_fullStr | Technical note: A new automated radiolarian image acquisition, stacking, processing, segmentation and identification workflow |
| title_full_unstemmed | Technical note: A new automated radiolarian image acquisition, stacking, processing, segmentation and identification workflow |
| title_short | Technical note: A new automated radiolarian image acquisition, stacking, processing, segmentation and identification workflow |
| title_sort | technical note a new automated radiolarian image acquisition stacking processing segmentation and identification workflow |
| url | https://cp.copernicus.org/articles/16/2415/2020/cp-16-2415-2020.pdf |
| work_keys_str_mv | AT mtetard technicalnoteanewautomatedradiolarianimageacquisitionstackingprocessingsegmentationandidentificationworkflow AT rmarchant technicalnoteanewautomatedradiolarianimageacquisitionstackingprocessingsegmentationandidentificationworkflow AT rmarchant technicalnoteanewautomatedradiolarianimageacquisitionstackingprocessingsegmentationandidentificationworkflow AT gcortese technicalnoteanewautomatedradiolarianimageacquisitionstackingprocessingsegmentationandidentificationworkflow AT ygally technicalnoteanewautomatedradiolarianimageacquisitionstackingprocessingsegmentationandidentificationworkflow AT tdegaridelthoron technicalnoteanewautomatedradiolarianimageacquisitionstackingprocessingsegmentationandidentificationworkflow AT lbeaufort technicalnoteanewautomatedradiolarianimageacquisitionstackingprocessingsegmentationandidentificationworkflow |