Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Recurso digital |
| Lingua: | |
| Pubblicazione: |
Zenodo
2026
|
| Accesso online: | https://doi.org/10.5281/zenodo.19126698 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866902113872773120 |
|---|---|
| author | Vašinková, Markéta Jochymek, Lukáš Gajdoš, Petr |
| author_facet | Vašinková, Markéta Jochymek, Lukáš Gajdoš, Petr |
| contents | <h1>Curated multi-source dataset for AI-based peripheral blood cell classification and computer-assisted diagnosis with an unified 14-class taxonomy</h1> <div>This dataset is a curated collection of single-cell images of peripheral blood cells, designed for machine learning and artificial intelligence applications in hematological image analysis and computer-aided diagnosis.</div> <div> </div> <div>The dataset integrates three publicly available sources to ensure variability in staining protocols, imaging conditions, and annotation granularity:</div> <div>- AML-Cytomorphology dataset (Matek et al.)</div> <div>- Raabin-WBC dataset (Kouzehkanan et al.)</div> <div>- Multi-focus WBC dataset (Park et al.)</div> <div> </div> <div>All images correspond to cropped single-cell regions extracted from peripheral blood smears and annotated by expert hematopathologists or trained specialists.</div> <div> </div> <div>The primary contribution of this dataset lies in the harmonization of heterogeneous annotation schemes into a unified 14-class taxonomy that preserves biologically meaningful relationships between cell types and their developmental stages.</div> <div> </div> <h2>Source Datasets</h2> <h3>1. AML-Cytomorphology Dataset</h3> <div>Matek et al. (2019) provide a dataset of single-cell images from peripheral blood smears of patients with acute myeloid leukemia and non-malignant controls.</div> <div>Reference:</div> <div>Matek, C., Schwarz, S., Spiekermann, K., & Marr, C. (2019). Human-level recognition of blast cells in acute myeloid leukemia with convolutional neural networks. *Nature Machine Intelligence*, 1(11), 538–544.</div> <div> </div> <div>From this dataset, the following classes were retained:</div> <div>- Basophil</div> <div>- Eosinophil</div> <div>- Erythroblast</div> <div>- Lymphocyte (typical)</div> <div>- Atypical lymphocyte</div> <div>- Monocyte</div> <div>- Myelocyte</div> <div>- Metamyelocyte</div> <div>- Band neutrophil</div> <div>- Neutrophil (segmented)</div> <div>- Promyelocyte</div> <div>- Myeloblast</div> <p> </p> <h3>2. Raabin-WBC Dataset</h3> <div>Kouzehkanan et al. (2022) provide a large-scale dataset of peripheral blood smear images designed for automated WBC classification.</div> <div>Reference:</div> <div>Kouzehkanan, Z. M., Saghiri, M. A., et al. (2022). Raabin-WBC: a large free access dataset of white blood cells from normal peripheral blood. *Scientific Reports*, 12, 12584.</div> <div> </div> <div>From this dataset, the original five-class subset was used:</div> <div>- Basophil</div> <div>- Eosinophil</div> <div>- Lymphocyte</div> <div>- Monocyte</div> <div>- Neutrophil</div> <div> </div> <h3>3. Multi-focus WBC Dataset</h3> <div>Park et al. (2024) provide a dataset of leukocytes captured across multiple focal planes (z-stack imaging).</div> <div> </div> <div>Reference:</div> <div>Park, J. et al. (2024). A large multi-focus dataset for white blood cell classification. *Scientific Data*.</div> <div>Each sample consists of 10 focal slices (200×200 px) acquired with a 50× objective and 400 nm step size.</div> <div> </div> <div>From this dataset, the following classes were retained:</div> <div>- Immature WBC</div> <div>- Blast</div> <div>- Basophil</div> <div>- Band neutrophil</div> <div>- Metamyelocyte</div> <div>- Myelocyte</div> <div>- Promyelocyte</div> <div> </div> <h2> Class Harmonization</h2> <div>Due to differences in annotation schemes across datasets, all samples were mapped to a unified 14-class taxonomy.</div> <div>The harmonization process followed biologically and morphologically grounded rules:</div> <h3>Lymphoid lineage</h3> <div> <ul> <li>Typical lymphocytes (AML dataset) and lymphocytes (Raabin-WBC) were merged into a unified <em>Lymphocyte</em> class.</li> </ul> </div> <div> <ul> <li>Abnormal or reactive lymphocytes were preserved as a separate <em>Atypical lymphocyte</em> class.</li> </ul> </div> <h3>Blast cells</h3> <ul> <li>Monoblasts were not treated as a separate category due to limited morphological distinguishability and were merged into the broader <em>Blast</em> class.</li> </ul> <div> <ul> <li>Myeloblasts were retained as a distinct class due to consistent annotation and identifiable morphology.</li> </ul> </div> <h3>Erythroid lineage</h3> <ul> <li>Nucleated red blood cells (NRBCs) from the multi-focus dataset were mapped to the <em>Erythroblast</em> class.</li> </ul> <h3>Granulocytic maturation</h3> <div>Granulocyte development stages were preserved:</div> <div>- Promyelocyte</div> <div>- Myelocyte</div> <div>- Metamyelocyte</div> <div>- Band neutrophil</div> <div>- Neutrophil</div> <div> </div> <h3>Immature cells</h3> <ul> <li>The <em>Immature WBC </em>category from the multi-focus dataset was retained as a separate class to capture early or ambiguous precursor states.</li> </ul> <h3>Excluded or merged categories</h3> <div> <ul> <li>Rare or ambiguous subclasses and artifacts (e.g., smudge cells) were excluded.</li> </ul> </div> <div> <ul> <li>Subclasses not consistently represented across datasets were merged into broader biologically meaningful categories.</li> </ul> </div> <h2>Final Class Taxonomy</h2> <div>The dataset contains the following 14 classes:</div> <div>- Basophil</div> <div>- Blast</div> <div>- Eosinophil</div> <div>- Erythroblast</div> <div>- Immature WBC</div> <div>- Lymphocyte</div> <div>- Atypical lymphocyte</div> <div>- Metamyelocyte</div> <div>- Monocyte</div> <div>- Myeloblast</div> <div>- Myelocyte</div> <div>- Neutrophil</div> <div>- Band neutrophil</div> <div>- Promyelocyte</div> <div> </div> <div>This taxonomy explicitly captures hierarchical relationships within hematopoiesis, particularly the granulocytic maturation continuum.</div> <div> </div> <h2>Intended Use</h2> <div>This dataset is suitable for:</div> <div>- AI-based white blood cell classification</div> <div>- hierarchical and hyperbolic representation learning</div> <div>- multi-class classification under class imbalance</div> <div>- biomedical image analysis</div> <div>- computer-assisted hematological diagnosis</div> <div> </div> <h2>Licensing and Usage</h2> <div>This dataset is a derived work based on publicly available datasets. Users must comply with the original licenses of:</div> <div> </div> <div> <ul> <li>AML-Cytomorphology dataset (Matek et al.)</li> </ul> </div> <div> <ul> <li>Raabin-WBC dataset (Kouzehkanan et al.)</li> </ul> </div> <div> <ul> <li>Multi-focus WBC dataset (Park et al.)</li> </ul> </div> <p> </p> <div>The creators of this curated dataset do not claim ownership of the original images.</div> <div> </div> <h2>Citation</h2> <div>If you use this dataset, please cite:</div> <div>1. The original datasets listed above</div> <div>2. The associated publication (to be added)</div> <div> </div> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_19126698 |
| institution | Zenodo |
| language | |
| publishDate | 2026 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | Curated multi-source dataset for AI-based peripheral blood cell classification and computer-aided diagnosis with a unified 14-class taxonomy Vašinková, Markéta Jochymek, Lukáš Gajdoš, Petr <h1>Curated multi-source dataset for AI-based peripheral blood cell classification and computer-assisted diagnosis with an unified 14-class taxonomy</h1> <div>This dataset is a curated collection of single-cell images of peripheral blood cells, designed for machine learning and artificial intelligence applications in hematological image analysis and computer-aided diagnosis.</div> <div> </div> <div>The dataset integrates three publicly available sources to ensure variability in staining protocols, imaging conditions, and annotation granularity:</div> <div>- AML-Cytomorphology dataset (Matek et al.)</div> <div>- Raabin-WBC dataset (Kouzehkanan et al.)</div> <div>- Multi-focus WBC dataset (Park et al.)</div> <div> </div> <div>All images correspond to cropped single-cell regions extracted from peripheral blood smears and annotated by expert hematopathologists or trained specialists.</div> <div> </div> <div>The primary contribution of this dataset lies in the harmonization of heterogeneous annotation schemes into a unified 14-class taxonomy that preserves biologically meaningful relationships between cell types and their developmental stages.</div> <div> </div> <h2>Source Datasets</h2> <h3>1. AML-Cytomorphology Dataset</h3> <div>Matek et al. (2019) provide a dataset of single-cell images from peripheral blood smears of patients with acute myeloid leukemia and non-malignant controls.</div> <div>Reference:</div> <div>Matek, C., Schwarz, S., Spiekermann, K., & Marr, C. (2019). Human-level recognition of blast cells in acute myeloid leukemia with convolutional neural networks. *Nature Machine Intelligence*, 1(11), 538–544.</div> <div> </div> <div>From this dataset, the following classes were retained:</div> <div>- Basophil</div> <div>- Eosinophil</div> <div>- Erythroblast</div> <div>- Lymphocyte (typical)</div> <div>- Atypical lymphocyte</div> <div>- Monocyte</div> <div>- Myelocyte</div> <div>- Metamyelocyte</div> <div>- Band neutrophil</div> <div>- Neutrophil (segmented)</div> <div>- Promyelocyte</div> <div>- Myeloblast</div> <p> </p> <h3>2. Raabin-WBC Dataset</h3> <div>Kouzehkanan et al. (2022) provide a large-scale dataset of peripheral blood smear images designed for automated WBC classification.</div> <div>Reference:</div> <div>Kouzehkanan, Z. M., Saghiri, M. A., et al. (2022). Raabin-WBC: a large free access dataset of white blood cells from normal peripheral blood. *Scientific Reports*, 12, 12584.</div> <div> </div> <div>From this dataset, the original five-class subset was used:</div> <div>- Basophil</div> <div>- Eosinophil</div> <div>- Lymphocyte</div> <div>- Monocyte</div> <div>- Neutrophil</div> <div> </div> <h3>3. Multi-focus WBC Dataset</h3> <div>Park et al. (2024) provide a dataset of leukocytes captured across multiple focal planes (z-stack imaging).</div> <div> </div> <div>Reference:</div> <div>Park, J. et al. (2024). A large multi-focus dataset for white blood cell classification. *Scientific Data*.</div> <div>Each sample consists of 10 focal slices (200×200 px) acquired with a 50× objective and 400 nm step size.</div> <div> </div> <div>From this dataset, the following classes were retained:</div> <div>- Immature WBC</div> <div>- Blast</div> <div>- Basophil</div> <div>- Band neutrophil</div> <div>- Metamyelocyte</div> <div>- Myelocyte</div> <div>- Promyelocyte</div> <div> </div> <h2> Class Harmonization</h2> <div>Due to differences in annotation schemes across datasets, all samples were mapped to a unified 14-class taxonomy.</div> <div>The harmonization process followed biologically and morphologically grounded rules:</div> <h3>Lymphoid lineage</h3> <div> <ul> <li>Typical lymphocytes (AML dataset) and lymphocytes (Raabin-WBC) were merged into a unified <em>Lymphocyte</em> class.</li> </ul> </div> <div> <ul> <li>Abnormal or reactive lymphocytes were preserved as a separate <em>Atypical lymphocyte</em> class.</li> </ul> </div> <h3>Blast cells</h3> <ul> <li>Monoblasts were not treated as a separate category due to limited morphological distinguishability and were merged into the broader <em>Blast</em> class.</li> </ul> <div> <ul> <li>Myeloblasts were retained as a distinct class due to consistent annotation and identifiable morphology.</li> </ul> </div> <h3>Erythroid lineage</h3> <ul> <li>Nucleated red blood cells (NRBCs) from the multi-focus dataset were mapped to the <em>Erythroblast</em> class.</li> </ul> <h3>Granulocytic maturation</h3> <div>Granulocyte development stages were preserved:</div> <div>- Promyelocyte</div> <div>- Myelocyte</div> <div>- Metamyelocyte</div> <div>- Band neutrophil</div> <div>- Neutrophil</div> <div> </div> <h3>Immature cells</h3> <ul> <li>The <em>Immature WBC </em>category from the multi-focus dataset was retained as a separate class to capture early or ambiguous precursor states.</li> </ul> <h3>Excluded or merged categories</h3> <div> <ul> <li>Rare or ambiguous subclasses and artifacts (e.g., smudge cells) were excluded.</li> </ul> </div> <div> <ul> <li>Subclasses not consistently represented across datasets were merged into broader biologically meaningful categories.</li> </ul> </div> <h2>Final Class Taxonomy</h2> <div>The dataset contains the following 14 classes:</div> <div>- Basophil</div> <div>- Blast</div> <div>- Eosinophil</div> <div>- Erythroblast</div> <div>- Immature WBC</div> <div>- Lymphocyte</div> <div>- Atypical lymphocyte</div> <div>- Metamyelocyte</div> <div>- Monocyte</div> <div>- Myeloblast</div> <div>- Myelocyte</div> <div>- Neutrophil</div> <div>- Band neutrophil</div> <div>- Promyelocyte</div> <div> </div> <div>This taxonomy explicitly captures hierarchical relationships within hematopoiesis, particularly the granulocytic maturation continuum.</div> <div> </div> <h2>Intended Use</h2> <div>This dataset is suitable for:</div> <div>- AI-based white blood cell classification</div> <div>- hierarchical and hyperbolic representation learning</div> <div>- multi-class classification under class imbalance</div> <div>- biomedical image analysis</div> <div>- computer-assisted hematological diagnosis</div> <div> </div> <h2>Licensing and Usage</h2> <div>This dataset is a derived work based on publicly available datasets. Users must comply with the original licenses of:</div> <div> </div> <div> <ul> <li>AML-Cytomorphology dataset (Matek et al.)</li> </ul> </div> <div> <ul> <li>Raabin-WBC dataset (Kouzehkanan et al.)</li> </ul> </div> <div> <ul> <li>Multi-focus WBC dataset (Park et al.)</li> </ul> </div> <p> </p> <div>The creators of this curated dataset do not claim ownership of the original images.</div> <div> </div> <h2>Citation</h2> <div>If you use this dataset, please cite:</div> <div>1. The original datasets listed above</div> <div>2. The associated publication (to be added)</div> <div> </div> |
| title | Curated multi-source dataset for AI-based peripheral blood cell classification and computer-aided diagnosis with a unified 14-class taxonomy |
| url | https://doi.org/10.5281/zenodo.19126698 |