CellsiLab
CoDAlab Research Line
CellsiLab
Computational Hematopathology & Biomedical AI
CellsiLab (Cells and image analysis) brings together researchers of UPC and clinicians of the Hospital Clínic de Barcelona / IDIBAPS. It develops computational models from microscope images to recognise blood cells and to support the diagnosis of haematological diseases.
Institutions
UPC - EEBE
Hospital Clínic de Barcelona / IDIBAPS
Co-Leads
Anna Merino
José Rodellar
Challenge
Knowledge & Quality of Life
System in deployment
CellsiMatic
The team combines clinical pathology, haematology laboratory practice, cytology, applied mathematics, statistics, biomedical engineering, image analysis and machine learning, aligned with the Knowledge and Quality of Life challenge.
It has two objectives: to develop computational models that recognise blood cells automatically from microscope images of peripheral blood, and to support clinical laboratory medicine through the automated diagnosis of haematological diseases.
CellsiMatic, the system that integrates these models, has been validated in a multicentre proof of concept in five hospitals and is deployed on an industrial platform. The group continues to develop it as a clinical decision-support tool.
Areas of expertise
Team
Current members are listed on the Group Members page. Former members: Dr. Andrea Acevedo, Dr. Laura Boldú, Dr. Laura Puigví, Dr. Maria Delgado.
Meet the team →Blood morphology and image-based artificial intelligence
The line develops new quantitative features from images to describe and standardise the morphology of blood cells: normal cells, reactive and abnormal neoplastic lymphocytes, blast cells, and abnormal red cells.
It also builds computer methods and algorithms that support haematological and non-haematological diagnosis through digital image analysis, machine learning and deep learning, and applies them to the automatic recognition of the cell types listed below in peripheral blood.
System in deployment
CellsiMatic: 5-hospital proof of concept, industrial platform, leukaemia and lymphoma.
Methods. Deep Learning, GANs / Diffusion Models, Computer Vision, Explainable AI, Stain Normalisation, Digital Pathology.
What the models recognise
Work over time, 2022-2026
Since 2022 the group has developed deep learning models for blast and atypical lymphoid cells, neutrophil abnormalities, synthetic cell images and stain normalisation. Outputs by year:
Generative AI in blood morphology. Review on synthetic image creation and digital stain normalization (International Journal of Laboratory Hematology).
MDS detection. MDS-AN, an attention-based explainable model that detects myelodysplastic syndrome from haematological variables (Biomedical Signal Processing and Control).
Carcinocythemia. Literature review on circulating tumour cells in peripheral blood smears (Advances in Laboratory Medicine).
SAM and transformer models. Red cell abnormalities, malaria parasites and NPM1-mutated acute myeloid leukaemia (congress abstracts).
CAR-T outcome. Quantification of reactive lymphocytes as a predictor of clinical outcome after CAR-T therapy (congress abstract).
Federated learning. Generalisation of blood cell classification models in multicentre settings (congress abstract).
Vision-language models. Fine-tuned models that generate morphological descriptions of blood cells (congress abstract).
Interpretable models. Interpretable identification of botryoid neutrophils (congress abstract).
MDS detection. Machine learning model based on haematological parameters (congress abstract).
Neutrophil abnormalities. Recognition of abnormalities in the cytoplasm of neutrophils (Computers in Biology and Medicine).
CAR-T therapy monitoring. Automatic detection of reactive lymphocytes after CAR-T therapy with sequential convolutional networks (congress abstract).
Activated monocytes. Quantification and biomarker development in an endotoxemia sepsis model (congress abstracts).
Morphological descriptions. Deep learning multimodels to describe and recognise blood cell images (congress abstract).
SyntheticCellGAN. Generation of synthetic images of leukocytes and leukaemic cells with GANs (Computer Methods and Programs in Biomedicine).
Stain normalisation. Digital colour staining normalisation with GANs for the recognition of abnormal blood cells (Computer Methods and Programs in Biomedicine).
Reactive lymphocytes in COVID-19. Deep learning recognition of reactive lymphocytes in patients with COVID-19 infection (Bioengineering).
Next steps
Consolidate CellsiMatic as a clinical decision-support tool, with iterative multicentre validation and data harmonisation using diffusion models. New models for the detection of myelodysplastic syndromes (MDS) and for CAR-T therapy monitoring.
Projects of the group between 2012 and 2027, drawn on a common time axis.
Funded research projects
xAI-HEALTH
2024 - 2027
Explainable deep learning in medical image analysis, with Hospital Clínic and Hospital Sant Joan de Déu.
Funded by: Agencia Estatal de Investigación (€128,750)
CellsiMaticPlus
2022 - 2024
Proof of concept and valorisation roadmap for a computer system that supports the diagnosis of leukaemia and lymphoma from blood images.
Funded by: Agencia Estatal de Investigación (€90,000)
Deep learning for haematological diagnosis
2020 - 2023
Deep learning methods to diagnose haematological diseases from blood cell images.
Funded by: Ministerio de Ciencia, Innovación y Universidades
Infectious diseases from blood cells
2019 - 2021
Image-based recognition of blood cells to detect infectious diseases with machine learning.
Funded by: Healthcare SLU
Leukaemic cells
2014 - 2019
Characterisation and automatic classification of leukaemic cells using digital image processing and pattern recognition.
Funded by: Ministerio de Economía y Competitividad
Doctoral and postdoctoral projects
Mireia García López
2026 - present
Automatic detection of erythrocyte morphological alterations in peripheral blood images using explainable deep learning
Jithin Velayudhan (postdoctoral)
2026 - present
Kevin Barrera
2019 - 2024
Generative models for the recognition of reactive and abnormal peripheral blood cells
Angel Molina
2018 - 2021
Detection and classification of abnormal red cells and malaria
Laura Boldú
2017 - 2020
Classification of myeloid, monocytic, promyelocytic and lymphoid blast cells
Andrea Acevedo
2016 - 2020
Deep learning system to classify normal and dysplastic blood cells
Laura Puigví
2016 - 2019
Quantitative features for the morphological characterisation of lymphoid and blast cells by image analysis
Santiago Alférez
2012 - 2015
Methodology to classify atypical lymphoid cells using image processing and machine learning
JournalMerino A., Barrera K., Alférez S. & Rodellar J.: "Generative artificial intelligence in peripheral blood morphology: synthetic image creation and digital stain normalization." International Journal of Laboratory Hematology. Read →
JournalLaguna J. & Merino A.: "Carcinocythemia detection in peripheral blood smears: a literature review." Advances in Laboratory Medicine, 7(2), 90-104. Read →
JournalBarrera K., Merino A., Díaz-Beya M., Molina Á. & Rodellar J.: "MDS-AN: An attention-based explainable model for detecting myelodysplastic syndrome using hematological variables." Biomedical Signal Processing and Control, 115, 109416. Read →
CongressBarrera K., García M., Molina A., Rodellar J. & Merino A.: "Discrimination of malaria parasites from other red blood cell inclusions using a SAM-ConvNeXt-based AI model." Int J Lab Hematol 48.
CongressBarrera K., Navarro C., Molina A., Rodellar J. & Merino A.: "Automated detection of red blood cell abnormalities in peripheral blood smears using SAM and ResNet." Int J Lab Hematol 48.
CongressBarrera K., Merino A. & Rodellar J.: "Automatic identification of acute myeloid leukemia with nucleophosmin gene mutation using a vision transformer-based artificial intelligence model." Int J Lab Hematol 48.
CongressCalvera A., Barrera K., Rodellar J., Martinez-Cibrian N., Albiol N., Oliver A. et al.: "Deep learning quantification of reactive lymphocytes in peripheral blood predicts clinical outcome after CAR-T therapy." Int J Lab Hematol 48.
CongressMerino A., Barrera K., Verdu G., San Jose P., Vallbe M., Molina A. & Rodellar J.: "Interpretable artificial intelligence model for the automatic identification of botryoid neutrophils." Int J Lab Hematol 48.
CongressRuiz L., Merino A., Barrera K., Rodellar J. & Alférez S.: "Automatic generation of morphological descriptions of peripheral blood cells with finetuned vision-language models." Int J Lab Hematol 48.
CongressBarrera K., Lorenzo N., Merino A. & Rodellar J.: "A federated learning approach to generalize blood cell image classification models in multicenter settings." Int J Lab Hematol 48.
CongressMerino A., Barrera K., Molina A., Alférez S. & Rodellar J.: "A new model based on machine learning and hematological parameters for myelodysplastic syndrome detection." Int J Lab Hematol 47.
JournalBarrera K., Rodellar J., Alférez S. & Merino A.: "A deep learning approach for automatic recognition of abnormalities in the cytoplasm of neutrophils." Computers in Biology and Medicine, 178, 108691. 17 citations (Google Scholar) Read →
CongressBarrera K., Merino A., Rubbini D., García S., Alférez S., Isamat M. et al.: "Artificial intelligence model for the quantification of activated monocytes in an endotoxemia sepsis model." Int J Lab Hematol 46, 12.
CongressMerino A., Rubbini D., Isamat M., Molina A., Castillo F., Lozano F. et al.: "Morphological changes in monocytes using an in vitro endotoxemia model: contribution to the detection of patients with early sepsis." Int J Lab Hematol 46, 67-68.
CongressCampanale M., Barrera K., Rodellar J. & Merino A.: "Automatic detection of reactive lymphocytes after CAR-T cell therapy using sequential convolutional neural networks." Int J Lab Hematol 46, 20-21.
CongressAlférez S., Merino A., Rubbini D., Isamat M., García L., Lozano F., Barrera K. et al.: "Development of a biomarker using computer vision and machine learning for the discrimination of lipopolysaccharide-induced activated monocytes." Int J Lab Hematol 46, 6-7.
CongressMotamedi L., Merino A., Barrera K., Alférez S. & Rodellar J.: "Deep learning multimodels for morphological description and recognition of peripheral blood cell images." Int J Lab Hematol 46, 72.
CongressMoncada Y., Merino A., Rodellar J., Caicedo A. & Alférez S.: "Peripheral blood cell imaging generation with latent diffusion models." Int J Lab Hematol 46, 70.
JournalBarrera K., Rodellar J., Alférez S. & Merino A.: "Automatic normalized digital color staining in the recognition of abnormal blood cells using generative adversarial networks." Computer Methods and Programs in Biomedicine, 240, 107629. IF 4.9, Q1. FWCI 1.57, Top 10% Popularity & Influence Read →
JournalBarrera K., Merino A., Molina A. & Rodellar J.: "Automatic generation of artificial images of leukocytes and leukemic cells using generative adversarial networks (SyntheticCellGAN)." Computer Methods and Programs in Biomedicine, 229, 107314. IF 4.6, Q1. 58 citations, FWCI 3.22 (94th percentile) Read →
CongressBarrera K., Merino A., Alférez S., Molina A. & Rodellar J.: "Vision transformer for automatic image recognition of peripheral blood cells." Int J Lab Hematol 45, 11.
CongressCalafat M.A., Alférez S., Molina A., Barrera K., Merino A. & Rodellar J.: "Custom automatic segmentation of normal and malignant lymphoid cell images for morphologic feature analysis." Int J Lab Hematol 45, 19-20.
CongressBarrera K., Merino A., Rodríguez-García M., Laguna J., Molina A. & Rodellar J.: "Automatic recognition of hypogranulation and cytoplasmic inclusions in neutrophils using sequential convolutional neural networks." Int J Lab Hematol 45, 16.
CongressBarrera K., Merino A., Laguna J., Rodríguez M., Molina Á. & Rodellar J.: "Hypogranulation and inclusions in neutrophils: deep learning models for the generation of artificial images and automatic recognition." Clin Chem Lab Med 61 (s1), s324-s358.
CongressMontcada Y., Merino A., Rodellar J., Caicedo A. & Alférez S.: "Peripheral blood cells imaging generation with diffusion models." Int J Lab Hematol 45, 81-82.
CongressYu H., Forster F., Panambur A.B., Merino A., Maier A. & Marquardt G.: "Unstained white blood cell classification using deep learning." Int J Lab Hematol 45, 132-133.
CongressTomczak A., Boldú L., Brock J.P., Merino A., Engel T. & Marquardt G.: "Identification of artificially stained images of peripheral blood cells." Int J Lab Hematol 45, 181-182.
JournalRodellar J. et al.: "A deep learning approach for the morphological recognition of reactive lymphocytes in patients with COVID-19 infection." Bioengineering, 9(5), 229. Read →
JournalMerino A. & Rodellar J.: "Quantitative features to assist in the diagnostic assessment of chronic lymphocytic leukemia progression." The Journal of Pathology, 257(1), 1-4. Read →
JournalMolina A., Rodellar J., Boldú L., Acevedo A., Alférez S. & Merino A.: "Automatic identification of malaria and other red blood cell inclusions using convolutional neural networks." Computers in Biology and Medicine, 136, 104680. 72 citations (Google Scholar) Read →
JournalAcevedo A., Merino A., Boldú L., Molina A., Alférez S. & Rodellar J.: "A new convolutional neural network predictive model for the automatic recognition of hypogranulated neutrophils in myelodysplastic syndromes." Computers in Biology and Medicine, 134, 104479. 50 citations (Google Scholar) Read →
JournalBoldú L., Merino A., Acevedo A., Molina A. & Rodellar J.: "A deep learning model (ALNet) for the diagnosis of acute leukaemia lineage using peripheral blood cell images." Computer Methods and Programs in Biomedicine, 202, 105999. 185 citations (Google Scholar) Read →
CongressBarrera K., Merino A., Acevedo A., Boldú L., Molina A. & Rodellar J.: "Generation of artificial images of leukocytes by means of new generative adversarial networks." International Society for Laboratory Hematology (ISLH).
CongressMerino A., Laguna J., Boldú L., Molina A., Barrera K. et al.: "Morphology and automatic recognition of atypical lymphocytes in peripheral blood in COVID-19 patients." Int J Lab Hematol.
CongressAcevedo A., Merino A., Boldú L., Molina A., Barrera K. & Rodellar J.: "Modular system for neutrophil recognition and myelodysplastic syndrome diagnosis." Int J Lab Hematol.
JournalMolina A., Alférez S., Boldú L., Acevedo A., Rodellar J. & Merino A.: "Sequential classification system for recognition of malaria infection using peripheral blood cell images." Journal of Clinical Pathology, 73(10), 665-670. 41 citations (Google Scholar) Read →
JournalDelgado-Ortet M., Molina A., Alférez S., Rodellar J. & Merino A.: "A deep learning approach for segmentation of red blood cell images and malaria detection." Entropy, 22(6), 657. 118 citations (Google Scholar) Read →
JournalAcevedo A., Merino A., Alférez S., Molina Á., Boldú L. & Rodellar J.: "A dataset of microscopic peripheral blood cell images for development of automatic recognition systems." Data in Brief, 30, 105474. 566 citations (Google Scholar) Read →
CongressBarrera K., Pérez A., Acevedo A., Alférez S., Boldú L. et al.: "Automatic generation of artificial images of peripheral blood cells using generative adversarial networks (GANs)." Int J Lab Hematol.
CongressPérez A., Alférez S., Barrera K., Merino A. & Rodellar J.: "Generative adversarial networks for peripheral cell blood images standardization." XXXVIII Congreso Anual de la Sociedad Española de Ingeniería Biomédica.
JournalAcevedo A. et al.: "Recognition of peripheral blood cell images using convolutional neural networks." Computer Methods and Programs in Biomedicine. Read →
JournalBoldú L. et al.: "Automatic recognition of different types of acute leukaemia in peripheral blood by image analysis." Journal of Clinical Pathology. Read →
JournalPuigví L. et al.: "Quantitative cytological descriptors to differentiate CLL, Sézary, Granular and Villous lymphocytes through image analysis." American Journal of Clinical Pathology. Read →
JournalRodellar J. et al.: "Image processing and machine learning in the morphological analysis of blood cells." International Journal of Laboratory Hematology. Read →
Books, patent and theses
Books and chapters
2019 Merino A. Manual de Citología de Sangre Periférica y Líquidos Biológicos, 2nd ed. Editorial Panamericana.
2011 Citología de Líquidos Biológicos: Experiencia del Grupo Catalán de Citología Hematológica. Editorial Acción Médica.
2005 Microorganismos de interés en Citología Hematológica. Club Català de Citologia Hematològica, Editorial Acción Médica.
2005 Merino A. Manual de Citología de Sangre Periférica. Editorial Acción Médica.
Patent
Alférez S., Merino A., Rodellar J., Mujica L. & Ruiz M.: computer-implemented method for the recognition and classification of abnormal blood cells.
P201331348, Spain. Priority date 9 May 2013.
Doctoral theses
2024 Kevin Barrera: Generative models for the automatic recognition of reactive and abnormal peripheral blood cells with diagnosis purposes.
2019 Laura Puigví: Morphological characterization of normal lymphoid cells, reactive, abnormal and blast cells in peripheral blood by means of digital image processing.
2015 Santiago Alférez: Methodology for automatic classification of atypical lymphoid cells from peripheral blood cell images.
2012 Gabriela Gutiérrez: Quality external evaluation in peripheral blood cytology: diagnostic competence of the Spanish laboratories by means of a new indicator.
Consultancy
Analysis of blood cell images with digital imaging analysers
Guidance on digital image processing methods
Answers on peripheral blood and body fluid cytology
Statistical studies
Training for specialists
Teaching
Courses in peripheral blood and body fluid cytology, through aulaclinic.com
University expert programme in peripheral blood cytology, biological fluids and image analysis methods
Projects
Collaboration in projects on digital morphology and haematology analysers
Courses
Dates and registration for each edition are published by the organiser.
Links of interest