Keras R-CNN: library for cell detection in biological images using deep neural networks

Autor(es):Mostrar menos -

Hung, Jane ^{[1, 2]} ; Goodman, Allen ^[2] ; Ravel, Deepali ^[3] ; Lopes, Stefanie C. P. ^{[4, 5]} ; Rangel, Gabriel W. ^[3] ; Nery, Odailton A. ^[6] ; Malleret, Benoit ^{[7, 8]} ; Nosten, Francois ^{[9, 10]} ; Lacerda, Marcus V. G. ^{[4, 5]} ; Ferreira, Marcelo U. ^[6] ; Renia, Laurent ^[8] ; Duraisingh, Manoj T. ^[3] ; Costa, Fabio T. M. ^[11] ; Marti, Matthias ^{[3, 12]} ; Carpenter, Anne E. ^[2]

Afiliação do(s) autor(es):Mostrar menos -

^[1] MIT, Dept Chem Engn, Cambridge, MA 02139 - USA

^[2] Broad Inst, Cambridge, MA 02142 - USA

^[3] Harvard TH Chan Sch Publ Hlth, Boston, MA - USA

^[4] Fundacao Oswaldo Cruz FIOCRUZ, Inst Leonidas & Maria Deane, Manaus, Amazonas - Brazil

^[5] Fundacao Med Trop Dr Heitor Vieira Dourado, Gerencia Malaria, Manaus, Amazonas - Brazil

^[6] Univ Sao Paulo, Sao Paulo - Brazil

^[7] Natl Univ Singapore, Yong Loo Lin Sch Med, Dept Microbiol & Immunol, Singapore 119077 - Singapore

^[8] Agcy Sci Res & Technol, Singapore Immunol Network SIgN, Singapore 138632 - Singapore

^[9] Mahidol Univ, Fac Trop Med, Shoklo Malaria Res Unit, Mahidol Oxford Trop Med Res Unit, Mae Sot - Thailand

^[10] Nuffield, Ctr Trop Med & Global Hlth, Oxford - England

^[11] Univ Estadual Campinas, Dept Genet Evolut Microbiol & Immunol, Campinas, SP - Brazil

^[12] Univ Glasgow, Coll Med Vet & Life Sci, Wellcome Ctr Integrat Parasitol Inst Infect Immun, Glasgow, Lanark - Scotland

• Tipo de documento: Artigo Científico
• Fonte: BMC Bioinformatics; v. 21, n. 1 JUL 11 2020.
• Citações Web of Science: 0

Resumo

Background: A common yet still manual task in basic biology research, high-throughput drug screening and digital pathology is identifying the number, location, and type of individual cells in images. Object detection methods can be useful for identifying individual cells as well as their phenotype in one step. State-of-the-art deep learning for object detection is poised to improve the accuracy and efficiency of biological image analysis. Results: We createdKeras R-CNNto bring leading computational research to the everyday practice of bioimage analysts.Keras R-CNNimplements deep learning object detection techniques using Keras and Tensorflow (https://github.com/ broadinstitute/keras-rcnn). We demonstrate the command line tool's simplified Application Programming Interface on two important biological problems, nucleus detection and malaria stage classification, and show its potential for identifying and classifying a large number of cells. For malaria stage classification, we compare results with expert human annotators and find comparable performance. Conclusions: Keras R-CNNis a Python package that performs automated cell identification for both brightfield and fluorescence images and can process large image sets. Both the package and image datasets are freely available on GitHub and the Broad Bioimage Benchmark Collection. (AU)

• Processo FAPESP: 17/18611-7 - Desenvolvimento de novas ferramentas para busca e validação de alvos moleculares para terapia contra Plasmodium vivax
• Beneficiário: Fabio Trindade Maranhão Costa
• Modalidade de apoio: Auxílio à Pesquisa - Temático