Efficient Segmentation of Cell Nuclei in Histopathological Images

Autor(es):Mostrar menos -

Linares, Oscar Cuadros ; Soriano-Vargas, Aurea Aurea ; Faical, Bruno S. ; Hamann, Bernd ; Fabro, Alexandre T. ; Traina, Agma J. M. ; DeHerrera, AGS ; Gonzalez, AR ; Santosh, KC ; Temesgen, Z ; Kane, B ; Soda, P

• Tipo de documento: Artigo Científico
• Fonte: 2020 IEEE 33RD INTERNATIONAL SYMPOSIUM ON COMPUTER-BASED MEDICAL SYSTEMS(CBMS 2020); v. N/A, p. 6-pg., 2020-01-01.

Resumo

Computer-aided cell nuclei segmentation in histology images is essential for image analysis. There is a demand for methods that accurately detect cell nuclei in large images. We introduce the FECS method for automatic cell nuclei segmentation in Hematoxylin and Eosin (H&E) stained histology images. Our method accurately segments cell nuclei, even in large images, efficiently. We use bimodal-like histograms to perform image binarization via the fast Otsu algorithm. We introduce a superpixel based filter for cell nuclei boundary detection. A Gaussian blur filter allows us to identify cell nuclei centers, which are understood as local minima in the individual cell nuclei regions. We have evaluated our method for two publicly available datasets. Out tests have produced average Jaccard index values of 0.963 and 0.914, respectively, supporting a high degree of segmentation accuracy. We have compared our method against a state-of-theart method; our method produced better results for both datasets. The average processing time of FECS was approximately just one second for images of 1k x 1k pixel resolution and about three minutes for larger images of 15k x 15k pixel resolution. (AU)

• Processo FAPESP: 16/17078-0 - Mineração, indexação e visualização de Big Data no contexto de sistemas de apoio à decisão clínica (MIVisBD)
• Beneficiário: Agma Juci Machado Traina
• Modalidade de apoio: Auxílio à Pesquisa - Temático
• Processo FAPESP: 18/06228-7 - Detecção de padrões e anomalias em dados médicos usando Modelagem Matemática
• Beneficiário: Bruno Squizato Faiçal
• Modalidade de apoio: Bolsas no Brasil - Pós-Doutorado
• Processo FAPESP: 18/06074-0 - Content-Based Image Retrieval using Selective Visual Attention
• Beneficiário: Oscar Alonso Cuadros Linares
• Modalidade de apoio: Bolsas no Brasil - Pós-Doutorado
• Processo FAPESP: 20/07200-9 - Analisando dados complexos vinculados a COVID-19 para apoio à tomada de decisão e prognóstico
• Beneficiário: Agma Juci Machado Traina
• Modalidade de apoio: Auxílio à Pesquisa - Regular