Nuclei Detection and Segmentation of Histopathological Images Using a Feature Pyramidal Network Variant of a Mask R-CNN.
Vignesh Ramakrishnan, Annalena Artinger, Laura Alexandra Daza Barragan, Jimmy Daza, Lina Winter, Tanja Niedermair, Timo Itzel, Pablo Arbelaez, Andreas Teufel, Cristina L Cotarelo, Christoph Brochhausen
Author Information
Vignesh Ramakrishnan: Institute of Pathology, University Regensburg, Franz-Josef-Strau��-Allee 11, 93053 Regensburg, Germany.
Annalena Artinger: Institute of Pathology, Medical Faculty Mannheim, Heidelberg University, Theodor-Kutzer-Ufer 1-3, 68167 Mannheim, Germany. ORCID
Laura Alexandra Daza Barragan: Center for Research and Formation in Artificial Intelligence (CinfonIA), Universidad de Los Andes, Cra. 1 E No. 19A-40, Bogot�� 111711, Colombia.
Jimmy Daza: Department of Internal Medicine II, Division of Hepatology, Medical Faculty Mannheim, Theodor-Kutzer-Ufer 1-3, 68167 Mannheim, Germany. ORCID
Lina Winter: Institute of Pathology, Medical Faculty Mannheim, Heidelberg University, Theodor-Kutzer-Ufer 1-3, 68167 Mannheim, Germany. ORCID
Tanja Niedermair: Institute of Pathology, University Regensburg, Franz-Josef-Strau��-Allee 11, 93053 Regensburg, Germany.
Timo Itzel: Department of Internal Medicine II, Division of Hepatology, Medical Faculty Mannheim, Theodor-Kutzer-Ufer 1-3, 68167 Mannheim, Germany.
Pablo Arbelaez: Center for Research and Formation in Artificial Intelligence (CinfonIA), Universidad de Los Andes, Cra. 1 E No. 19A-40, Bogot�� 111711, Colombia. ORCID
Andreas Teufel: Department of Internal Medicine II, Division of Hepatology, Medical Faculty Mannheim, Theodor-Kutzer-Ufer 1-3, 68167 Mannheim, Germany. ORCID
Cristina L Cotarelo: Institute of Pathology, Medical Faculty Mannheim, Heidelberg University, Theodor-Kutzer-Ufer 1-3, 68167 Mannheim, Germany.
Christoph Brochhausen: Institute of Pathology, Medical Faculty Mannheim, Heidelberg University, Theodor-Kutzer-Ufer 1-3, 68167 Mannheim, Germany. ORCID
Cell nuclei interpretation is crucial in pathological diagnostics, especially in tumor specimens. A critical step in computational pathology is to detect and analyze individual nuclear properties using segmentation algorithms. Conventionally, a semantic segmentation network is used, where individual nuclear properties are derived after post-processing a segmentation mask. In this study, we focus on showing that an object-detection-based instance segmentation network, the Mask R-CNN, after integrating it with a Feature Pyramidal Network (FPN), gives mature and reliable results for nuclei detection without the need for additional post-processing. The results were analyzed using the Kumar dataset, a public dataset with over 20,000 nuclei annotations from various organs. The dice score of the baseline Mask R-CNN improved from 76% to 83% after integration with an FPN. This was comparable with the 82.6% dice score achieved by modern semantic-segmentation-based networks. Thus, evidence is provided that an end-to-end trainable detection-based instance segmentation algorithm with minimal post-processing steps can reliably be used for the detection and analysis of individual nuclear properties. This represents a relevant task for research and diagnosis in digital pathology, which can improve the automated analysis of histopathological images.
