Review/Book

DKFZ-2019-02955

http://join2-wiki.gsi.de/foswiki/pub/Main/Artwork/join2_logo100x88.png Black-Box Hyperparameter Optimization for Nuclei Segmentation in Prostate Tissue Images

Handels, H. (Editor) ; Deserno, T. M. (Editor) ; Maier, A. (Editor) ; Maier-Hein, K. H. (Editor) ; Palm, C. (Editor) ; Tolxdorff, T. (Editor) ; Wollmann, T. ; Bernhard, P. ; Gunkel, M. ; Braun, D.DKFZ* ; Meiners, J. ; Simon, R. ; Sauter, G. ; Erfle, H. ; Rippe, K.DKFZ* ; Rohr, K.

2019
Springer Fachmedien Wiesbaden Wiesbaden
ISBN: 978-3-658-25325-7, 978-3-658-25326-4 (electronic)

Abstract: Segmentation of cell nuclei is essential for analyzing highcontent histological screens. Often, parameters of automatic approaches need to be optimized, which is tedious and difficult to perform manually. We propose a novel hyperparameter optimization framework, which formulates optimization as a combination of candidate sampling and an optimization strategy. We present a clustering based and a deep neural network based pipeline for nuclei segmentation, for which the parameters are optimized using state of the art optimizers as well as a novel optimizer. The pipelines were applied to challenging prostate cancer tissue images. We performed a quantitative evaluation using 28,388 parameter settings. It turned out that the deep neural network outperforms the clustering based pipeline, while the results for different optimizers vary slightly.

Note: Missing Journal: = 1431-472X (import from CrossRef Book Series)

---

Contributing Institute(s):

1. Chromatin-Netzwerke (B066)

Research Program(s):

1. 899 - ohne Topic (POF3-899) (POF3-899)

---