Black-Box Hyperparameter Optimization for Nuclei Segmentation in Prostate Tissue Images
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.
Unable to display preview. Download preview PDF.
- 1.Wang Y, Du S, Balakrishnan S, et al. Stochastic Zeroth-order Optimization in High Dimensions. arXiv:1710.10551; 2017.
- 3.Hansen N, Auger A, Ros R, et al.; ACM. Comparing results of 31 algorithms from the black-box optimization benchmarking BBOB-2009. Proc GECCO. 2010; p. 1689-1696.Google Scholar
- 4.Teodoro G, Kurҫ TM, Taveira LF, et al. Algorithm sensitivity analysis and parameter tuning for tissue image segmentation pipelines. Bioinformatics. 2016;33(7):1064-1072.Google Scholar
- 5.Snoek J, Larochelle H, Adams RP. Practical bayesian optimization of machine learning algorithms. Proc Adv Neural Inf Process Syst. 2012; p. 2951-2959.Google Scholar
- 6.Bergstra J, Yamins D, Cox DD; Citeseer. Hyperopt: a python library for optimizing the hyperparameters of machine learning algorithms. Proc SciPy. 2013; p. 13-20.Google Scholar
- 7.Komer B, Bergstra J, Eliasmith C. Hyperopt-sklearn: automatic hyperparameter configuration for scikit-learn. Proc ICML Workshop AutoML. 2014; p. 2825-2830.Google Scholar
- 8.Golovin D, Solnik B, Moitra S, et al.; ACM. Google vizier: a service for black-box optimization. Proc SIGKDD. 2017; p. 1487-1495.Google Scholar
- 9.Ronneberger O, Fischer P, Brox T; Springer. U-net: convolutional networks for biomedical image segmentation. Proc MICCAI. 2015; p. 234-241.Google Scholar
- 10.Hutter F, Hoos HH, Leyton-Brown K; Springer. Sequential model-based optimization for general algorithm configuration. Proc LION. 2011; p. 507-523.Google Scholar
- 11.Chen T, Guestrin C; ACM. Xgboost: a scalable tree boosting system. Proc SIGKDD. 2016; p. 785-794.Google Scholar
- 12.Goldberg DE. Genetic Algorithms in Search, Optimization, and Machine Learning. Addison-Wesley; 1989.Google Scholar