Learning The Optimal Threshold : A Novel Method to Segment Histopathology Images
Article Sidebar
Main Article Content
Abstract
In the last decade, a lot of segmentation techniques had been proposed. Most of them include using an encoder-decoder network, such as the UNet model, to predict the mask for a certain image. The issue with UNet models and fully convolutional networks, in general, is that they require a substantial amount of data. In this study, we propose a novel technique to segment nuclei from breast histology patches. Instead of learning to generate a mask from an input image, we simplify the problem by using a regression network that learns to predict the optimal threshold value. To find the best threshold value for a certain histology image, we introduce a similarity guided search algorithm that compares the embeddings of the ground truth mask with the embedding of all masks generated from different possible threshold values. The proposed network manages to segment the nuclei from a small histology dataset with high accuracy i.e. 90%.
Article Details
This work is licensed under a Creative Commons Attribution 4.0 International License.
References
Sun YS, Zhao Z, Yang ZN, Xu F, Lu HJ, Zhu ZY, Shi W, Jiang J, Yao PP, Zhu HP. Risk factors and preventions of breast cancer. International journal of biological sciences. 2017 Nov 1;13(11):1387. doi: 10.7150/ijbs.21635
Zhai J, Zhang S, Chen J, He Q. Autoencoder and its various variants. In2018 IEEE international conference on systems, man, and cybernetics (SMC) 2018 Oct 7 (pp. 415-419). IEEE. doi: 10.1109/SMC.2018.00080
Xia P, Zhang L, Li F. Learning similarity with cosine similarity ensemble. Information sciences. 2015 Jun 20;307:39-52. https://doi.org/10.1016/j.ins.2015.02.024
Targ S, Almeida D, Lyman K. Resnet in resnet: Generalizing residual architectures. arXiv preprint arXiv:1603.08029. 2016 Mar 25. https://doi.org/10.48550/arXiv.1603.08029
Chai T, Draxler RR. Root mean square error (RMSE) or mean absolute error (MAE). Geoscientific model development discussions. 2014 Feb 28;7(1):1525-34. doi:10.5194/gmdd-7-1525-2014
Rezaei S, Emami A, Zarrabi H, Rafiei S, Najarian K, Karimi N, Samavi S, Soroushmehr SR. Gland segmentation in histopathology images using deep networks and handcrafted features. In2019 41st annual international conference of the IEEE Engineering in Medicine and Biology Society (EMBC) 2019 Jul 23 (pp. 1031-1034). IEEE. doi: 10.1109/EMBC.2019.8856776
Li W, Manivannan S, Akbar S, Zhang J, Trucco E, McKenna SJ. Gland segmentation in colon histology images using hand-crafted features and convolutional neural networks. In2016 IEEE 13th international symposium on biomedical imaging (ISBI) 2016 Apr 13 (pp. 1405-1408). IEEE. doi: 10.1109/ISBI.2016.7493530
Tang J, Li J, Xu X. Segnet-based gland segmentation from colon cancer histology images. In2018 33rd Youth Academic Annual Conference of Chinese Association of Automation (YAC) 2018 May 18 (pp. 1078-1082). IEEE. doi: 10.1109/YAC.2018.8406531
[ 9] Khatun R, Chatterjee S. Machine learning approach for segmenting glands in colon histology images using local intensity and texture features. In2018 IEEE 8th International Advance Computing Conference (IACC) 2018 Dec 14 (pp. 314-320). IEEE. doi: 10.1109/IADCC.2018.8692135
Jack NP, Thomas W, Marick L, Fabien R. Segmentation of Nuclei in Histopathology Images by deep regression of the distance map. (No Title). 2018 Feb 16. https://cir.nii.ac.jp/crid/1882553967389743744
Nunn EJ, Khadivi P, Samavi S. Compound frechet inception distance for quality assessment of gan created images. arXiv preprint arXiv:2106.08575. 2021 Jun 16. https://doi.org/10.48550/arXiv.2106.08575
Brunet D, Vrscay ER, Wang Z. On the mathematical properties of the structural similarity index. IEEE Transactions on Image Processing. 2011 Oct 24;21(4):1488-99. doi: 10.1109/TIP.2011.2173206
Xu X, Xu S, Jin L, Song E. Characteristic analysis of Otsu threshold and its applications. Pattern recognition letters. 2011 May 1;32(7):956-61. https://doi.org/10.1016/j.patrec.2011.01.021