Improved cell segmentation using deep learning in label-free optical microscopy images

Aydin Ayanzadeh*, Özden Yalçin Özuysal, Devrim Pesen Okvur, Sevgi Önal, Behçet Uğur TÖreyİn, Devrim Ünay

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

3 Citations (Scopus)


The recently popular deep neural networks (DNNs) have a significant effect on the improvement of segmentation accuracy from various perspectives, including robustness and completeness in comparison to conventional methods. We determined that the naive U-Net has some lacks in specific perspectives and there is high potential for further enhancements on the model. Therefore, we employed some modifications in different folds of the U-Net to overcome this problem. Based on the probable opportunity for improvement, we develop a novel architecture by using an alternative feature extractor in the encoder of U-Net and replacing the plain blocks with residual blocks in the decoder. This alteration makes the model superconvergent yielding improved performance results on two challenging optical microscopy image series: a phase-contrast dataset of our own (MDA-MB-231) and a brightfield dataset from a well-known challenge (DSB2018). We utilized the U-Net with pretrained ResNet-18 as the encoder for the segmentation task. Hence, following the modifications, we redesign a novel skip-connection to reduce the semantic gap between the encoder and the decoder. The proposed skip-connection increases the accuracy of the model on both datasets. The proposed segmentation approach results in Jaccard Index values of 85.0% and 89.2% on the DSB2018 and MDA-MB-231 datasets, respectively. The results reveal that our method achieves competitive results compared to the state-of-the-art approaches and surpasses the performance of baseline approaches.

Original languageEnglish
Pages (from-to)2855-2868
Number of pages14
JournalTurkish Journal of Electrical Engineering and Computer Sciences
Publication statusPublished - 2021

Bibliographical note

Publisher Copyright:


This work has been supported by the Scientific and Technological Research Council of Turkey (TÜBİTAK) under Grant 119E578. The data used in this study is collected under the Marie Curie IRG grant (no: FP7 PIRG08-GA-2010-27697). Aydın Ayanzadeh’s work is supported, in part, by Vodafone Turkey, under project no. ITUVF20180901P04 within the context of ITU Vodafone Future Lab R&D program.

FundersFunder number
ITU Vodafone Future Lab R&D program
Vodafone TurkeyITUVF20180901P04
Türkiye Bilimsel ve Teknolojik Araştirma Kurumu119E578, FP7 PIRG08-GA-2010-27697


    • Breast cancer
    • Brightfield
    • Convolutional neural networks
    • Optical microscopy
    • Phase-contrast
    • Segmentation


    Dive into the research topics of 'Improved cell segmentation using deep learning in label-free optical microscopy images'. Together they form a unique fingerprint.

    Cite this