Swin-Integrated Cell Detection Model for Embryo Imaging and Morphological Evaluation

International Journal of Electronics and Communication Engineering

© 2025 by SSRG - IJECE Journal

Volume 12 Issue 12

Year of Publication : 2025

Authors : R. Barkavi, G. Yamuna, C. Jayaram

10.14445/23488549/IJECE-V12I12P107

How to Cite?

R. Barkavi, G. Yamuna, C. Jayaram, "Swin-Integrated Cell Detection Model for Embryo Imaging and Morphological Evaluation," SSRG International Journal of Electronics and Communication Engineering, vol. 12,  no. 12, pp. 75-87, 2025. Crossref, https://doi.org/10.14445/23488549/IJECE-V12I12P107

Abstract:

Accurate analysis of embryonic cell structures is important to evaluate the embryo quality during early developmental stages. The overlapping blastomeres, fragmentation, and irregular morphologies in microscopic images complicate the cell counting and centroid localization procedure. Conventional methods that utilize convolutions and progressive upsampling struggle to model long-range dependencies and adapt to the dense cell regions’ spatial irregularities. This research work is proposed with the objective of developing a precise and adaptable model to overcome the limitations in conventional methods. The proposed SwinDePeriNet is a combination of a hierarchical Swin Transformer encoder with a deformable perceptual module, which is specifically developed to capture the dynamic spatial relationships and structural variations. Additionally, a Gaussian-based probabilistic estimator is incorporated to generate localized confidence maps for accurate centroid detection. The proposed model is trained and tested using a benchmark cell image dataset and exhibited 95.3% accuracy, 97.6% R² score, 26 MAE, 250 MSE, 93.1% perfect localization rate, 5.2 pixels mean Euclidean error, and a false positive rate of 2.6% which is better than conventional models.

Keywords:

Microscopic Embryo Imaging, Cell Count Regression, Centroid Detection, Framework, Spatial Confidence Mapping, Transformer-Based Encoding, Deformable Context Fusion, Gaussian Localization.

References:

[1] Anna Cecchele et al., “Cellular and Molecular Nature of Fragmentation of Human Embryos,” International Journal of Molecular Sciences, vol. 23, no. 3, pp. 1-16, 2022.
[CrossRef] [Google Scholar] [Publisher Link]
[2] Reza Moradi Rad et al., “Cell-Net: Embryonic Cell Counting and Centroid Localization via Residual Incremental Atrous Pyramid and Progressive Upsampling Convolution,” IEEE Access, vol. 7, pp. 81945-81955, 2019.
[CrossRef] [Google Scholar] [Publisher Link]
[3] Chang Qiao et al., “Evaluation and Development of Deep Neural Networks for Image Super-Resolution in Optical Microscopy,” Nature Methods, vol. 18, pp. 194-202, 2021.
[CrossRef] [Google Scholar] [Publisher Link]
[4] Pavan Chandra Konda et al., “Fourier Ptychography: Current Applications and Future Promises,” Optics Express, vol. 28, no. 7, pp. 9603-9630, 2020.
[CrossRef] [Google Scholar] [Publisher Link]
[5] Nguyen Duy Tan et al., “From Images to Insights: Cell Counting and Uniformity Grading of Day 3 Embryos,” Computers in Biology and Medicine, vol. 196, pp. 1-16, 2025.
[CrossRef] [Google Scholar] [Publisher Link]
[6] Qingnan Ma et al., “DCO-Net: Deformable Convolution-based O-Shape Network for Fully Automated Placenta Segmentation,” Digital Signal Processing, vol. 145, 2024.
[CrossRef] [Google Scholar] [Publisher Link]
[7] Chengyang Zhang et al., “Difference-Deformable Convolution with Pseudo Scale Instance Map for Cell Localization,” IEEE Journal of Biomedical and Health Informatics, vol. 28, no. 1, pp. 355-366, 2024.
[CrossRef] [Google Scholar] [Publisher Link]
[8] Jiazheng Liu et al., “A Robust Transformer-Based Pipeline of 3D Cell Alignment, Denoise and Instance Segmentation on Electron Microscopy Sequence Images,” Journal of Plant Physiology, vol. 297, pp. 1-16, 2024.
[CrossRef] [Google Scholar] [Publisher Link]
[9] Xi Hu et al., “Transformer-based Deep Learning for Accurate Detection of Multiple Base Modifications using Single Molecule Real Time Sequencing,” Communications Biology, vol. 8, pp. 1-12, 2025.
[CrossRef] [Google Scholar] [Publisher Link]
[10] Wided S. Miled et al., “Semantic Segmentation of Human Blastocyst Images using Deep CNNs and Vision Transformers,” Network Modeling Analysis in Health Informatics and Bioinformatics, vol. 14, 2025.
[CrossRef] [Google Scholar] [Publisher Link]
[11] Hongju Jo et al., “A Novel Method for Effective Cell Segmentation and Tracking in Phase Contrast Microscopic Images,” Sensors, vol. 21, no. 10, pp. 1-13, 2021.
[CrossRef] [Google Scholar] [Publisher Link]
[12] Hasita V. Nalluri et al., “Optimizing Colocalized Cell Counting using Automated and Semiautomated Methods,” Neuroinformatics, vol. 23, pp. 1-12, 2025.
[CrossRef] [Google Scholar] [Publisher Link]
[13] Yajie Chen et al., “Cell Localization and Counting using Direction Field Map,” IEEE Journal of Biomedical and Health Informatics, vol. 26, no. 1, pp. 359-368, 2022.
[CrossRef] [Google Scholar] [Publisher Link]
[14] Syed Nauyan Rashid, and Muhammad Moazam Fraz, “Nuclei Probability and Centroid Map Network for Nuclei Instance Segmentation in Histology Images,” Neural Computing and Applications, vol. 35, pp. 15447-15460, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[15] Huan Liu et al., “UPBAS-Net: An Upsampling-Powered Boundary-Aware Segmentation Network for Fluorescent Spots in Microscopy Images,” Analytical Chemistry, vol. 97, no. 40, pp. 22200-22210, 2025.
[Google Scholar] [Publisher Link]
[16] Yue Guo et al., “SAU-Net: A Unified Network for Cell Counting in 2D and 3D Microscopy Images,” IEEE/ACM Transactions on Computational Biology and Bioinformatics, vol. 19, no. 4, pp. 1920-1932, 2022.
[CrossRef] [Google Scholar] [Publisher Link]
[17] Rui Liu et al., “Interactive Dual Network with Adaptive Density Map for Automatic Cell Counting,” IEEE Transactions on Automation Science and Engineering, vol. 21, no. 4, pp. 6731-6743, 2024.
[CrossRef] [Google Scholar] [Publisher Link]
[18] Qianqian Zhang et al., “Automatic Cell Counting from Stimulated Raman Imaging using Deep Learning,” Plos One, vol. 16, no. 7, pp. 1-18, 2021.
[CrossRef] [Google Scholar] [Publisher Link]
[19] Roberto Morelli et al., “Automating Cell Counting in Fluorescent Microscopy through Deep Learning with C-ResUnet,” Scientific Reports, vol. 11, pp. 1-11, 2021.
[CrossRef] [Google Scholar] [Publisher Link]
[20] López Flórez et al., “Automatic Cell Counting with YOLOv5: A Fluorescence Microscopy Approach,” International Journal of Interactive Multimedia and Artificial Intelligence, vol. 8, no. 3, pp. 1-15, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[21] Synthetic Cell Images and Masks, Kaggle. [Online]. Available: https://www.kaggle.com/datasets/vbookshelf/synthetic-cell-images-and masks-bbbc005-v1