Lightweight Instance Segmentation Network for Microscopic Images of Suspension‐Cultured B Cells Integrating Edge Priors and Multi‐Scale Attention
Microscopy Research and Technique · 29.10.2026
DOI: 10.1002/jemt.70195 ↗Kurz gefasst
Fachartikel in der Zeitschrift Microscopy Research and Technique, online erschienen am 29.10.2026. Verfasst von Aoyang Tang und 8 weiteren Autor:innen. Beteiligte Einrichtung: School of Mechanical Engineering, Hubei University of Technology, Wuhan, People's Republic of China. Die Arbeit nennt 37 Literaturangaben. Gefördert von: International Science and Technology Joint Research Project of Hubei, China, Hubei Province Science and Technolog y Innovation Talent Program.
Abstract (Originalsprache)
Precise single‐cell analysis is important for life sciences and precision medicine. In antibody drug discovery and immunotherapy, automated segmentation and quantitative analysis of suspension‐cultured B cells in label‐free bright‐field microscopic images can support image‐based characterization of cell populations under the recorded culture conditions. However, accurate segmentation remains challenging because of weak boundary gradients, heterogeneous cell sizes, background interference, and high‐density clustering. Existing deep learning models may also face a trade‐off between segmentation performance and model compactness, which can limit their use in resource‐constrained automated microscopy systems. To address these challenges, we propose EPMSA‐Net, a lightweight instance segmentation network that integrates edge priors and multi‐scale attention for B‐cell image analysis. The proposed framework incorporates two complementary structures: the Edge‐Aware Feature Preservation (EAFP) module supplements weak boundary‐related and local spatial cues during early feature extraction, whereas the Multi‐Scale Contextual Attention (MSCA) module subsequently models multi‐scale contextual information and recalibrates downstream feature responses. Evaluation on the current suspension‐cultured B‐cell dataset yielded an F1‐score of 94.44% ± 0.53% and an Aggregated Jaccard Index (AJI) of 89.87% ± 0.78% over five independent training runs, with a parameter count of 3.276 M. Density‐stratified evaluation showed relatively small F1‐score variation across the predefined low‐, medium‐, and high‐density subsets, with a mean CV F1 of 0.84% ± 0.13%. The resulting instance masks were further used to calculate image‐level cell count and segmented area ratio (SAR). These results indicate that EPMSA‐Net achieves competitive segmentation performance with a compact model size on the current dataset. Further validation across independent culture batches, imaging systems, and hardware platforms is required to evaluate its broader applicability and deployment performance.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- School of Mechanical Engineering, Hubei University of Technology, Wuhan, People's Republic of China.
- Yunnan Kunming Blood Center, Kunming, People's Republic of China.
- Biomedical Engineering Department, College of Medicine, Muhimbili University of Health and Allied Sciences, Dar es Salaam, Tanzania.
- School of Mechanical Science and Engineering, Huazhong University of Science and Technology, Wuhan, People's Republic of China.
- Hubei Modern Manufacturing Quality Laboratory, Wuhan, People's Republic of China.
ORCID-Profile: Aoyang Tang, Zihan Li, Wenzhe Gong, Yujiao Song, Twaha Kabika, Zhiyong Liu, Guanglan Liao, Lei Nie
Förderung
- International Science and Technology Joint Research Project of Hubei, China
- Hubei Province Science and Technolog y Innovation Talent Program
Zitieren
APA
Tang, A., Li, Z., Gong, W., Song, Y., Du, X., Kabika, T., Liu, Z., Liao, G., & Nie, L. (2026). Lightweight Instance Segmentation Network for Microscopic Images of Suspension‐Cultured B Cells Integrating Edge Priors and Multi‐Scale Attention. Microscopy Research and Technique. https://doi.org/10.1002/jemt.70195
Vancouver
Tang A, Li Z, Gong W, Song Y, Du X, Kabika T, et al. Lightweight Instance Segmentation Network for Microscopic Images of Suspension‐Cultured B Cells Integrating Edge Priors and Multi‐Scale Attention. Microscopy Research and Technique. 2026. doi:10.1002/jemt.70195
BibTeX
@article{Tang2026,
author = {Tang, Aoyang and Li, Zihan and Gong, Wenzhe and Song, Yujiao and Du, Xia and Kabika, Twaha and Liu, Zhiyong and Liao, Guanglan and Nie, Lei},
title = {Lightweight Instance Segmentation Network for Microscopic Images of Suspension‐Cultured B Cells Integrating Edge Priors and Multi‐Scale Attention},
journal = {Microscopy Research and Technique},
year = {2026},
doi = {10.1002/jemt.70195}
}
EUVIMED