A lightweight network based on dual-stream feature fusion and dual-domain attention for white blood cells segmentation

Luo, Yang1; Wang, Yingwei1; Zhao, Yongda1; Guan, Wei2; Shi, Hanfeng3; Fu, Chong4,5,6; Jiang, Hongyang3

2023-09-04

10.3389/fonc.2023.1223353

FRONTIERS IN ONCOLOGY   Impact Factor & Quartile

2234-943X

IntroductionAccurate white blood cells segmentation from cytopathological images is crucial for evaluating leukemia. However, segmentation is difficult in clinical practice. Given the very large numbers of cytopathological images to be processed, diagnosis becomes cumbersome and time consuming, and diagnostic accuracy is also closely related to experts' experience, fatigue and mood and so on. Besides, fully automatic white blood cells segmentation is challenging for several reasons. There exists cell deformation, blurred cell boundaries, and cell color differences, cells overlapping or adhesion.MethodsThe proposed method improves the feature representation capability of the network while reducing parameters and computational redundancy by utilizing the feature reuse of Ghost module to reconstruct a lightweight backbone network. Additionally, a dual-stream feature fusion network (DFFN) based on the feature pyramid network is designed to enhance detailed information acquisition. Furthermore, a dual-domain attention module (DDAM) is developed to extract global features from both frequency and spatial domains simultaneously, resulting in better cell segmentation performance.ResultsExperimental results on ALL-IDB and BCCD datasets demonstrate that our method outperforms existing instance segmentation networks such as Mask R-CNN, PointRend, MS R-CNN, SOLOv2, and YOLACT with an average precision (AP) of 87.41%, while significantly reducing parameters and computational cost.DiscussionOur method is significantly better than the current state-of-the-art single-stage methods in terms of both the number of parameters and FLOPs, and our method has the best performance among all compared methods. However, the performance of our method is still lower than the two-stage instance segmentation algorithms. in future work, how to design a more lightweight network model while ensuring a good accuracy will become an important problem.

white blood cells segmentation instance segmentation YOLACT-CIS dual-stream feature fusion network (DFFN) dual-domain attention module (DDAM)

SCI

English

Corresponding

Department of education in Liaoning Province China[LJKMZ20221811] ; Doctoral Scientific Research Foundation of Anshan Normal University[22b08] ; 14th Five-Year Plan Special Research Project of Anshan Normal University[sszx013]

Oncology

Oncology

WOS:001066944600001

FRONTIERS MEDIA SA

Web of Science

1.Anshan Normal Univ, Sch Math & Informat Sci, Anshan, Liaoning, Peoples R China
2.Anshan Normal Univ, Sch Appl Technol, Anshan, Liaoning, Peoples R China
3.Southern Univ Sci & Technol, Dept Comp Sci & Engn, Shenzhen, Guangdong, Peoples R China
4.Northeastern Univ, Sch Comp Sci & Engn, Shenyang, Peoples R China
5.Minist Educ, Engn Res Ctr Secur Technol Complex Network Syst, Shenyang, Peoples R China
6.Northeastern Univ, Key Lab Intelligent Comp Med Image, Minist Educ, Shenyang, Peoples R China

Luo, Yang,Wang, Yingwei,Zhao, Yongda,et al. A lightweight network based on dual-stream feature fusion and dual-domain attention for white blood cells segmentation[J]. FRONTIERS IN ONCOLOGY,2023,13.

Luo, Yang.,Wang, Yingwei.,Zhao, Yongda.,Guan, Wei.,Shi, Hanfeng.,...&Jiang, Hongyang.(2023).A lightweight network based on dual-stream feature fusion and dual-domain attention for white blood cells segmentation.FRONTIERS IN ONCOLOGY,13.

Luo, Yang,et al."A lightweight network based on dual-stream feature fusion and dual-domain attention for white blood cells segmentation".FRONTIERS IN ONCOLOGY 13(2023).