Abstract

In this paper, this study proposes Expectation-Maximization (EM) algorithm to segment pure Ground-Glass Opacity (GGO) nodules and solid part in part-solid GGO nodules and obtain their volumes as the features of GGO nodules. The features of GGO nodules are helpful to determine whether the GGO nodule is malignant. EM algorithm is used to segment pure GGO nodules and solid part in part-solid GGO nodules with different parameters. Experiments show EM algorithm can quickly segment solid part in part-solid GGO nodules and pure GGO nodules to reduce the burden of doctors.

Highlights

  • Lung cancer has a higher mortality rate than others in World Health Organization (1)

  • This paper proposes Expectation-Maximization (EM) algorithm to segment pure Ground-Glass Opacity (GGO) nodules and solid part in part-solid GGO nodules and obtain their volumes as the features of GGO nodules

  • Experiments show EM algorithm can quickly segment solid part in part-solid GGO nodules and pure GGO nodules to reduce the burden of doctors

Read more

Summary

Introduction

Lung cancer has a higher mortality rate than others in World Health Organization (1). Survival rate of lung cancer is less than 16% according to the survey in 5 years. Survival rate of lung cancer becomes 90% in 5 years if tumors of lung have been found early. When tumors process to the stages of malignant, survival rate of lung cancer in I stage is 60%, survival rate of lung cancer reduce 5% of IV stage from 40% of II stage. GGO nodules were segmented from the images using EM algorithm. The pure GGO nodules can be segmented directly and obtained its volume. The part-solid GGO nodules need to use EM algorithm again to segment solid part and no solid part in part-solid GGO nodules, and their volumes can be obtained. This paper proposes EM algorithm in order to find the clusters of GGO nodules.

Methods
Results
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call