Segmentation and Labeling of Vertebra Using SegFormer Architecture

Authors

Archan Ghosh, University of Calcutta
Debgandhar Ghosh, Maulana Abul Kalam Azad University of Technology, West Bengal
Somoballi Ghoshal, University of Calcutta
Amlan Chakrabarti, University of Calcutta
Susmita Sur-Kolay, Indian Statistical Institute, Kolkata

Document Type

Conference Article

Publication Title

Communications in Computer and Information Science

Abstract

Vertebra segmentation and labeling in MR images of the spine play a vital role in the identification of diseases or anomalies. MRI captures the tissue structure of a spine accurately, hence it is essential to demarcate and identify the vertebra in the MRI image. There are both supervised and unsupervised methods for vertebra segmentation and labeling. However, the acquisition of requisite data is a challenge to designing methods with very high accuracy. In this work, we have modified a transformer-based architecture called Segformer for semantic segmentation of 3D sliced data. Our method leverages transfer learning on low-population data. With a new advanced masking logic, we achieve 99% accuracy for segmentation and labeling of lumbar spine MR images.

First Page

160

Last Page

171

DOI

10.1007/978-3-031-58174-8_15

Publication Date

1-1-2024

Recommended Citation

Ghosh, Archan; Ghosh, Debgandhar; Ghoshal, Somoballi; Chakrabarti, Amlan; and Sur-Kolay, Susmita, "Segmentation and Labeling of Vertebra Using SegFormer Architecture" (2024). Conference Articles. 898.
https://digitalcommons.isical.ac.in/conf-articles/898