Generative style transfer for MR image segmentation: A case of glioma segmentation in sub-saharan africa

---

Rancy Chepchirchir

Strathmore University, Kenya

: J Comput Eng Inf Technol

Abstract

---

In resource-constrained Sub-Saharan Africa (SSA), the utilization of lower-quality MRI technology raises questions about the applicability of machine learning (ML) methods for clinical tasks. The challenges associated with inadequate image contrast and resolution in SSA-acquired brain MRI scans demand advanced pre-processing to enhance image quality before employing ML techniques for tasks like tumor segmentation and classification. Addressing this, the BRaTS-Africa 2023 challenge aims to provide a robust deep learning-based brain tumor segmentation method tailored for the SSA population. This study presents a threefold approach. Firstly, the impact of domain shift from SSA training data on model efficacy is highlighted, revealing a discernible influence despite the small dataset size. Secondly, a comparative analysis of 3D and 2D full-resolution models using the nnU-Net framework indicates the superiority of a 2D model trained for 300 epochs achieving a five-fold cross-validation score of 0.9278. Lastly, addressing the performance gap observed in SSA validation as opposed to GLI validation, two strategies are proposed: fine-tuning SSA cases using the GLI+SSA best-pretrained 2D fullres model at 300 epochs, and introducing a novel neural style transfer-based data augmentation technique for the SSA cases. The fusion of these strategies significantly improves SSA validation results within computational limitations. This investigation underscores the potential of enhancing brain tumor prediction within SSA’s unique healthcare landscape. Recent Publications: 1. Chepchirchir Rancy and Olukuru John (2017) Effectiveness of Commodity Futures in Curbing Spot Volatility. Journal of Finance and Economics. 2017, 5(3), 85-95. DOI: 10.12691/jfe-5-3-1 Pub. Date: April 06, 2017. 2. Chepchirchir Rancy (2018) Currency portfolio optimization with an innovative covariance matrix estimator (QMLE). Journal of Economics and International Finance.

Biography

---

Rancy Chepchirchir interest in AI stems from my recent experience in a hackathon at the 4th Big data school, where I worked on detecting out-of-distribution malaria samples. My team and I emerged as the winners, and this experience further ignited my passion for exploring the application of computer vision in the medical field. To build on this experience, I joined the SPARK Academy for further training in preparation for the MICCAI 2023 for the Brain Tumor Segmentation (BraTS challenge). This training has provided me with a solid foundation in the fundamental concepts and techniques required for working with medical imaging data.