zeynelsamak / fema

This is the official implementation of paper FeMA: Feature Matching Auto-encoder for Predicting Ischaemic Stroke Evolution and Treatment Outcome

FeMA: Feature Matching Auto-encoder for Predicting Ischaemic Stroke Evolution and Treatment Outcome [Paper]

by Zeynel Abidin Samak, Philip Clatworthy and Majid Mirmehdi

Abstract

Although, predicting ischaemic stroke evolution and treatment outcome provide important information one step towards individual treatment planning, classifying functional outcome and modelling the brain tissue evolution remains a challenge due to data complexity and visually subtle changes in the brain. We propose a novel deep learning approach, Feature Matching Auto-encoder (FeMA) that consists of two stages, predicting ischaemic stroke evolution at one week without voxel-wise annotation and predicting ischaemic stroke treatment outcome at 90 days from a baseline scan. In the first stage, we introduce feature similarity and consistency objective, and in the second stage, we show that adding stroke evolution information increase the performance of functional outcome prediction. Comparative experiments demonstrate that our proposed method is more effective to extract representative follow-up features and achieves the best results for functional outcome of stroke treatment.

Code

The code will be released upon the paper publication.

Cite

@article{samak2022fema,
  title={FeMA: Feature matching auto-encoder for predicting ischaemic stroke evolution and treatment outcome},
  author={Samak, Zeynel A and Clatworthy, Philip and Mirmehdi, Majid},
  journal={Computerized Medical Imaging and Graphics},
  volume={99},
  pages={102089},
  year={2022},
  publisher={Elsevier}
}