Toward more frugal models for functional cerebral networks automatic recognition with resting-state fMRI

TL;DR

This paper explores encoding techniques like supervoxels and graphs to create more efficient CNN models for recognizing resting-state functional brain networks in brain tumor patients, reducing complexity significantly while maintaining performance.

Contribution

It introduces a novel approach combining supervoxels and graph encoding to reduce CNN model complexity in brain network recognition tasks.

Findings

Graph encoding preserves activation characteristics.

Model complexity reduced by 26 times.

Performance comparable to classical CNN models.

Abstract

We refer to a machine learning situation where models based on classical convolutional neural networks have shown good performance. We are investigating different encoding techniques in the form of supervoxels, then graphs to reduce the complexity of the model while tracking the loss of performance. This approach is illustrated on a recognition task of resting-state functional networks for patients with brain tumors. Graphs encoding supervoxels preserve activation characteristics of functional brain networks from images, optimize model parameters by 26 times while maintaining CNN model performance.