Verifiable and Energy Efficient Medical Image Analysis with Quantised Self-attentive Deep Neural Networks

TL;DR

This paper introduces quantised self-attentive neural networks as an energy-efficient and hardware-friendly alternative to CNNs for medical image analysis, achieving significant reductions in model size, parameters, and energy consumption.

Contribution

It proposes replacing convolutional layers with stand-alone self-attention layers and applying quantisation, demonstrating improved efficiency in medical image classification and segmentation tasks.

Findings

50-80% reduction in model size

60-80% fewer parameters

40-85% fewer FLOPs

Abstract

Convolutional Neural Networks have played a significant role in various medical imaging tasks like classification and segmentation. They provide state-of-the-art performance compared to classical image processing algorithms. However, the major downside of these methods is the high computational complexity, reliance on high-performance hardware like GPUs and the inherent black-box nature of the model. In this paper, we propose quantised stand-alone self-attention based models as an alternative to traditional CNNs. In the proposed class of networks, convolutional layers are replaced with stand-alone self-attention layers, and the network parameters are quantised after training. We experimentally validate the performance of our method on classification and segmentation tasks. We observe a $50-80\%$ reduction in model size, $60-80\%$ lesser number of parameters, $40-85\%$ fewer FLOPs and $65-80\%$ more energy efficiency during inference on CPUs. The code will be available at \href {https://github.com/Rakshith2597/Quantised-Self-Attentive-Deep-Neural-Network}{https://github.com/Rakshith2597/Quantised-Self-Attentive-Deep-Neural-Network}.