Deep Structure Learning using Feature Extraction in Trained Projection Space

TL;DR

This paper introduces PiNet, a neural network architecture that uses a Radon-transform-based feature extraction method to reduce memory and data requirements while maintaining competitive performance in volumetric segmentation tasks.

Contribution

The paper presents PiNet, a novel deep learning framework that employs a data-dependent Radon-transform for efficient feature extraction in lower-dimensional space.

Findings

Achieves comparable segmentation results with fewer parameters.

Demonstrates superior memory efficiency and processing speed.

Effective on public volumetric datasets.

Abstract

Over the last decade of machine learning, convolutional neural networks have been the most striking successes for feature extraction of rich sensory and high-dimensional data. While learning data representations via convolutions is already well studied and efficiently implemented in various deep learning libraries, one often faces limited memory capacity and insufficient number of training data, especially for high-dimensional and large-scale tasks. To overcome these limitations, we introduce a network architecture using a self-adjusting and data dependent version of the Radon-transform (linear data projection), also known as x-ray projection, to enable feature extraction via convolutions in lower-dimensional space. The resulting framework, named PiNet, can be trained end-to-end and shows promising performance on volumetric segmentation tasks. We test proposed model on public datasets to show that our approach achieves comparable results only using fractional amount of parameters. Investigation of memory usage and processing time confirms PiNet's superior efficiency compared to other segmentation models.