Residual-Recursion Autoencoder for Shape Illustration Images

TL;DR

This paper introduces the Residual-Recursion Autoencoder (RRAE), a novel neural network framework that enhances autoencoder performance in reconstructing shape illustration images by recursively refining residuals.

Contribution

The paper proposes RRAE, a recursive autoencoder framework that improves low-dimensional feature extraction and reconstruction accuracy for shape illustration images, applicable as a wrapper over existing autoencoders.

Findings

Reconstruction loss decreased by 86.47% for convolutional autoencoder.

Reconstruction loss decreased by 10.77% for variational autoencoder.

Reconstruction loss decreased by 8.06% for conditional variational autoencoder.

Abstract

Shape illustration images (SIIs) are common and important in describing the cross-sections of industrial products. Same as MNIST, the handwritten digit images, SIIs are gray or binary and containing shapes that are surrounded by large areas of blanks. In this work, Residual-Recursion Autoencoder (RRAE) has been proposed to extract low-dimensional features from SIIs while maintaining reconstruction accuracy as high as possible. RRAE will try to reconstruct the original image several times and recursively fill the latest residual image to the reserved channel of the encoder's input before the next trial of reconstruction. As a kind of neural network training framework, RRAE can wrap over other autoencoders and increase their performance. From experiment results, the reconstruction loss is decreased by 86.47% for convolutional autoencoder with high-resolution SIIs, 10.77% for variational autoencoder and 8.06% for conditional variational autoencoder with MNIST.