Creation of a Deep Convolutional Auto-Encoder in Caffe

TL;DR

This paper presents the development of a deep convolutional auto-encoder in Caffe, demonstrating its effectiveness in dimensionality reduction on MNIST without using pooling layers.

Contribution

It introduces a simple method to create a deep convolutional auto-encoder in Caffe, highlighting its comparable performance to classic auto-encoders.

Findings

Achieved comparable accuracy to classic auto-encoders on MNIST

Designed a deep convolutional auto-encoder without pooling layers

Validated the model's effectiveness through experimental results

Abstract

The development of a deep (stacked) convolutional auto-encoder in the Caffe deep learning framework is presented in this paper. We describe simple principles which we used to create this model in Caffe. The proposed model of convolutional auto-encoder does not have pooling/unpooling layers yet. The results of our experimental research show comparable accuracy of dimensionality reduction in comparison with a classic auto-encoder on the example of MNIST dataset.