A Powerful Generative Model Using Random Weights for the Deep Image Representation

TL;DR

This paper demonstrates that untrained, randomly weighted convolutional neural networks can effectively perform deep visualization tasks, challenging the notion that training is essential for high-quality image representation and synthesis.

Contribution

The study introduces a novel approach using untrained random networks for deep visualization, achieving results comparable to trained models across multiple tasks.

Findings

Reconstructed images from untrained networks surpass those from trained networks.

Synthesized textures are nearly indistinguishable from natural textures.

Generated artistic images are highly competitive with prior trained network methods.

Abstract

To what extent is the success of deep visualization due to the training? Could we do deep visualization using untrained, random weight networks? To address this issue, we explore new and powerful generative models for three popular deep visualization tasks using untrained, random weight convolutional neural networks. First we invert representations in feature spaces and reconstruct images from white noise inputs. The reconstruction quality is statistically higher than that of the same method applied on well trained networks with the same architecture. Next we synthesize textures using scaled correlations of representations in multiple layers and our results are almost indistinguishable with the original natural texture and the synthesized textures based on the trained network. Third, by recasting the content of an image in the style of various artworks, we create artistic images with high perceptual quality, highly competitive to the prior work of Gatys et al. on pretrained networks. To our knowledge this is the first demonstration of image representations using untrained deep neural networks. Our work provides a new and fascinating tool to study the representation of deep network architecture and sheds light on new understandings on deep visualization.