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

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.

10 pages, 10 figures, submited to NIPS 2016 conference. Computer Vision and Pattern Recognition, Neurons and Cognition, Neural and Evolutionary Computing

Cited by in corpus (15)

• Audio style transfer
• Adversarial Reprogramming of Neural Networks
• Robust and Generalizable Visual Representation Learning via Random Convolutions
• Error-feedback stochastic modeling strategy for time series forecasting with convolutional neural networks
• Unifying local and non-local signal processing with graph CNNs
• Texture Interpolation for Probing Visual Perception
• Deep Stochastic Configuration Networks with Universal Approximation Property
• DiCOVA-Net: Diagnosing COVID-19 using Acoustics based on Deep Residual Network for the DiCOVA Challenge 2021
• Synthesizing Diverse, High-Quality Audio Textures
• MC-LSTM: Mass-Conserving LSTM
• Randomness in Deconvolutional Networks for Visual Representation
• How Powerful Are Randomly Initialized Pointcloud Set Functions?
• Generic Perceptual Loss for Modeling Structured Output Dependencies
• Image Synthesis and Style Transfer
• Image classification and retrieval with random depthwise signed convolutional neural networks