Texture Synthesis Using Convolutional Neural Networks
This provides a new tool for generating stimuli in neuroscience and offers insights into deep network representations, though it is incremental in applying existing CNNs to texture synthesis.
The paper tackled the problem of generating natural textures by using feature correlations from convolutional neural networks trained for object recognition, resulting in high perceptual quality samples that capture statistical properties of natural images.
Here we introduce a new model of natural textures based on the feature spaces of convolutional neural networks optimised for object recognition. Samples from the model are of high perceptual quality demonstrating the generative power of neural networks trained in a purely discriminative fashion. Within the model, textures are represented by the correlations between feature maps in several layers of the network. We show that across layers the texture representations increasingly capture the statistical properties of natural images while making object information more and more explicit. The model provides a new tool to generate stimuli for neuroscience and might offer insights into the deep representations learned by convolutional neural networks.