Unsupervised Learning of Artistic Styles with Archetypal Style Analysis

Wynen, Daan; Schmid, Cordelia; Mairal, Julien

doi:10.48550/arxiv.1805.11155

Cited by 3 publications

(12 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These include three linear archetypal analysis methods: [11] (i.e. PCHA), [14], and [13] as well as two non-linear AA methods: kernel PCHA [11] and PCHA on the latent layer of a neural network [15]. For [15] we exchanged the classifier framework for an autoencoder and refer to the method as "PCHA on AE".…”

Section: Resultsmentioning

confidence: 99%

“…PCHA), [14], and [13] as well as two non-linear AA methods: kernel PCHA [11] and PCHA on the latent layer of a neural network [15]. For [15] we exchanged the classifier framework for an autoencoder and refer to the method as "PCHA on AE". We did this modification in order to be able to decode back to the data space, which is required for quantifying the performance of the methods, and because most of our data did not have labels.…”

Section: Resultsmentioning

confidence: 99%

“…There, an algorithm is provided for AA applied to the kernel space of a dataset. In [15], the authors perform archetypal analysis on the representation found in a hidden layer of an image classification neural network in order to define image styles. Although these methods extend AA to non-linear feature spaces, both apply a fixed transformation to the data space.…”

Section: Previous Work and Backgroundmentioning

confidence: 99%

See 2 more Smart Citations

Finding Archetypal Spaces Using Neural Networks

Dijk¹,

Burkhardt²,

Amodio³

et al. 2019

Preprint

View full text Add to dashboard Cite

Archetypal analysis is a data decomposition method that describes each observation in a dataset as a convex combination of "pure types" or archetypes. These archetypes represent extrema of a data space in which there is a trade-off between features, such as in biology where different combinations of traits provide optimal fitness for different environments. Existing methods for archetypal analysis work well when a linear relationship exists between the feature space and the archetypal space. However, such methods are not applicable to systems where the feature space is generated non-linearly from the combination of archetypes, such as in biological systems or image transformations. Here, we propose a reformulation of the problem such that the goal is to learn a non-linear transformation of the data into a latent archetypal space. To solve this problem, we introduce Archetypal Analysis network (AAnet), which is a deep neural network framework for learning and generating from a latent archetypal representation of data. We demonstrate stateof-the-art recovery of ground-truth archetypes in non-linear data domains, show AAnet can generate from data geometry rather than from data density, and use AAnet to identify biologically meaningful archetypes in single-cell gene expression data.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

Section: Previous Work and Backgroundmentioning

confidence: 99%

See 1 more Smart Citation

Finding Archetypal Spaces Using Neural Networks

Dijk¹,

Burkhardt²,

Amodio³

et al. 2019

Preprint

View full text Add to dashboard Cite

show abstract

“…This is better illustrated in Fig. 1, where a stateof-the-art feature transformation method, that is based on the Whitening Coloring Transformation (WCT) [13,24], was used for transferring the style. Note that the WCT-based method produces artifacts, e.g., the color of the rocks leaks into the sea, due to its inability to adequately model the separate substyles that exist in the style image.…”

Section: Stylized Content Imagementioning

confidence: 99%

“…which they were pre-trained [3,8,11,22,23]. The aforementioned limitations were recently resolved by universal style transfer methods that employ one single trainable model to transfer arbitrary artistic styles via feature manipulations using a shared high-level feature space [6, 13,20,24]. Even though these universal style transfer approaches are capable of performing style transfer of arbitrary styles in a style-agnostic manner via feature transforms in (almost) realtime, they usually lead to less impressive results than the other algorithms [7].…”

Section: Introductionmentioning

confidence: 99%

Style Decomposition for Improved Neural Style Transfer

Pegios¹,

Passalis²,

Tefas³

2018

Preprint

View full text Add to dashboard Cite

Universal Neural Style Transfer (NST) methods are capable of performing style transfer of arbitrary styles in a style-agnostic manner via feature transforms in (almost) real-time. Even though their unimodal parametric style modeling approach has been proven adequate to transfer a single style from relatively simple images, they are usually not capable of effectively handling more complex styles, producing significant artifacts, as well as reducing the quality of the synthesized textures in the stylized image. To overcome these limitations, in this paper we propose a novel universal NST approach that separately models each sub-style that exists in a given style image (or a collection of style images). This allows for better modeling the subtle style differences within the same style image and then using the most appropriate sub-style (or mixtures of different sub-styles) to stylize the content image. The ability of the proposed approach to a) perform a wide range of different stylizations using the sub-styles that exist in one style image, while giving the ability to the user to appropriate mix the different sub-styles, b) automatically match the most appropriate sub-style to different semantic regions of the content image, improving existing state-of-the-art universal NST approaches, and c) detecting and transferring the sub-styles from collections of images are demonstrated through extensive experiments.

show abstract

Anisotropic Stroke Control for Multiple Artists Style Transfer

Chen

Yan

Liu

et al. 2020

Proceedings of the 28th ACM International Conference on Multimedia

View full text Add to dashboard Cite

Though significant progress has been made in artistic style transfer, semantic information is usually difficult to be preserved in a fine-grained locally consistent manner by most existing methods, especially when multiple artists styles are required to transfer within one single model. To circumvent this issue, we propose a Stroke Control Multi-Artist Style Transfer framework. On the one hand, we design an Anisotropic Stroke Module (ASM) which realizes the dynamic adjustment of style-stroke between the non-trivial and the trivial regions. ASM endows the network with the ability of adaptive semantic-consistency among various styles. On the other hand, we present an novel Multi-Scale Projection Discriminator to realize the texture-level conditional generation. In contrast to the single-scale conditional discriminator, our discriminator is able to capture multi-scale texture clue to effectively distinguish a wide range of artistic styles. Extensive experimental results well demonstrate the feasibility and effectiveness of our approach. Our framework can transform a photograph into different artistic style oil painting via only ONE single model. Furthermore, the results are with distinctive artistic style and retain the anisotropic semantic information. CCS CONCEPTS • Computing methodologies → Computational photography; Computer vision representations.

show abstract

Unsupervised Learning of Artistic Styles with Archetypal Style Analysis

Cited by 3 publications

References 15 publications

Finding Archetypal Spaces Using Neural Networks

Finding Archetypal Spaces Using Neural Networks

Style Decomposition for Improved Neural Style Transfer

Anisotropic Stroke Control for Multiple Artists Style Transfer

Contact Info

Product

Resources

About