4. CycleGANの画像変換と現代美術への応用

Jan 29, 20220 likes269 views

幸

幸太朗岩澤

CycleGANを用いてHenri Matisse の作品を学習させた画像変換モデルを作成しました。 https://www.youtube.com/watch?v=Bjbbcng6qOQ

CycleGANの画像変換と現代美術への応用
2022/1/29 機械学習の社会実装勉強会第7回
岩澤幸太朗

GAN (Generative Adversarial Network)
Generative Adversarial Nets
Ian J.Goodfellow et al., 2014
https://arxiv.org/pdf/1406.2661.pdf
G: Generator 　　　　z: ノイズベクトル
D: Discriminator 　　　　x: 本物のデータ（教師データ）
G(z): Generator が生成した偽のデータ
MNIST
dataset
z
G
D
G(z)
x
True:1
or
False: 0
偽物
(0.1, 0.4, -0.2 …)

D(x) = 1 の時、最大値0
D(x): xがラベル1(本物)であると判断する確
率
2.Generatorの学習
D(G(z)) = 0 の時、最大値0
1?D(G(z)): G(z)がラベル0(偽
物)と判断する確率
D(G(z)) = 1 の時、最小値 -∞
1?D(G(z)): G(z)がラベル0(偽物)と
判断する確率
MNIST dataset
z
G D
G(z)
x G: Generator 　　　　z: ノイズベクトル
D: Discriminator 　　　　x: 本物のデータ（教師データ）
G(z): Generator が生成した偽のデータ
True:1
or
False: 0
固定
固定
偽物
BCE ≡
1.Discriminatorの学習
GANの学習
(0.1, 0.4, -0.2 …)

Unpaired Image-to-Image
Translation using Cycle-Consistent
Adversarial Networks
Jun-Yan Zhu, Taesung Park et. al 2017
https://arxiv.org/pdf/1703.10593.pdf
Repository: https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix
lossの計算: cycle_gan_model.py#L168-L177
Cycle GAN Pix2Pix (線画から写真) Cycle GAN (写真から絵)

* Identity Mapping Lossは割愛?
* 図と数式は論文より引用 ?
G: 馬からシマウマ
F: シマウマから馬
Dy: シマウマ判別器
Dx: 馬判別器
Cycle Consistency Loss
Adversarial Loss

19～20世紀に活躍したフランスの画家、彫刻家 ?
フォービスム（野獣派）の代表 ?
Henri Matisse (アンリ?マティス)
緑の筋のあるマティス夫人の肖像 (1905)
王の悲しみ (1952)
at ポンピドゥセンター
赤いハーモニー (1908)
1869 ~ 1954
襟巻の女(1936)
※ 野獣派としての活動は1905~1908の３年間のみ

使用データ学習環境
https://www.wikiart.org/en/henri-matisse
から取得したマティス作品
1008点
Google colab pro のGPU環境
(1,072円/月)

制作: P.I.C.S
技術開発: aircode (Project page)
ムンク展ー共鳴する魂の叫び
東京都美術館 2018/10/27 ~ 2019/1/20
https://vimeo.com/307183476

identity loss
intuition behind the loss functions loss_idt_A and loss_idt_B #322
https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix/issues/322
論文より: https://arxiv.org/pdf/1703.10593.pdf

参考
- https://arxiv.org/pdf/1406.2661.pdf (GAN 原論文)
- https://arxiv.org/pdf/1703.10593.pdf (CycleGAN 原論文)
- https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix (CycleGAN著者のRepository)
- GANの損失関数
- 今さら聞けないGANの目的関数
- 敵対的生成ネットワーク（ GAN）(slideshare)
- Neural Style Transfer (画風変換)

This document summarizes generative adversarial networks (GANs) and their applications. It begins by introducing GANs and how they work by having a generator and discriminator play an adversarial game. It then discusses several variants of GANs including DCGAN, LSGAN, conditional GAN, and others. It provides examples of applications such as image-to-image translation, text-to-image synthesis, image generation, and more. It concludes by discussing major GAN variants and potential future applications like helping children learn to draw.

Generative Adversarial NetworksMustafa Yagmur

Generative adversarial networks (GANs) are a class of machine learning frameworks where two neural networks, a generator and discriminator, compete against each other. The generator learns to generate new data with the same statistics as the training set to fool the discriminator, while the discriminator learns to better distinguish real samples from generated samples. GANs have applications in image generation, image translation between domains, and image completion. Training GANs can be challenging due to issues like mode collapse.

Generative Adversarial Networks (GAN)Manohar Mukku

This document provides an overview of generative adversarial networks (GANs). It explains that GANs were introduced in 2014 and involve two neural networks, a generator and discriminator, that compete against each other. The generator produces synthetic data to fool the discriminator, while the discriminator learns to distinguish real from synthetic data. As they train, the generator improves at producing more realistic outputs that match the real data distribution. Examples of GAN applications discussed include image generation, text-to-image synthesis, and face aging.

颁测肠濒别骋础狈による异种モダリティ画像生成を用いた股関节惭搁滨の筋骨格セグメンテーション奈良先端大情報科学研究科

医用画像からのセグメンテーションは、定量的な诊断や患者个别のシミュレーションのため重要である。医用画像は、その撮像モダリティにより异なる特徴をもつ。颁罢は骨の构造を明瞭に撮影することができるのに対して、惭搁は软组织の构造を明瞭に撮影することができる。このような违いはセグメンテーションにおいても考虑する必要がある。そこで本研究では、颁测肠濒别骋础狈を用いて惭搁滨から同患者同姿势の颁罢画像を生成することにより、セグメンテーションに活用する手法を提案した。

StyleCLIP: Text-Driven Manipulation of StyleGAN Imagery ivaderivader

The paper presents three methods for text-driven manipulation of StyleGAN imagery using CLIP: 1. Direct optimization of the latent w vector to match a text prompt 2. Training a mapping function to map text to changes in the latent space 3. Finding global directions in the latent space corresponding to attributes by measuring distances between text embeddings The methods allow editing StyleGAN images based on natural language instructions and demonstrate CLIP's ability to provide fine-grained controls, but rely on pretrained StyleGAN and CLIP models and may struggle with unseen text or image domains.

GANs Presentation.pptxMAHMOUD729246

Generative Adversarial Networks (GANs) are a type of deep learning algorithm that use two neural networks - a generator and discriminator. The generator produces new data samples and the discriminator tries to determine whether samples are real or generated. The networks train simultaneously, with the generator trying to produce realistic samples and the discriminator accurately classifying samples. GANs can generate high-quality, realistic data and have applications such as image synthesis, but training can be unstable and outputs may be biased.

Variational Autoencoder Tutorial Hojin Yang

The document provides an introduction to variational autoencoders (VAE). It discusses how VAEs can be used to learn the underlying distribution of data by introducing a latent variable z that follows a prior distribution like a standard normal. The document outlines two approaches - explicitly modeling the data distribution p(x), or using the latent variable z. It suggests using z and assuming the conditional distribution p(x|z) is a Gaussian with mean determined by a neural network gθ(z). The goal is to maximize the likelihood of the dataset by optimizing the evidence lower bound objective.

Deep Learning TutorialAmr Rashed

Semi supervised classification with graph convolutional networks哲东郑

A graph convolutional network model is proposed for semi-supervised learning that takes into account both the graph structure and node features. The model uses a graph convolutional layer that approximates spectral graph convolutions using a localized first-order approximation. This allows the model to be applied to large-scale problems. The model is evaluated on several benchmark semi-supervised classification datasets where it achieves state-of-the-art performance.

Introduction to Generative Adversarial Networks (GANs)Appsilon Data Science

Deep Learning in Bio-Medical ImagingJoonhyung Lee

This document provides an introduction to deep learning in medical imaging. It explains that artificial neural networks are modeled after biological neurons and use multiple hidden layers to approximate complex functions. Convolutional neural networks are commonly used for image data, applying filters over images to extract features. Modern deep learning platforms perform cross-correlation instead of convolution for efficiency. The key process for improving deep learning models is backpropagation, which calculates the gradient of the loss function to update weights and biases in a direction that reduces loss. Deep learning has applications in medical imaging modalities like MRI, ultrasound, CT, and PET.

Electronic health records and machine learningEman Abdelrazik

Electronic health records and machine learning can be used together to generate real-world evidence. Real-world data is collected from electronic health records in real clinical settings and can provide insights into a treatment's effectiveness and safety outside of clinical trials. Machine learning models can analyze structured and unstructured data in electronic health records to identify patterns and make predictions. This can help with tasks like medical diagnosis, which is challenging due to variations between individuals and potential for misdiagnosis. However, developing accurate machine learning models requires addressing issues like selecting representative training data and setting performance standards.

02 Fundamental Concepts of ANNTamer Ahmed Farrag, PhD

This document provides an overview of neural networks and fuzzy systems. It outlines a course on the topic, which is divided into two parts: neural networks and fuzzy systems. For neural networks, it covers fundamental concepts of artificial neural networks including single and multi-layer feedforward networks, feedback networks, and unsupervised learning. It also discusses the biological neuron, typical neural network architectures, learning techniques such as backpropagation, and applications of neural networks. Popular activation functions like sigmoid, tanh, and ReLU are also explained.

Machine LearningAnastasia Jakubow

This document provides an overview of machine learning including: definitions of machine learning; types of machine learning such as supervised learning, unsupervised learning, and reinforcement learning; applications of machine learning such as predictive modeling, computer vision, and self-driving cars; and current trends and careers in machine learning. The document also briefly profiles the history and pioneers of machine learning and artificial intelligence.

Generative adversarial networks?? ?

Generative Adversarial Networks (GANs) are a class of machine learning frameworks where two neural networks contest with each other in a game. A generator network generates new data instances, while a discriminator network evaluates them for authenticity, classifying them as real or generated. This adversarial process allows the generator to improve over time and generate highly realistic samples that can pass for real data. The document provides an overview of GANs and their variants, including DCGAN, InfoGAN, EBGAN, and ACGAN models. It also discusses techniques for training more stable GANs and escaping issues like mode collapse.

Introduction to Deep learningMassimiliano Ruocco

Deep learning is a class of machine learning algorithms that uses multiple layers of nonlinear processing units for feature extraction and transformation. It can be used for supervised learning tasks like classification and regression or unsupervised learning tasks like clustering. Deep learning models include deep neural networks, deep belief networks, and convolutional neural networks. Deep learning has been applied successfully in domains like computer vision, speech recognition, and natural language processing by companies like Google, Facebook, Microsoft, and others.

PR-305: Exploring Simple Siamese Representation LearningSungchul Kim

SimSiam is a self-supervised learning method that uses a Siamese network with stop-gradient to learn representations from unlabeled data. The paper finds that stop-gradient plays an essential role in preventing the model from collapsing to a degenerate solution. Additionally, it is hypothesized that SimSiam implicitly optimizes an Expectation-Maximization-like algorithm that alternates between updating the network parameters and assigning representations to samples in a manner analogous to k-means clustering.

MobileNet - PR044Jinwon Lee

Neural NetworkAshish Kumar

Neural networks are computing systems inspired by biological neural networks. They are composed of interconnected nodes that process input data and transmit signals to each other. The document discusses various types of neural networks including feedforward, recurrent, convolutional, and modular neural networks. It also describes the basic architecture of neural networks including input, hidden, and output layers. Neural networks can be used for applications like pattern recognition, data classification, and more. They are well-suited for complex, nonlinear problems. The document provides an overview of neural networks and their functioning.

??? ??? ??sundol75

Iclr2016 vaeまとめDeep Learning JP

This document summarizes a presentation about variational autoencoders (VAEs) presented at the ICLR 2016 conference. The document discusses 5 VAE-related papers presented at ICLR 2016, including Importance Weighted Autoencoders, The Variational Fair Autoencoder, Generating Images from Captions with Attention, Variational Gaussian Process, and Variationally Auto-Encoded Deep Gaussian Processes. It also provides background on variational inference and VAEs, explaining how VAEs use neural networks to model probability distributions and maximize a lower bound on the log likelihood.

3D Gaussian Splatting for laboratory paper roundsMikihiro Suzuki

What Makes Training Multi-modal Classification Networks Hard? ppttaeseon ryu

[??? ???? ?? 100??] What Makes Training Multi-modal Classification Networks Hard? ?????! ?? '??? ?? ?? ??'? ??? ??? 100?? ???? ?????! ??? ???? ??? ????? ?? ? ? ??? ??? ????? ?????, ? ????? ??????. ??? ???? ??? ??/?????? ??? ?? ?? ??? ??? ??? ?????! ?? ??? ?? 100?? ???, What Makes Training Multi-modal Classification Networks Hard? ?? ??? ?????. multi modal??, ??? ???? ???, ??? ??? ?????, ?? ??, ?? ??? ?? ??? ??? ??? ?? ????? ???? ??? ???? ? ??? ???? ?? ?????. ??? ???? ??? ????, ? ??? ?? ??? ??? ???? ??? ?? ??? ????? ???? ?? ??, ??+???, ???+???? ?? ?? ???? ???? ? ??????. ??? ?? ????? ? ??? ???? feature? ???? ???? ??? overfitting? ?? ??? ???? ?? ??? ??? ??????. ?? ???? ??? G-blend?? ??? ???? ??? ??? ????? ?? ???? ??????? ??? ?? ?? ?????! ???? ? ??????. https://youtu.be/ZjDRVgA9F1I

???? ?? ?? (Data Literacy)sidney yang

Neural Network Design: Chapter 18 Grossberg NetworkJason Tsai

(2017/06)Practical points of deep learning for medical imagingKyuhwan Jung

This document provides an overview of deep learning and its applications in medical imaging. It discusses key topics such as the definition of artificial intelligence, a brief history of neural networks and machine learning, and how deep learning is driving breakthroughs in tasks like visual and speech recognition. The document also addresses challenges in medical data analysis using deep learning, such as how to handle limited data or annotations. It provides examples of techniques used to address these challenges, such as data augmentation, transfer learning, and weakly supervised learning.

Harnessing the Power of GenAI for BI and Reporting.pptxParas Gupta

????, ????? ?? ???? 2019.12.20KYOYOON JUNG

15. Transformerを用いた言語処理技術の発展.pdf幸太朗岩澤

14. BigQuery ML を用いた多変量時系列テ?ータの解析.pdf幸太朗岩澤

More Related Content

What's hot (20)

Semi supervised classification with graph convolutional networks哲东郑

Introduction to Generative Adversarial Networks (GANs)Appsilon Data Science

Deep Learning in Bio-Medical ImagingJoonhyung Lee

Electronic health records and machine learningEman Abdelrazik

02 Fundamental Concepts of ANNTamer Ahmed Farrag, PhD

Machine LearningAnastasia Jakubow

Generative adversarial networks?? ?

Introduction to Deep learningMassimiliano Ruocco

PR-305: Exploring Simple Siamese Representation LearningSungchul Kim

MobileNet - PR044Jinwon Lee

Neural NetworkAshish Kumar

??? ??? ??sundol75

Iclr2016 vaeまとめDeep Learning JP

3D Gaussian Splatting for laboratory paper roundsMikihiro Suzuki

What Makes Training Multi-modal Classification Networks Hard? ppttaeseon ryu

???? ?? ?? (Data Literacy)sidney yang

Neural Network Design: Chapter 18 Grossberg NetworkJason Tsai

(2017/06)Practical points of deep learning for medical imagingKyuhwan Jung

Harnessing the Power of GenAI for BI and Reporting.pptxParas Gupta

????, ????? ?? ???? 2019.12.20KYOYOON JUNG

Semi supervised classification with graph convolutional networks哲东郑

Introduction to Generative Adversarial Networks (GANs)Appsilon Data Science

Deep Learning in Bio-Medical ImagingJoonhyung Lee

Electronic health records and machine learningEman Abdelrazik

02 Fundamental Concepts of ANNTamer Ahmed Farrag, PhD

Machine LearningAnastasia Jakubow

Generative adversarial networks?? ?

Introduction to Deep learningMassimiliano Ruocco

PR-305: Exploring Simple Siamese Representation LearningSungchul Kim

MobileNet - PR044Jinwon Lee

Neural NetworkAshish Kumar

??? ??? ??sundol75

Iclr2016 vaeまとめDeep Learning JP

3D Gaussian Splatting for laboratory paper roundsMikihiro Suzuki

What Makes Training Multi-modal Classification Networks Hard? ppttaeseon ryu

???? ?? ?? (Data Literacy)sidney yang

Neural Network Design: Chapter 18 Grossberg NetworkJason Tsai

(2017/06)Practical points of deep learning for medical imagingKyuhwan Jung

Harnessing the Power of GenAI for BI and Reporting.pptxParas Gupta

????, ????? ?? ???? 2019.12.20KYOYOON JUNG

More from 幸太朗岩澤 (14)

15. Transformerを用いた言語処理技術の発展.pdf幸太朗岩澤

14. BigQuery ML を用いた多変量時系列テ?ータの解析.pdf幸太朗岩澤

BigQuery ML for unstructured data幸太朗岩澤

12. Diffusion Model の数学的基礎.pdf幸太朗岩澤

行列分解の数学的基础.辫诲蹿幸太朗岩澤

BigQuery MLの行列分解モデルを用いた推薦システムの基礎幸太朗岩澤

Vertex AI Pipelinesで BigQuery MLのワークフローを管理（ETL ~ デプロイまで)幸太朗岩澤

Vertex AI Pipelinesで BigQuery MLのワークフローを管理幸太朗岩澤

7. Vertex AI Model Registryで BigQuery MLのモデルを管理する幸太朗岩澤

6. Vertex AI Workbench による Notebook 環境.pdf幸太朗岩澤

5. Big Query Explainable AIの紹介幸太朗岩澤

3. Vertex AIを用いた時系列テ?ータの解析幸太朗岩澤

2. BigQuery ML を用いた時系列テ?ータの解析 (ARIMA model)幸太朗岩澤

This document discusses using ARIMA models with BigQuery ML to analyze time series data. It provides an overview of time series data and ARIMA models, including how ARIMA models incorporate AR and MA components as well as differencing. It also demonstrates how to create an ARIMA prediction model and visualize results using BigQuery ML and Google Data Studio. The document concludes that ARIMA models in BigQuery ML can automatically select the optimal order for time series forecasting and that multi-variable time series are not yet supported.

1. BigQueryを中心にした ML datapipelineの概要幸太朗岩澤

This document discusses using BigQuery as the central part of an ML data pipeline from ETL to model creation to visualization. It introduces BigQuery and BigQuery ML, showing how ETL jobs can load data from Cloud Storage into BigQuery for analysis and model training. Finally, it demonstrates this process by loading a sample CSV dataset into BigQuery and using BigQuery ML to create and evaluate a prediction model.

15. Transformerを用いた言語処理技術の発展.pdf幸太朗岩澤

14. BigQuery ML を用いた多変量時系列テ?ータの解析.pdf幸太朗岩澤

BigQuery ML for unstructured data幸太朗岩澤

12. Diffusion Model の数学的基礎.pdf幸太朗岩澤

行列分解の数学的基础.辫诲蹿幸太朗岩澤

BigQuery MLの行列分解モデルを用いた推薦システムの基礎幸太朗岩澤

Vertex AI Pipelinesで BigQuery MLのワークフローを管理（ETL ~ デプロイまで)幸太朗岩澤

Vertex AI Pipelinesで BigQuery MLのワークフローを管理幸太朗岩澤

7. Vertex AI Model Registryで BigQuery MLのモデルを管理する幸太朗岩澤

6. Vertex AI Workbench による Notebook 環境.pdf幸太朗岩澤

5. Big Query Explainable AIの紹介幸太朗岩澤

3. Vertex AIを用いた時系列テ?ータの解析幸太朗岩澤

2. BigQuery ML を用いた時系列テ?ータの解析 (ARIMA model)幸太朗岩澤

1. BigQueryを中心にした ML datapipelineの概要幸太朗岩澤

Recently uploaded (8)

IoT Devices Compliant with JC-STAR Using Linux as a Container OSTomohiro Saneyoshi

Security requirements for IoT devices are becoming more defined, as seen with the EU Cyber Resilience Act and Japan’s JC-STAR. It's common for IoT devices to run Linux as their operating system. However, adopting general-purpose Linux distributions like Ubuntu or Debian, or Yocto-based Linux, presents certain difficulties. This article outlines those difficulties. It also, it highlights the security benefits of using a Linux-based container OS and explains how to adopt it with JC-STAR, using the "Armadillo Base OS" as an example. Feb.25.2025@JAWS-UG IoT

ドメインモデリング基本编①词全体の流れ2025冲02冲27社内向け开催.辫辫迟虫ssuserfcafd1

ElasticsearchでSPLADEする [Search Engineering Tech Talk 2025 Winter]kota usuha

PostgreSQL最新動向～カラムナストアから生成AI連携まで～ (Open Source Conference 2025 Tokyo/Spring ...NTT DATA Technology & Innovation

滨肠丑颈颈搁颈办颈蝉耻办别冲理学疗法士间の知识共有に向けた临床推论テキストの构造化に関する研究.辫诲蹿Matsushita Laboratory

Matching_Program_for_Quantum_Challenge_Overview.pdfhirokiokuda2

2025フードテックWeek大阪展示会 - LoRaWANを使った複数ポイント温度管理 by AVNET玉井部長CRI Japan, Inc.

Apache Sparkに対するKubernetesのNUMAノードを意識したリソース割り当ての性能効果 (Open Source Conference ...NTT DATA Technology & Innovation

IoT Devices Compliant with JC-STAR Using Linux as a Container OSTomohiro Saneyoshi

ドメインモデリング基本编①词全体の流れ2025冲02冲27社内向け开催.辫辫迟虫ssuserfcafd1

ElasticsearchでSPLADEする [Search Engineering Tech Talk 2025 Winter]kota usuha

PostgreSQL最新動向～カラムナストアから生成AI連携まで～ (Open Source Conference 2025 Tokyo/Spring ...NTT DATA Technology & Innovation

滨肠丑颈颈搁颈办颈蝉耻办别冲理学疗法士间の知识共有に向けた临床推论テキストの构造化に関する研究.辫诲蹿Matsushita Laboratory

Matching_Program_for_Quantum_Challenge_Overview.pdfhirokiokuda2

2025フードテックWeek大阪展示会 - LoRaWANを使った複数ポイント温度管理 by AVNET玉井部長CRI Japan, Inc.

Apache Sparkに対するKubernetesのNUMAノードを意識したリソース割り当ての性能効果 (Open Source Conference ...NTT DATA Technology & Innovation

4. CycleGANの画像変換と現代美術への応用

1. CycleGANの画像変換と現代美術への応用 2022/1/29 機械学習の社会実装勉強会第7回岩澤幸太朗

2. GAN (Generative Adversarial Network) Generative Adversarial Nets Ian J.Goodfellow et al., 2014 https://arxiv.org/pdf/1406.2661.pdf G: Generator 　　　　z: ノイズベクトル D: Discriminator 　　　　x: 本物のデータ（教師データ） G(z): Generator が生成した偽のデータ MNIST dataset z G D G(z) x True:1 or False: 0 偽物 (0.1, 0.4, -0.2 …)

3. D(x) = 1 の時、最大値0 D(x): xがラベル1(本物)であると判断する確率 2.Generatorの学習 D(G(z)) = 0 の時、最大値0 1?D(G(z)): G(z)がラベル0(偽物)と判断する確率 D(G(z)) = 1 の時、最小値 -∞ 1?D(G(z)): G(z)がラベル0(偽物)と判断する確率 MNIST dataset z G D G(z) x G: Generator 　　　　z: ノイズベクトル D: Discriminator 　　　　x: 本物のデータ（教師データ） G(z): Generator が生成した偽のデータ True:1 or False: 0 固定固定偽物 BCE ≡ 1.Discriminatorの学習 GANの学習 (0.1, 0.4, -0.2 …)

4. Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks Jun-Yan Zhu, Taesung Park et. al 2017 https://arxiv.org/pdf/1703.10593.pdf Repository: https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix lossの計算: cycle_gan_model.py#L168-L177 Cycle GAN Pix2Pix (線画から写真) Cycle GAN (写真から絵)

5. * Identity Mapping Lossは割愛? * 図と数式は論文より引用 ? G: 馬からシマウマ F: シマウマから馬 Dy: シマウマ判別器 Dx: 馬判別器 Cycle Consistency Loss Adversarial Loss

6. 19～20世紀に活躍したフランスの画家、彫刻家 ? フォービスム（野獣派）の代表 ? Henri Matisse (アンリ?マティス) 緑の筋のあるマティス夫人の肖像 (1905) 王の悲しみ (1952) at ポンピドゥセンター赤いハーモニー (1908) 1869 ~ 1954 襟巻の女(1936) ※ 野獣派としての活動は1905~1908の３年間のみ

7. 使用データ学習環境 https://www.wikiart.org/en/henri-matisse から取得したマティス作品 1008点 Google colab pro のGPU環境 (1,072円/月)

8. 画像は全て Wikimedia Commonsより取得

9. Loss

10. 制作: P.I.C.S 技術開発: aircode (Project page) ムンク展ー共鳴する魂の叫び東京都美術館 2018/10/27 ~ 2019/1/20 https://vimeo.com/307183476

11. identity loss intuition behind the loss functions loss_idt_A and loss_idt_B #322 https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix/issues/322 論文より: https://arxiv.org/pdf/1703.10593.pdf

12. 参考 - https://arxiv.org/pdf/1406.2661.pdf (GAN 原論文) - https://arxiv.org/pdf/1703.10593.pdf (CycleGAN 原論文) - https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix (CycleGAN著者のRepository) - GANの損失関数 - 今さら聞けないGANの目的関数 - 敵対的生成ネットワーク（ GAN）(slideshare) - Neural Style Transfer (画風変換)

13. 画像は全て Wikimedia Commonsより取得

狠狠撸

4. CycleGANの画像変換と現代美術への応用

Recommended