Non-Stationary Texture Synthesis by Adversarial Expansion

Slides:



Advertisements
Similar presentations
Hurieh Khalajzadeh Mohammad Mansouri Mohammad Teshnehlab
Advertisements

Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Evolving Reactive NPCs for the Real-Time Simulation Game.
NTNU Speech and Machine Intelligence Laboratory 1 Autoregressive product of multi-frame predictions can improve the accuracy of hybrid models 2016/05/31.
Scene Parsing With Integration of Parametric and Non-Parametric Models BING SHUAI, ZHEN ZUO, GANG WANG AND BING WANG 2016 IEEE TRANSACTIONS ON IMAGE PROCESSING.
Conditional Generative Adversarial Networks
Deeply-Recursive Convolutional Network for Image Super-Resolution
Generative Adversarial Nets
Deep Learning for Dual-Energy X-Ray
Generative Adversarial Network (GAN)
ALADDIN A Locality Aligned Deep Model for Instance Search
Deep Neural Net Scenery Generation
an introduction to: Deep Learning
Photorealistic Image Colourization with Generative Adversarial Nets
Perceptual Loss Deep Feature Interpolation for Image Content Changes
Regularizing Face Verification Nets To Discrete-Valued Pain Regression
Adversaries.
Part-Based Room Categorization for Household Service Robots
Chaoyun Zhang, Xi Ouyang, and Paul Patras
Synthesis of X-ray Projections via Deep Learning
Sujay Yadawadkar, Virginia Tech
Multiple Wavelet Coefficients Fusion in Deep Residual Networks for Fault Diagnosis
Ying Kin Yu, Kin Hong Wong and Siu Hang Or
Generative adversarial networks (GANs) for edge detection
Unsupervised Conditional Generation
Authors: Jun-Yan Zhu*, Taesun Park*, Phillip Isola, Alexei A. Efros
Adversarially Tuned Scene Generation
A Convolutional Neural Network Cascade For Face Detection
By: Kevin Yu Ph.D. in Computer Engineering
Data-driven methods: Texture 2 (Sz 10.5)
NormFace:
A New Boosting Algorithm Using Input-Dependent Regularizer
CS 4501: Introduction to Computer Vision Training Neural Networks II
ECE 599/692 – Deep Learning Lecture 1 - Introduction
Progressive Cross-media Correlation Learning
Presented by Shixing Chen
Object Detection + Deep Learning
8-3 RRAM Based Convolutional Neural Networks for High Accuracy Pattern Recognition and Online Learning Tasks Z. Dong, Z. Zhou, Z.F. Li, C. Liu, Y.N. Jiang,
KFC: Keypoints, Features and Correspondences
Image to Image Translation using GANs
GAN Applications.
Outline Background Motivation Proposed Model Experimental Results
Lip movement Synthesis from Text
RCNN, Fast-RCNN, Faster-RCNN
Iterative Crowd Counting
View Inter-Prediction GAN: Unsupervised Representation Learning for 3D Shapes by Learning Global Shape Memories to Support Local View Predictions 1,2 1.
Machine Learning based Data Analysis
A Dynamic System Analysis of Simultaneous Recurrent Neural Network
ImageNet Classification with Deep Convolutional Neural Networks
Advances in Deep Audio and Audio-Visual Processing
Heterogeneous convolutional neural networks for visual recognition
TPGAN overview.
Abnormally Detection
An introduction to: Deep Learning aka or related to Deep Neural Networks Deep Structural Learning Deep Belief Networks etc,
Ch 14. Generative adversarial networks (GANs) for edge detection
Chuan Wang1, Haibin Huang1, Xiaoguang Han2, Jue Wang1
Image Processing and Multi-domain Translation
Learning and Memorization
Object Detection Implementations
Multi-Modal Multi-Scale Deep Learning for Large-Scale Image Annotation
Angel A. Cantu, Nami Akazawa Department of Computer Science
CVPR 2019 Tutorial on Map Synchronization
Cengizhan Can Phoebe de Nooijer
Fig. 2. Examples showing the ability of deep learning to generate realistic fake images. (a) Representative test images from the trained network for generating.
Example of training and deployment of deep convolutional neural networks. Example of training and deployment of deep convolutional neural networks. During.
GIF2Video: Color Dequantization and Temporal Interpolation of GIF images Yang Wang, Haibin Huang, Chuan Wang, Tong He, Jue Wang, Minh Hoai. Stony Brook.
Deep Video Quality Assessor: From Spatio-temporal Visual Sensitivity to A convolutional Neural Aggregation Network Woojae Kim1, Jongyoo Kim2, Sewoong Ahn1,Jinwoo.
Recent Developments on Super-Resolution
SDSEN: Self-Refining Deep Symmetry Enhanced Network
ICLR, 2019 Jiahe Li
CVPR2019 Jiahe Li SiamRPN introduces the region proposal network after the Siamese network and performs joint classification and regression.
Presentation transcript:

Non-Stationary Texture Synthesis by Adversarial Expansion YANG ZHOU∗ , Shenzhen University and Huazhong University of Science & Technology ZHEN ZHU and XIANG BAI, Huazhong University of Science and Technology DANI LISCHINSKI, The Hebrew University of Jerusalem DANIEL COHEN-OR, Shenzhen University and Tel Aviv University HUI HUANG† , Shenzhen University HW6 606410004 黃立緯 606410010 劉玉蕙

INTRODUCTION Example-based texture synthesis has been an active area of research for over two decades. Despite excellent results for many classes of textures, example-based synthesis of significantly nonstationary textures remains an open problem.

INTRODUCTION (cont.) Recently, deep learning based approaches for texture synthesis have begun to gain popularity. However, none of the existing state-of-the-art methods are able to successfully cope with significantly non-stationary input exemplars.

Goal Given that our ultimate goal is to generate larger instances that perceptually resemble a smaller input texture exemplar, the main idea is to teach a fully convolutional generator network how to do just that.

Generator A generator architecture that satisfies our requirements was, in fact, already proposed by Johnson et al. [2016], who demonstrated its effectiveness for neural style transfer and for super-resolution. Thus, we adopt a similar architecture for our generator. Justin Johnson, Alexandre Alahi, and Li Fei-Fei. 2016. Perceptual Losses for Real-Time Style Transfer and Super-Resolution. In Proc. ECCV 2016, Bastian Leibe, Jiri Matas, Nicu Sebe, and Max Welling (Eds.), Vol. Part II. 694–711.

Discriminator We adopt the PatchGAN discriminator [Isola et al. 2017] . Phillip Isola, Jun-Yan Zhu, Tinghui Zhou, and Alexei A Efros. 2017. Image-to-Image Translation with Conditional Adversarial Networks. In Proc. CVPR 2017.

Conclusion We have presented an example-based texture synthesis method capable of expanding an exemplar texture, while faithfully preserving the global structures therein. The generator is able to faithfully reproduce local patterns, as well as their global arrangements. The trained model is stable enough for repeated application, enabling generating diverse results of different sizes.