• 沒有找到結果。

基於合作學習的神經網路進行圖片轉換 - 政大學術集成

N/A
N/A
Protected

Academic year: 2021

Share "基於合作學習的神經網路進行圖片轉換 - 政大學術集成"

Copied!
2
0
0

加載中.... (立即查看全文)

全文

(1)

2019 智慧企業資訊應用發展國際研討會 台灣,2019/06/29(六) 基於合作學習的神經網路進行圖片轉換 翁健豪 政治大學資訊管理學系研究所 106356034@nccu.edu.tw 黃存宇 政治大學資訊管理學系研究所 106356036@nccu.edu.tw 郁方 政治大學資訊管理學系研究所 yuf@nccu.edu.tw 摘要 本論文提出了一個新的圖像轉換方法,主要的功能是能實現將某一組被特定標籤後的資料 還原近似原始資料的樣子。例如我們將一張真實照片做成一張標籤化的圖片,透過大量資料集 的訓練,我們可以讓我們的神經網路學習到標籤化的圖片該如何還原成原來的真實照片,在本 文中的實驗部分我們會看到這樣的結果。 在之前的論文中已經有人提出對於這個問題的解法,最著名的是pix2pix,此作者使用了 conditional-GAN 的概念來解決此問題。在我們所使用的方法中,我們提出了一種新的方法, 不同於GAN 的生成式對抗網路,我們採用生成式合作網路的概念,在本文中我們將這方法命 名為Coop_pix2pix。生成式合作網路在之前的作品中,最著名的是 CoopNet,我們的作法也是 由此出發,將CoopNet 中合作網路的概念拿來解圖片轉換的題目,我們可以在後面的章節看到 使用這方法產生的圖片所帶來的效果。 關鍵詞:生成式合作網路、圖像轉換、深度學習、神經網路

(2)

2019 智慧企業資訊應用發展國際研討會 台灣,2019/06/29(六) Image-to-Image Translation with Cooperative Learning Networks

Chien-Hao Weng

Department of Management Information Systems, National Chengchi University 106356034@nccu.edu.tw

Cun Yu Huang

Department of Management Information Systems, National Chengchi University 106356036@nccu.edu.tw

Fang Yu

Department of Management Information Systems, National Chengchi University yuf@nccu.edu.tw

Abstract

This paper proposes a new Image-to-Image translation method. Our work is to restore the original picture from the picture with specific labels as close as possible. For example, we label a real photo to a labeled image. After training by a large number of data sets, we can let our neural network learn how to convert the labeled image to the original photo. We will show the result in the experiment part.

There are several previous works solving this problem. For example, pix2pix is a famous work, which uses the concept of conditional-GAN to solve this problem. In our works, we propose a new method. Different from GAN's generative adversarial network, we use concept of the cooperative learning network. We name this method Coop_pix2pix. CoopNet is a famous example in previous works of the cooperative learning network. Our work based on the cooperative neural networks concept, using this concept to solve Image-to-Image translation problem. We will demonstrate our work’s performance in our experiments part.

Keywords: Cooperative learning networks, Image-to-Image Translation, deep learning, neural network

參考文獻

相關文件

In this paper, we build a new class of neural networks based on the smoothing method for NCP introduced by Haddou and Maheux [18] using some family F of smoothing functions.

The temperature angular power spectrum of the primary CMB from Planck, showing a precise measurement of seven acoustic peaks, that are well fit by a simple six-parameter

To complete the “plumbing” of associating our vertex data with variables in our shader programs, you need to tell WebGL where in our buffer object to find the vertex data, and

• We will look at ways to exploit the text using different e-learning tools and multimodal features to ‘Level Up’ our learners’ literacy skills.. • Level 1 –

• In the present work, we confine our discussions to mass spectro metry-based proteomics, and to study design and data resources, tools and analysis in a research

CAST: Using neural networks to improve trading systems based on technical analysis by means of the RSI financial indicator. Performance of technical analysis in growth and small

CAST: Using neural networks to improve trading systems based on technical analysis by means of the RSI financial indicator. Performance of technical analysis in growth and small

▪ Can we decide whether a problem is “too hard to solve” before investing our time in solving it.. ▪ Idea: decide which complexity classes the problem belongs to