簡易檢索 / 詳目顯示

回結果列表
研究生： 羅文懋
Lo, Wen-Mao
論文名稱： 映射函數之壓縮於樣本學習基礎的超解析演算法
Data Compression of Mapping Function in Example Learning-based Super Resolution
指導教授： 賴源泰
Lai, yen-Tai
學位類別： 碩士
Master
系所名稱： 電機資訊學院 - 電機工程學系
Department of Electrical Engineering
論文出版年： 2016
畢業學年度： 104
語文別： 英文
論文頁數： 41
中文關鍵詞： 局部多梯度模式 、映射矩陣 、學習基礎 、超解析
外文關鍵詞： LMGP, mapping function, Example learning-based, super-resolution
相關次數： 點閱：89  下載：0 
分享至:
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報

    • 單張影像超解析是一個能將模糊的低解析度影像轉為清楚高解析的方法，是一個有效將低品質的訊號源轉為高清的方法，像是在網路鏡頭，手機或監視系統等。在眾多超解析演算法當中，以學習基礎的方法還原較為快速，且可以藉由資料庫訓練的方式取得圖片細節，使用線性回歸的方式找出轉換圖片所需的映射函數，但此函數所需的儲存空間較為龐大。
    　　因此本篇論文主要探討的主題是映射矩陣的大小壓縮，以改善超解析在行動裝置上的實用性，利用演算法中局部多梯度模式的特性，分類出每個補塊的頻率高低，而給予各種映射矩陣不同的量化程度，在維持原本圖片品質下達到壓縮的效果。

    The objective of single image super-resolution (SR) is to restore a visually pleasing high-resolution (HR) image from a single low-resolution (LR) input. SR reconstruction is an effective signal recovery technique that produces high quality images from low-cost imaging systems (e.g., webcams or mobile phones) and limited environmental conditions (e.g., security surveillance or remote sensing imaging). There are many method nowadays, and the example-learning-based approach is faster than others and restore the detail by using training dataset. It uses linear regression to find the mapping functions for transforming images, but the problem is that storage requirement is large.
    In this thesis, we focus on the compression of mapping function to improve applicability of mobile device. This work presents local multi-gradient level pattern (LMGP) to describe the patches, and mapping function can be classified to different image frequency. This thesis compresses storage space remaining original image quality by giving different quantization to values in mapping function.

    Chapter 1　Introduction 1 1.1 Motivation 1 1.2 Introduction 1 1.3 Related Researches 1 1.4 Interpolation-based Method 2 1.5 Example Learning-based Method 2 1.5.1 Neighbor Embedding-base Methods 3 1.5.2 The Sparse Coding Approach Methods 4 1.5.3 Regression-based Methods 4 1.6 Mapping Function 5 1.7 Thesis Organization 6 Chapter 2　Background 7 2.1 Local Binary Pattern 7 2.1.1 Local Multi-Gradient Level Pattern 8 2.2 Direct Super-Resolution by Simple Function 9 2.2.1 Dataset 10 2.2.2 Training phase 11 2.2.3 Implementation 12 2.3 Linear Regression 13 2.4 Double Precision Floating Point Format 16 2.5 Contrast Sensitive Function (CSF) 17 2.6 Quantization 19 Chapter 3　Proposed Method 20 3.1 Overview 20 3.2 Compression 21 3.2.1 Exponent Part 21 3.2.2 Mantissa Part 22 3.2.3 Compression Flow 25 3.3 Decompression 27 Chapter 4　Experimental Results and Discussions 30 4.1 Quality Assessment 30 4.1.1 PSNR 30 4.1.2 SSIM 31 4.1.3 IFC 32 4.2 Experiment Results 32 4.2.1 Quality Comparison 33 4.2.2 Time complexity and Compression Ratio 35 4.2.3 Pictures Comparison 37 Chapter 5　Conclusions 39 Reference 40

    [1] C.-Y. Yang and M.-H. Yang, “Fast direct super-resolution by simple functions,” in Proc. IEEE Int. Conf. Comput. Vis. Dec. 2013, pp. 561–568.
    [2] D. Glasner, S. Bagon, and M. Irani. Super-resolution from a single image. In ICCV, 2009.
    [3] H. Chang, D.-Y. Yeung, and Y. Xiong. Super-resolution through neighbor embedding. In CVPR, 2004.
    [4] W. T. Freeman, T. R. Jones, and E. C. Pasztor. Example-based super-resolution.
    IEEE Computer Graphics and Applications, pages 56–65,
    2002.
    [5] H. Chang, D.-Y. Yeung, and Y. Xiong, “Super-resolution through neighbor embedding,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. Jun/Jul. 2004, pp. 275–282.
    [6] K. I. Kim and Y. Kwon, “Single-image super-resolution using sparse regression and natural image prior,” IEEE Trans. Pattern Anal. Mach. Intell. vol. 32, no. 6, pp. 1127–1133, Jun. 2010.
    [7] S. Lloyd, “Least squares quantization in PCM,” IEEE Trans. Inf. Theory, vol. 28, no. 2, pp. 129–137, Mar. 1982.
    [8] K. Zhang, D. Tao, X. Gao, X. Li, and Z. Xiong, “Learning multiple linear mappings for efficient single image super-resolution,” IEEE Trans. Image Process. vol. 24, no. 3, pp. 846–861, Mar. 2015.
    [9] C. W. Chen, F. K. Hsu, D. W. Yang, J. Wang, and M. D. Shieh, “Fast model searching and combining for example learning-based super-resolution,” in Proc. IEEE Int. Symp. Circuits Syst. May. 2016.
    [10] T. Ojala, M. Pietikainen and T. Maenpaa, "Multiresolution Gray-Scale and Rotation Invariant Texture Classification with Local Binary Patterns," IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 24, no. 7, pp. 971-987, July 2002.
    [11] Hamid Rahim sheikh and Alan Conrad Bovik, “An Information Fidelity Criterion for Image Quality Assessment Using Natural Scene Statistics, “IEEE Trans. Image Processing.” .VOL. 14, NO. 12, pp.2117-2128. ,December 2005

    無法下載圖示 校內：立即公開
    校外：不公開
    電子論文尚未授權公開，紙本請查館藏目錄
    QR CODE