Selective Block Compensation for Scalable Super-resolution
碩士 === 清雲科技大學 === 電機工程研究所 === 96 === In this thesis, we propose a selective block compensation super-resolution mechanism to transmit a low quality bit-stream over the Internet, and end users can use the fast SR mechanism to reconstruct a high video quality sequence. The fast SR mechanism employs ei...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | zh-TW |
Published: |
2008
|
Online Access: | http://ndltd.ncl.edu.tw/handle/54206975927220009748 |
id |
ndltd-TW-096CYU00442005 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-096CYU004420052016-05-16T04:09:53Z http://ndltd.ncl.edu.tw/handle/54206975927220009748 Selective Block Compensation for Scalable Super-resolution 選擇式區塊補償技術達到可調性超解析度 Tsung-Hsuan Tsai 蔡宗軒 碩士 清雲科技大學 電機工程研究所 96 In this thesis, we propose a selective block compensation super-resolution mechanism to transmit a low quality bit-stream over the Internet, and end users can use the fast SR mechanism to reconstruct a high video quality sequence. The fast SR mechanism employs eigenvalue statistics (ES) to obtain the features in each Group of Pictures (GOP), and variance is applied to compute all of blocks’ differences between high-resolution (HR) frame and low-resolution (LR) frame. To reduce transmission overhead, the index selection technique is then applied to those blocks to compensate for blocks that have variances larger than a threshold. The results of index numbers from index selection are regarded as SR reconstruction information. These features and index parameters are transmitted to clients as side information. At the receiver end, a fast block compensation mechanism is applied for SR compensation to achieve high visual quality. Simulation results show that the proposed method has a fast, low complexity, and fine performance to retrieve the visual quality in video sequences. Ta-Te Lu Wan-Ching Chen 魯大德 陳萬清 2008 學位論文 ; thesis 75 zh-TW |
collection |
NDLTD |
language |
zh-TW |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 清雲科技大學 === 電機工程研究所 === 96 === In this thesis, we propose a selective block compensation super-resolution mechanism to transmit a low quality bit-stream over the Internet, and end users can use the fast SR mechanism to reconstruct a high video quality sequence. The fast SR mechanism employs eigenvalue statistics (ES) to obtain the features in each Group of Pictures (GOP), and variance is applied to compute all of blocks’ differences between high-resolution (HR) frame and low-resolution (LR) frame. To reduce transmission overhead, the index selection technique is then applied to those blocks to compensate for blocks that have variances larger than a threshold. The results of index numbers from index selection are regarded as SR reconstruction information. These features and index parameters are transmitted to clients as side information. At the receiver end, a fast block compensation mechanism is applied for SR compensation to achieve high visual quality. Simulation results show that the proposed method has a fast, low complexity, and fine performance to retrieve the visual quality in video sequences.
|
author2 |
Ta-Te Lu |
author_facet |
Ta-Te Lu Tsung-Hsuan Tsai 蔡宗軒 |
author |
Tsung-Hsuan Tsai 蔡宗軒 |
spellingShingle |
Tsung-Hsuan Tsai 蔡宗軒 Selective Block Compensation for Scalable Super-resolution |
author_sort |
Tsung-Hsuan Tsai |
title |
Selective Block Compensation for Scalable Super-resolution |
title_short |
Selective Block Compensation for Scalable Super-resolution |
title_full |
Selective Block Compensation for Scalable Super-resolution |
title_fullStr |
Selective Block Compensation for Scalable Super-resolution |
title_full_unstemmed |
Selective Block Compensation for Scalable Super-resolution |
title_sort |
selective block compensation for scalable super-resolution |
publishDate |
2008 |
url |
http://ndltd.ncl.edu.tw/handle/54206975927220009748 |
work_keys_str_mv |
AT tsunghsuantsai selectiveblockcompensationforscalablesuperresolution AT càizōngxuān selectiveblockcompensationforscalablesuperresolution AT tsunghsuantsai xuǎnzéshìqūkuàibǔchángjìshùdádàokědiàoxìngchāojiěxīdù AT càizōngxuān xuǎnzéshìqūkuàibǔchángjìshùdádàokědiàoxìngchāojiěxīdù |
_version_ |
1718268047000600576 |