Use Partial Convolution to solve SRCNN’s pixel loss of picture
碩士 === 國立臺北大學 === 統計學系 === 108 === Super-Resolution imaging (SR) is a technique for improving image resolution. Images can be divided into movies and pictures. In this paper, we want to perform super-resolution imaging on single images . At present, the methods for improving the resolution are gener...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2019
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/47v8gr |
| id |
ndltd-TW-107NTPU0337045 |
|---|---|
| record_format |
oai_dc |
| spelling |
ndltd-TW-107NTPU03370452019-09-10T03:32:37Z http://ndltd.ncl.edu.tw/handle/47v8gr Use Partial Convolution to solve SRCNN’s pixel loss of picture 使用部分卷積運算實現 SRCNN 的完整像素 HSU, WEI-HAN 許巍瀚 碩士 國立臺北大學 統計學系 108 Super-Resolution imaging (SR) is a technique for improving image resolution. Images can be divided into movies and pictures. In this paper, we want to perform super-resolution imaging on single images . At present, the methods for improving the resolution are generally Nearest - neighborhood interpolation, Bilinear Interpolation, and Bicubic Interpolation. In this paper, we use the technique proposed by C. Dong et al. for performing super- resolution imaging using a convolutional neural network. However,their technique reveals a problem of pixel loss for the high resolution output. This paper used partial convolution to solve this problem so that the output high-resolution image has the same size as the input image. WANG, CHUN-CHAO 汪群超 2019 學位論文 ; thesis 67 zh-TW |
| collection |
NDLTD |
| language |
zh-TW |
| format |
Others
|
| sources |
NDLTD |
| description |
碩士 === 國立臺北大學 === 統計學系 === 108 === Super-Resolution imaging (SR) is a technique for improving image resolution. Images can be divided into movies and pictures. In this paper, we want to perform super-resolution imaging on single images .
At present, the methods for improving the resolution are generally Nearest - neighborhood interpolation, Bilinear Interpolation, and Bicubic Interpolation. In this paper, we use the technique proposed by C. Dong et al. for performing super- resolution imaging using a convolutional neural network. However,their technique reveals a problem of pixel loss for the high resolution output. This paper used partial convolution to solve this problem so that the output high-resolution image has the same size as the input image.
|
| author2 |
WANG, CHUN-CHAO |
| author_facet |
WANG, CHUN-CHAO HSU, WEI-HAN 許巍瀚 |
| author |
HSU, WEI-HAN 許巍瀚 |
| spellingShingle |
HSU, WEI-HAN 許巍瀚 Use Partial Convolution to solve SRCNN’s pixel loss of picture |
| author_sort |
HSU, WEI-HAN |
| title |
Use Partial Convolution to solve SRCNN’s pixel loss of picture |
| title_short |
Use Partial Convolution to solve SRCNN’s pixel loss of picture |
| title_full |
Use Partial Convolution to solve SRCNN’s pixel loss of picture |
| title_fullStr |
Use Partial Convolution to solve SRCNN’s pixel loss of picture |
| title_full_unstemmed |
Use Partial Convolution to solve SRCNN’s pixel loss of picture |
| title_sort |
use partial convolution to solve srcnn’s pixel loss of picture |
| publishDate |
2019 |
| url |
http://ndltd.ncl.edu.tw/handle/47v8gr |
| work_keys_str_mv |
AT hsuweihan usepartialconvolutiontosolvesrcnnspixellossofpicture AT xǔwēihàn usepartialconvolutiontosolvesrcnnspixellossofpicture AT hsuweihan shǐyòngbùfēnjuǎnjīyùnsuànshíxiànsrcnndewánzhěngxiàngsù AT xǔwēihàn shǐyòngbùfēnjuǎnjīyùnsuànshíxiànsrcnndewánzhěngxiàngsù |
| _version_ |
1719249401009930240 |