发明名称 |
Method for scaling channel of an image |
摘要 |
The present invention relates to a method for scaling a first channel, wherein the method comprises computing a low resolution second channel based on a transformation of the second channel with a transformation function used to transform high resolution channel into a low resolution channel; computing a correlation function between the low resolution second channel and the first channel; determining a predicted second channel having the high resolution from the low resolution second channel according to a prediction method; computing an high-pass second channel based on the difference between the second channel and the predicted second channel and based on the correlation function; determining a predicted first channel having the high resolution from the first channel according to the prediction method. |
申请公布号 |
US9288464(B2) |
申请公布日期 |
2016.03.15 |
申请号 |
US201414462330 |
申请日期 |
2014.08.18 |
申请人 |
Qualcomm Technologies, Inc. |
发明人 |
Bernard Christophe;Varoquaux Vincent;De Vulpillieres Marc |
分类号 |
H04N9/64;H04N11/20;H04N7/01;H04N11/24;G06T3/40 |
主分类号 |
H04N9/64 |
代理机构 |
Knobbe, Martens, Olson & Bear, LLP |
代理人 |
Knobbe, Martens, Olson & Bear, LLP |
主权项 |
1. A method for enhancing an image, comprising:
determining a first channel of low resolution and a second channel of high resolution based on a frame of the image; applying a transformation function to the second channel to compute a low-resolution second channel; determining a correlation between the low-resolution second channel and the first channel; computing a high-pass second channel based at least in part on the second channel and the correlation; and enhancing the image via scaling the first channel based at least in part on the high-pass second channel. |
地址 |
San Diego CA US |