More than 3 years have passed since last update.

【Super Resolution】ESPCN : deconvolutionの代わりに画像をsubpixel accuracyで推定する方法

Posted at 2021-02-03

Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network

ESPCN(efficient sub-pixel convolutional neural network)を紹介する

アルゴリズム

![image.png](https://qiita-image-store.s3.ap-northeast-1.amazonaws.com/0/482094/29827c6e-33e8-0963-540c-732869881c9f.png) convolutionでr^2のChannelを推定する。推定したr^2をr x rのsub-pixelに展開する。賢いな～

結果

![image.png](https://qiita-image-store.s3.ap-northeast-1.amazonaws.com/0/482094/f8a05256-c52e-7439-e413-38fa5673b727.png) 速度も速くて、精度も一番

結論

mono depthやstereo depthでsub-pixel毎にDisparity Mapを推定するのに使えそう。

参考文献

Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network https://arxiv.org/pdf/1609.05158.pdf

You get articles that match your needs
You can efficiently read back useful information
You can use dark theme

What you can do with signing up