Recently browsing GitHub When , I came across a very awesome open source library , Using machine learning algorithm to enlarge the picture losslessly N times , It's against the sky ！
This library is called video2x, There are 4300+ A star , Is based on waifu2x,Anime4K,SRMD and RealSR Development tools , It not only supports lossless video amplification , It can also support pictures and GIF Animation .
Let's take a look at the official example effect .
Enlarge video
This example takes the video from 360P Zoom in to 4K, We can see that the local features are clearer , This is how many times it's magnified , Do your own calculations , Not only does it not reduce the image quality of the video picture , And enhanced ！！！
With this tool , Don't worry about those low-quality videos anymore ！
Zoom in GIF
It's primitive GIF（160x120） , The following is the original GIF Zoom in 4 The effect after times （640x480）：
Zoom in 4 Still very clear after times , And it looks more interesting ！ When your friend sends you a bad motion picture , Immediately use this tool to make the motion picture clear , Then send it back to your friend , See how he reacts ？
Enlarge picture
Except for video and GIF After the picture is enlarged , This tool can also enlarge the picture ：
The picture is enlarged 8 times , I feel a lot clearer , The color and texture of the whole picture are also much better , And there is no ambiguity at all , Do you feel that this young lady has become beautiful .
I seem to see many men go back and secretly process the photos of the goddess with this tool , Ha ha ha ！
Usage mode
We can GitHub Download the source code of this tool , Then run the example in the source code , By directly changing the parameters, you can process your own target image or video ：
Another simple way , Is to use the information provided by the author GUI Interface , This is currently only Windows edition , You can download Windows Installation package , Decompress and run .
The software interface looks like this ：
We can see , In fact, this interface is to interface some configuration parameters . But the default configuration has been initialized , We just need to select the target file , Then click Run .
But the process here is a little slow , You need to eat melon seeds , Wait slowly .
We can also use the command line interface that pops up when the software is running , Look at the detailed process of running ：
Running this software is very expensive GPU Of , After all, it's image processing ！
in addition ,Python Language is really the first language of artificial intelligence , With the help of AI This huge engine can do many amazing things .
