Sometimes the perfect image can arrive in the wrong resolution. This is my daily case, when I find something that perfectly fits the theme of the post, but is only available at a poor 600px. If it happens to you too, you will be happy to know that there is a solution to this issue, and It is as simple as uploading the photo to this completely free artificial intelligence.
on the website of replicate.com You will find a multitude of artificial intelligence models. However, the one we are interested in is Latent-sr, which will allow you to resize images to a higher resolution than what is available. As an example, I used one of Woody, the popular character from toy story.
The artificial intelligence of replicate.com offered a magnificent result, increasing the image from a scant 320×480 pixels, to 1280×1920 pixels. In addition, he managed to maintain the composition and harmony of the details and textures, without noticeably deforming the different parts of the file. If you have used other AIs similar to this one, you will know how complicated it is, especially when dealing with animated characters.
To achieve such great results, the artificial intelligence runs on a series of NVIDIA T4 GPUs. Predictions take approximately 12 minutes to complete, although everything will depend on the complexity of the image you use. In my case, it took a couple of minutes to finish processing.
How to increase the resolution of images with this free artificial intelligence
Now that you’ve seen the example, you’ll probably want to do the same with your favorite low-resolution images. Well then, the process is very simple and does not even require you to register in the web.
The first thing you will have to do is enter artificial intelligence latent-sr. Once inside, he greets us with a sample image.
Use artificial intelligence it is as simple as dragging the image you want to the box marked by a segmented line located on the left side of the screen. You can also touch this area and the file browser will open, where you can browse and select the desired document.
Once uploaded, all that remains is to start configuring the AI. In the bottom box of Upscale Factor you can select the amount of resolution you want to increase. The higher the number, the better it will look, but it will also take longer to generate. On the other hand, we have the Sampling Stepswhich will allow artificial intelligence to generate a better prediction the higher the number, in exchange for performance and rendering time.
I have left everything as it comes by default, but if you want you can take advantage and play a little with the model. When everything is as you wish, just tap on the black button submission to start the process. Automatically, the right box will begin to display text, where you can also see the progress in a percentage.
Once the process is finished, you can see the full size image by tapping on the result or the button Download which appears just below. By right clicking you can save it to your internal storage, and it will be ready to be used.