Pixelate was done using Pyodide since i prefer writing code in Python, but an improved version is available here: <a target="_blank" rel="ugc" href="https://civarchive.com/models/2003107?modelVersionId=2267300">https://civarchive.com/models/2003107?modelVersionId=2267300</a>Thanks <a target="_blank" rel="ugc" href="https://civarchive.com/user/LizArt9">LizArt9 </a>for taking the time to rewrite it in pure JS!V1.5Added more options to select quantization method, kmeans and dithering and a slightly better interface using simpleCSS (plus, no more need to change the wheel url if it is hosted next to it)V1Since not everyone has access to either A1111/Forge or even Python and that compiling <a target="_blank" rel="ugc" href="https://pillow.readthedocs.io/en/stable/">Pillow </a>with <a target="_blank" rel="ugc" href="https://pngquant.org/">libimagequant </a>is a pain in the ass, here is a tool to apply the pixel method described in my previous <a target="_blank" rel="ugc" href="https://civarchive.com/articles/19480/pixelart-colors-for-real-hardware">article</a>... directly in your browser!Leveraging <a target="_blank" rel="ugc" href="https://pyodide.org/en/stable/index.html">Pyodide</a>, a python interpreter in WASM and a <a target="_blank" rel="ugc" href="https://github.com/nArn0/pyodide-custom-recipes">custom build Pillow wheel</a> with libimagequant, you can directly pixelate any picture simply.<h3 id="the-tool-is-available-directly-as-a-static-space-in-hf:-https:huggingface.cospacesn-arnopixelate">The tool is available directly as a static space in HF: <a target="_blank" rel="ugc" href="https://huggingface.co/spaces/n-Arno/pixelate">https://huggingface.co/spaces/n-Arno/pixelate</a></h3>If you want to modify and host it yourself, you can download the source as a zip file here but please note that you'll have to host it somewhere (both the html and whl files) and change "https://&lt;your-server-url&gt;" in the code to your correct url like "https://test.com/page"How to use:1) browse for a picture2) click run (it may take a few seconds at start in order to init Pyodide)3) click reload original before re-running if you want to change the options (or they will be applied to the current image, may be helpful to reduce step by step).Options:1) scale: resize down the picture by dividing the dimension by the scale2) colors: how many colors in the target result3) quantize color: leverage libimagequant or not4) apply RGB555: reduce colors to 16bit5) crop to SNES size: center crop the resize picture dimensions if they exceeed SNES screen size (256x224)6) rescale post-process: resize back up by the same scale the resulting picture

Files with SHA256: 9a4868fddfeeb369e2cfe5aa8313ab86addbb6c86d861747423156f9bfbac631