Memento is a compression algorithm for images that are currently trending on the Internet. We scraped images from popular subreddits and isolated the text in the images from the backgrounds with optical character recognition(OCR). Then we used Tesseract, an OCR engine, to parse out characters from our text images and stored the text from the image along with the url for the background picture in our .meme file format. This allows for a much smaller file than if the entire image file was stored pixel by pixel. We created a flexible viewing engine through the html5 canvas standard. You can find the viewing engine here: https://shrink-my-meme.herokuapp.com.

Built With

Share this project:
×

Updates