When the recent infamous/famous popularity of sarahah.com/, I sought out to find ways to remove all the links and images while scrolling through my facebook feed as soon as they came into being (Using Mutation Observer API) with the help of chrome extensions. I solved the links thing pretty quickly. But I couldn't find a fast and efficient way to do it with images.

Some of the other chrome extensions sought out to solve the same problem, but were too sloppy probably because the image processing that they were doing (for example with TessractJS) was too slow. How could it be achieved? (Perhaps through a machine learning algorithm? Can ML/DL algorithms run in browsers for relatively such a small task?)