These things pop up from time to time, but the detection-evading pattern is a function of the specific detector model (and most likely specific model weights used). It's totally useless when you don't know the specifics of what you adversary is using to detect humans.
I was hoping for something better than "cover your body with a large picture of a scene, and make sure to never turn your body to obscure the camouflage picture.
Once did a project like this, I wanted to make a pair of glasses that blocked facial recognition. Tried using leds to blind the camera without blinding others and worked a bit with making weird graphics like this to confuse the algorithm. My image was made up of other people’s faces though