SRL Archive - Research & Development Lounge

Hi. I made a simple java program to record myself clicking randomly generated rectangles appearing in random places in the screen.

My idea is to try using machine learning (neural networks) to predict your next mouse movement given several previous mouse movements and the target rectangle. This could in theory be used to generate human-like mouse movements but I'm not at all sure if it will work.

Quite busy so I don't know when I'll be able to test this. Just decided to share the idea and the program and the data i have generated so far.

Edit: Upload limits are so low that I will only share some of the data for now: http://pastebin.com/9xtUguRM
Data format: startX,startY,startTime|movement1X,movement1Y,movement1Time|… movements …|mousepresX, mousepressY, mousepressTime|mousereleaseX,mousereleaseY,mousereleaseTime|rectangleX1,rectangleY1,rectangleWidth,rectangleHeight

MouseRecorder:[ATTACH]24236[/ATTACH]
Warning: ugly code.

Do I need netbeans or something to run this? Sorry for being a noob, I am not trained in the java ways.

java build/classes/mouserecorder/MouseRecorder
should work I think. No netbeans required.

Quite interested to see how Trilez got on with his collection of human mouse movement data ;)

From: Kave
Quite interested to see how Trilez got on with his collection of human mouse movement data ;)

Did he actually get the mouse movements working?

Also something like this: https://github.com/CynicRus/SMR Written in Lazarus.

lol i recall something very similar to this being done on another botting platform, none the less good job!