Cassius Posted February 16, 2020 Posted February 16, 2020 Salut mes amis!! Excuse me for my english, it's my third language. I am a CS major at a university in my Master's I am making a bot that will work through screen grab technology. I will be training the model to identify pixel locations that contain iron ores, this will be the first ore I will work on. It will be designed as a F2P mining training bot (hence only iron ore for now). If you could, would this community like to share with me screenshots of their gameplay of them mining iron ores? Those who share their screenshots will get access to the bot. Effectively, I am going to try and accumulate at least 1,000 images of in gameplay images, then work on capturing real mouse movements to implement during gameplay later on. You are welcome to block out your username and password. You can upload them to here, no sign in required: https://drive.google.com/drive/folders/1Qy6dK-Oi3K_Ud0PeGXFaA3rwHZLZRc6M?usp=sharing I will post updates as the project progresses. Cassius
Quickscope420 Posted February 16, 2020 Posted February 16, 2020 would extracting the iron rock model and rendering from multiple angles also be a way of training the bot?
Cassius Posted February 16, 2020 Author Posted February 16, 2020 57 minutes ago, Quickscope420 said: would extracting the iron rock model and rendering from multiple angles also be a way of training the bot? This is potential but there is no background interference or text in the image (including the player inventory) so training on that wont give optimal results. I can make screenshots, but I was hoping to crowdsourced the task because it is very time consuming. I'm only really doing this because I know I can and I'm a bit bored.
