Method and device for generating confrontation chartlet for resisting AI self-aiming cheating

The invention discloses an adversarial map generation method and device for resisting AI auto-collimation cheating, and the method comprises the steps: inputting a noise data set # imgabs0 # into a detector which is pre-constructed based on a neural network, and obtaining a neural network detection result; calculating a loss value Loss according to the real classification label and a neural network detection result; performing back propagation on the loss value Loss, and updating the noise image n through a gradient descent method; iterating repeatedly until the loss value Loss converges, and outputting a trained noise image # imgabs1 #; the trained noise image # imgabs2 # is converted into an adversarial map; fusing the trained noise image # imgabs3 # with a game article map file, and fusing an adversarial map with a game ground map file to resist AI self-aiming cheating; the confrontation chartlet enables the AI self-aiming cheating program to recognize a game picture into a plurality of player targets, so