Skip to content

Milestones

List view

  • Testing the models using the three different manipulated dataset for the different attacks: - Patch Attack: Sticker attached to the road signs; - Bent Attack: Malformation of the structure of the sign (Only the sign itself, not the post); - Camera Manipulation Attack: Filter added to the camera to simulate hardware malfunction on the car.

    No due date
    5/5 issues closed
  • Training and validation of the four models - Random Forest; - CNN; - AlexNet; - YOLO The Training and Validation will be done with the original dataset (the non manipulated one)

    No due date
    4/4 issues closed
  • This phase will include three main steps: - Map and Object Customization; - Data Collection: Acquiring raw images; - Data Preparation: Image modification to make them compatible for the training, validation and testing.

    No due date
    4/4 issues closed