The paper "Invisible Reflections: Leveraging Infrared Laser Reflections to Target Traffic Sign Perception" by Takami Sato, Sri Hrushikesh Varma Bhupathiraju, Michael Clifford, Takeshi Sugawara, Qi Alfred Chen, and Sara Rampazzi explores a novel physical-world attack on traffic sign perception systems used in Connected Autonomous Vehicles (CAVs). The attack leverages the sensitivity of filterless image sensors and the properties of Infrared Laser Reflections (ILRs), which are invisible to humans. The researchers formulate a threat model and evaluate the effectiveness of ILR-based attacks using real-world experiments with four IR-sensitive cameras. The attack is designed to mislead CAV cameras and perception systems, leading to misclassification of traffic signs. The paper demonstrates that the attack can achieve up to 100% success rate in indoor, static scenarios and a ≥80.5% success rate in outdoor, moving vehicle scenarios. The authors also evaluate the effectiveness of the state-of-the-art certifiable defense, PatchCleanser, which is found to be ineffective against ILR attacks, mis-certifying ≥33.5% of cases. To address this, they propose a detection strategy based on the physical properties of IR laser reflections, achieving a 96% True Positive Rate (TPR) and 6.7% False Positive Rate (FPR) in proof-of-concept tests. The study highlights the importance of considering IR-based attacks in traffic sign recognition systems and proposes a novel defense strategy to mitigate their impact.The paper "Invisible Reflections: Leveraging Infrared Laser Reflections to Target Traffic Sign Perception" by Takami Sato, Sri Hrushikesh Varma Bhupathiraju, Michael Clifford, Takeshi Sugawara, Qi Alfred Chen, and Sara Rampazzi explores a novel physical-world attack on traffic sign perception systems used in Connected Autonomous Vehicles (CAVs). The attack leverages the sensitivity of filterless image sensors and the properties of Infrared Laser Reflections (ILRs), which are invisible to humans. The researchers formulate a threat model and evaluate the effectiveness of ILR-based attacks using real-world experiments with four IR-sensitive cameras. The attack is designed to mislead CAV cameras and perception systems, leading to misclassification of traffic signs. The paper demonstrates that the attack can achieve up to 100% success rate in indoor, static scenarios and a ≥80.5% success rate in outdoor, moving vehicle scenarios. The authors also evaluate the effectiveness of the state-of-the-art certifiable defense, PatchCleanser, which is found to be ineffective against ILR attacks, mis-certifying ≥33.5% of cases. To address this, they propose a detection strategy based on the physical properties of IR laser reflections, achieving a 96% True Positive Rate (TPR) and 6.7% False Positive Rate (FPR) in proof-of-concept tests. The study highlights the importance of considering IR-based attacks in traffic sign recognition systems and proposes a novel defense strategy to mitigate their impact.