Surveillance industry implies any kind of monitoring for the purpose of information gathering, for safety reasons mostly. Machine vision technologies are used in surveillance cameras that can recognize people, cars, and other objects. Still, observation via cameras plays a crucial role in this industry in general, and that is where facial recognition and face detection are deployed.
Face recognition systems can be fooled if an attacker presents a photo, a video with a certain person to the targeted camera: these are the most common types of attacks here due to the low cost and simplicity of the method. Below, there are other possible attack vectors on surveillance systems.
Smart surveillance cameras can overlook trespassers with the help of anti-detection methods. These attacks are performed by wearing special glasses, face masks and other accessories.
Facial recognition cameras can misidentify wanted fugitives if malicious actors exploit deceptive masks or other tools.
Audio surveillance systems can misclassify suspicious conversation in case of misleading keywords.
There were examples of anti-detection attacks during Hong Kong protests. The protesters were wearing specially-designed eyeglasses or masks. Therefore, it is possible to achieve evasion with the use of makeup, hyperrealistic 3D masks or plastic surgery to make the system recognize an attacker as a different person or not a person at all.
Our team of security professionals has deep knowledge and considerable skills in cyber security, AI algorithms, and models that underlie any content moderation system. Your algorithms can be tested against the most critical AI vulnerability categories that include Evasion, Poisoning, Inference, Trojans, Backdoors, and others.
We offer Solutions for Awareness, Assessment, and Assurance areas to provide 360-degree end-to-end visibility on the AI threat landscape.
Read more on the topic
