Driver monitoring is a recent trend for trucking companies and car fleets. Vendors like Smart Eye and Nauto have developed small video cameras that use artificial intelligence (AI) to monitor drivers, and alert them if they appear drowsy or distracted. The manufacturers claim that their systems prevent accidents, which saves both lives and money.
Undoubtedly there is a lot of value in these systems. Alerts can help protect the driver and passengers in-vehicle, while the video feed and analytics information can be used for training purposes. Using the data and video feeds recorded by the devices, car fleet owners can train drivers to recognize different scenarios, helping to prevent future dangerous situations.
However, facial images and video contain the driver’s most sensitive biometric data and personally identifiable information (PII). In an effort to protect the driver’s privacy, videos which can be used to train in-car monitoring AI or in-driver training exercises are being discarded. While this enhances video privacy protection, it denies companies the use of valuable data. This is unfortunate, considering the positive outcome that would result from analyzing driver videos.
Smart Video Anonymization Removes Personally Identifiable Information
Anonymization, the process that removes recognizable features from a video, helps to mitigate the privacy concern. Through video anonymization, the driver’s face is replaced with a computer-generated version. Using AI tools and deep learning models, like those developed by D-ID, the anonymized video will still retain facial features like age, gender, ethnicity, emotion and gaze, but the features will change so that the individual is not recognizable by both humans and facial recognition AI.
Facial anonymization creates several key benefits for trucking companies and car fleets. The video data that’s being generated can be reviewed by the company for analytics purposes, which could lead to the discovery of unknown issues and scenarios that haven’t been covered in the past. These can then be fed back to the in-car monitoring system, making the real-time AI more accurate and improving the safety of drivers and passengers. Second, the videos can be made available for training purposes without violating the driver’s privacy. Lastly, the data from videos, as well as the videos themselves, can be shared with other organizations without concern of driver privacy.
Learn more about D-ID’s Smart Anonymization