Automated video and image anonymization using AI

Michał Mikulski
Protecting personal data should be a priority for companies in 2022. Taking care of data privacy can be done easily, using automated video and image anonymization. It doesn't require hi-tech equipment or a dedicated specialist.

Data production and data collection are on the rise every year. Researchers from the Northwestern University graduate program estimated that a single person produces 1.7 MB of data every second. That’s 2.5 quintillion bytes of data every day. Another stunning piece of information is that 90% of data was created in just the last two years alone. To put data production into perspective, an average Londoner is caught by CCTV cameras 300 times a day. This data is stored somewhere. Personal privacy is closely related to consent, and it is certainly impossible to get consent from every individual that lives in London. As individuals, we consent for our location to be used while navigated by Google Maps. Do we consent to be caught in a CCTV crossfire?

GDPR is a good thing

European GDPR and similar regulations across the globe show that citizen privacy is essential and should be addressed in a thoughtful manner. The main building block for GDPR puts pressure on data collectors to make sure that data is not relatable to individuals. That means collecting, sharing, or working personal data is legal as long as it cannot be related to an individual person. Even when it is only for the eyes of data protection specialists employed in companies, data should be anonymized.

There is a good reason for that. We can easily imagine a scenario when a huge amount of data from a multinational company is acquired by hackers. Without data anonymization they can use for example internal CCTV videos to track company employees, blackmail CEOs, to demand ransom. To make things worse, such data collection can be further processed by AI facial recognition algorithms.

Thanks to automated image and video anonymization, big data can be used to make better solutions, but not to track a single person. We still want to have a great Google Maps experience, which requires a huge amount of images being collected. But properly anonymized data is not personal, thus secure from breaching an individual's privacy. The keyword is: properly anonymized.

Automated image anonymization

Anonymizing thousands of photos manually can be time-consuming, expensive, and inefficient. The same can be said about even more challenging video anonymization, with hours of CCTV video recorded by the average company's security cameras. But anonymization can be done swiftly and fast, without huge budgets and using specialized equipment. Thanks to AI using computer vision, the app is capable of detecting faces on images and videos. When it's done, will anonymize sensitive personal data in an instant.

The most important part is that doesn't require a high-end PC. A company data specialist or assigned employee just needs to download a demo version of software, choose multimedia to anonymize, a directory for saving new files, and press START. It's as easy as that.

The free, demo version of is available for download HERE .