The first privacy-preserving
video analytics system.


Video analytics has become more pervasive: consumers demand rights to personal privacy.

Video analysis algorithms are applied in a wide range of domains including entertainment, video retrieval, health care, manufacturing, retail, automotive, transport, livestock farming, home automation, safety, and security.

Today, consumers demand rights to personal privacy, and this is enforced with new data protection rules around the world, like GDPR and CCPA. This global change in how privacy is understood and enforced has a huge impact on vision systems deployed in public and private places: customers now expect that their sensitive data is handled securely.

Nowadays and even more in the future, applying the wrong data privacy strategy can cost an organization billions in fees and damages.

With Goateye we want to provide a
new layer of abstraction.

Producing anonymized video material while maintaining all the non-sensitive information useful for data analytics purposes.

Privacy, technology and budget

Discover the features
of Goateye.


Only captures non-sensitive data

Goateye is the next generation camera that ONLY captures non-sensitive data by producing an anonymized video stream where people have been dehumanized into anthropomorphic parametric models. In this way, everybody can process the video with no worries about privacy-related rules.


Useful metadata for different industries

Our anthropomorphic parametric models already carry out useful non-sensitive information such as posture, age, gender, or even actions performed. Those metadata can be useful across a wide range of industries like retail, smart cities, fitness, and media.


Keeping the scenario unaltered

For data analytics purposes that do not involve humans but target, for example, vehicles, machinery, animals, or buildings, our solution can provide a stronger level of privacy by completely erasing people while keeping the scenario unaltered.


Other features:


Output the processed video metadata only


Stand-alone software for video anonymization


Low power and cost friendly

How goateye workS?

Advanced and optimized Artificial
Intelligence and Computer Vision

Step 1

Human localization

It utilizes a GoatAI proprietary people detection algorithm for human localization specifically designed to have few false negatives.

Step 2

Data elaboration

Localized people are then selectively elaborated by a series of GoatAI proprietary algorithms for people segmentation, classification, pose estimation, and video inpainting.

Step 3


The collected information are then used to censor the video accordingly, outputting non-sensitive metadata and video stream.