The volume of video data has increased exponentially over the last decade or so, with applications ranging from surveillance to retail analysis to evidence and monitoring.
While this data has been extremely beneficial in terms of delivering valuable and important information, the unsafe and insecure manner in which it can be used is concerning.
Redaction, often equated with the censorship of sensitive material in text documents for legal and security reasons, has taken on a new significance in the current climate.
With consumers drowning in a sea of video data and new privacy legislation restricting access and usage without adequate protection, there is an essential need for the ability to conceal important information automatically, at the edge.
What is video redaction, and why is it required?
Video redaction is the process of locating, tracking, and concealing (often via the use of video blurring) any identifiable information, such as faces, people or license plates.
The ever-growing quantity of video data acquired from various sources, along with unsecure ways in which this data may be shared and utilized publicly, has resulted in a growing concern for individual privacy and protection rules.
This concern is entirely understandable, especially in light of recent technological advancements that have resulted in our face acting as the "key" to accessing our phones and bank accounts and our license plates operating as a "pass" to restricted places.
Concerns such as these have prompted the establishment of legislation such as GDPR, CCPA, APPI, HIPAA, granting individuals the right of access, entitling them to a copy of any footage in which they appear, and the right to be forgotten.
Such legislation allows them to have their information deleted or erased from any such media with what’s called a Data Subject Access Request (DSAR) form. Often, having the data deleted or removed is not always a viable option. This is where video redaction comes in since the technique enables you to retain data while following privacy regulations and requirements.
GDPR regulates video, too!
Yep, you read that correctly; GDPR includes video! Essentially, any company that has a video camera is not GDPR compliant until it has redaction (software) capabilities in place to handle "right of erasure" and "right to access" demands, when it comes to receiving that CCTV placard. In this case, GDPR, CCPA, APPI, and HIPAA are all applicable. You can read a case study here, about Triumph Motorcycles in the UK using Sighthound Redactor for CCTV and GDPR compliance.
A burgeoning need for live redaction
We've witnessed an uptick in demand for live redaction, and after receiving multiple requests from our clients, our Computer Vision team got to work on a solution to meet this need. Redacting a live stream may be useful in various businesses and a variety of cases, where a human operator needs to monitor or watch the stream for security purposes, but should not be viewing identifiable, confidential or private information. For example, live streaming surgery in the operating theater (blurring patient faces), Traffic & Safety Control and Smart Cities monitoring (blurring citizen faces and license plates) and CCTV, security and surveillance footage of offices, businesses, warehouses and distribution centers to name a few.
Additionally, it is also required, where privacy laws dictate, that image/video data cannot be saved or stored for more than a certain period unless sensitive information within this video data is redacted. For such scenarios, offline redaction is not applicable and thus the need for live redaction comes into play. For example, GDPR states that the legal limit to storing unredacted footage is a mere 30 days. Storing a video stream for only 30 days is often not long enough when it comes to loss-prevention, safety, security or other reasons. Hence the need for video redaction software that hides sensitive information and allows the user to store the video beyond the 30 day limit.
The solution would need to meet the following requirements:
Straightforward and easy to use
Easy to train personnel
Function at a speed, involving minimal delay
Good processing power
Fully compliant and meets all regulatory standards
High accuracy and low latency
How does Sighthound's Live Video Redaction work?
Sighthound is one of the only companies that provides Live Redaction. Sighthound's Artificial intelligence and machine learning technologies are based on deep neural networks incorporated in the computer vision software we've created over the last eight years.
The Software runs on Sighthound’s Deep Neural Network (DNN) Node Device connected to the already installed CCTV or surveillance camera. The DNN Node receives a live RTSP stream that an authorized user can view, and then once processed, a redacted stream is sent to the Video Management System (VMS) in order to be stored by the customer for more than 30 days (GDPR legal limit). An unredacted stream can also be stored in the VMS by the customer, generally for only 30 days.
Redacted video can either be recorded by Sighthound software, and/or served over an RTSP feed, to be consumed, recorded and displayed by the same endpoint recording the original stream -- but retained longer to stay compliant with GDPR.
Essentially, our technology allows for the augmentation of existing infrastructure to be upgraded and enhanced, rather than replacing it entirely.
Sighthound Redactor video redaction software can also redact audio if stakeholders need to view a situation but not listen to its content. Users may redact audio using Beep, Mute, or Scramble options. They can redact certain portions or the entire recording.
The benefits of live video redaction
Sighthound is focused on helping customers extract insights from video and using it to create business value. Storing a video stream for only 30 days is often not long enough when it comes to loss-prevention, safety, security, customer behavior, workflow optimization or other reasons. The video has real business value beyond 30 days, and thus Live Redaction allows for a longer storage period to extract insights, while still meeting compliance requirements.
Due to the fact that the footage is in real-time, there is also no delay in publishing, releasing, and distributing the video. By on-the-fly exhibiting or sharing video footage, you can increase trust and transparency, improve the customer/viewer experience, enhance compliance, and mitigate significant risk.
These elements have the ability to boost income, promote brand loyalty, and reduce fines.
As such, live redaction has two utilities:
Altering stream in real-time for immediate presentation to authorized viewers and quick publishing
Altering stream in real-time to provide an alternative view to the one provided by the camera, one that complies with longer retention times.
Sighthound would like to help you with any and all of your video redaction needs, whether it be live streams or recorded streams, image files or video files, we have built an easy-to-use tool that anyone can use to protect people’s privacy, reduce the risk of their company or organization and ensure important privacy protections and data compliance requirements are met.