Everyday, Truepic’s entire team works diligently to ensure our verification technologies achieve the highest possible level of credibility. Our mission is to ensure that every photo and video captured using Truepic’s patented Controlled Capture technology has verifiable origin, contents, and metadata. Our team aims to deliver on that promise through relentless focus on the security of our platform, the quality of our image forensics techniques, and the protection of our core technology.
Unfortunately, there are current technological limitations that do exist today which we feel should be disclosed publicly. These limitations are rooted in natural life-cycle of software on older mobile devices, and an age old photo forensic challenge called rebroadcasting. We address both on this page to ensure full transparency on technology’s current capabilities.
Legacy Devices & Software present critical security issues for all industries, developers, and end-users that depend on them. Mobile devices, their operating systems, and the camera software that runs on them all suffer from the same natural software lifecycle of deprecation. Security mechanisms in both iOS & Android have continued to evolve, with new provisions such as more secure operating systems & boot protocols to prevent tampering with built-in software. Unfortunately, many older devices do not receive these updates, leaving them open to potential vulnerabilities.
Thus, Truepic can not fully ascertain whether older devices conform to high enough security standards to ensure the data originating from them is trustworthy. We flag device integrity issues, if detected, on our verification pages as a yellow warning flag.
Our team will continue to invest in practices that allow us to better segment trustworthy devices vs. outdated devices to ensure the highest level of credibility in the images captured through our technology.
Rebroadcasting, very simply, is the act of taking photographs of other photos. By performing this photo re-capture technique, the photographer is essentially distorting the reality in-front of the camera lens. While Truepic can verify the image data coming into the camera lens is accurate, we can not be 100% certain that what’s been captured is a truthful representation of what’s actually transpired at a certain time & place.
This problem of rebroadcasting has existed since the beginning of photography, and has posed many challenges for image forensics experts. For those interested in learning more, Hany Farid, a member of Truepic’s advisory board and a world renowned expert in the field, has authored this paper describing the issue & various detection methods.
Truepic is working hard to ensure that rebroadcasting can not be performed without being detected through our Controlled Capture technology. To-date, we’ve invested heavily in computer vision & machine learning techniques to detect abnormalities in front of a Truepic enabled camera, and to flag these detections automatically. We’ve made great progress on these techniques, and currently flag a very large percentage of rebroadcasting attacks.
Unfortunately, some limitations do still exist in these methods, which requires ongoing development by our team. We’re hiring top talent to help us in this area, including senior computer vision & machine learning engineers. If you’re interested in helping our cause, please contact us here!
This page was last updated on January 1st, 2019. We will continuously update this page as developments transpire in an effort to remain a transparent & trustworthy verification source for our clients, partners, and consumers.