A treatise on the use of block chain to authenticate video or pictures as real
Eric Oyen
eric.oyen at icloud.com
Sun Oct 14 06:53:55 MST 2018
What follows is a treatise on the use of block chain technology to authenticate video and pictures to avoid the fake news problem that seems so pervasive of late. I already published this on LinkedIn, but would like your thoughts on it to flesh it out further. Who knows, even publish a full paper on it and really become a published citizen researcher. :)
***
How to determine if a video or picture is real or fake?
The question seems simple enough, doesn't it? However, with today's advancing computer technology in image manipulation, you can no longer be absolutely sure that the photo you are viewing, or the video being broadcast is the "real McCoy". There have already been examples, courtesy of the Russian Internet Research Agency, of video that appears real, yet has been entirely faked. This is called False Evidence Appearing real and it can be a real problem when trying to sort out real documented events from those which have been faked. So, how do you certify that a picture or video is real and unaltered? The answer is a recent technological advancement called "Block Chain".
What is Block Chain?
Block Chain technology is an outgrowth of cryptographic hashing in combination with cryptocurrencies. It is a method by which you can certify and reference a given event as real using the P.O.W. of the block chain itself. P.O.W. refers to the Proof of Work crypt-graphic hash function used to certify that a particular job was completed using a specific set of rules and other criterian. This model is used in cryptocurrencies to certify, by consensus, that a given chain of crypto hashes was done correctly and completed before anyone else did the same function.
OK, So how does this relate to using block chain to certify video or picture images as real?
Lately, there has been a number of incidents where the press (or some third party) has tried to claim that a given video (or picture) has been altered or is not as presented. This has created a situation where you cannot always take the evidence of your eyes and ears at face value. This is the problem of perception being reality, and as we all know, perception can be altered in a myriad of ways to make reality appear different than it really is. How do we overcome this conundrum? We use the concept of block chain to certify that a picture or video has been unaltered by taking a digest of information found in the picture, combined with external references (such as GPS, local clock, etc) and create a cryptographic block that represents the picture (or each frame of a video). With video, this would also include the sound track. Once done, the block chain would then be embedded within the picture or video in question as an integrated data stream that can be read by an external program. This program would take the exif data from the picture (or video) and use that to decrypt and certify the picture (or video or frame therein) by comparing the decrypted output to the exif data along with the digest of the image(s) involved. Of necessity, this would be both computationally complex and intensive, but the result would be a certified image (or set of images).
How to implement this in a real world scenario.
implementation is rather simple from a layman's point of view. The entire process would be hidden from the end user. The imaging device would have 2 additional components installed and running in background. The first is a gps receiver that provides location index data as well as a certified time code. The second device would be an ASIC (Application Specific integrated circuit that would handle hashing the images and generating the block chain. The second item in this case has yet to be incorporated into the latest smart electronics, but given the need for certification of video or images as real, the justification of it being installed is clear enough.
Implementation
Now, part of this has already been implemented. The use of GPS for location services is built into just about every modern imaging device on the market these days (except for the "low end cheap stuff"). This means that some of the data is already setup to be used. Now, until ASIC's are actually installed in hardware, a handoff to an external device can be implemented in the cloud. That device would read the image and the exif data included (GPS location, time, image size, color bit range, brightness, etc.) and generate a block chain of that image based on those criteria. There are several methods that can be used depending on the level of certification one chooses to use.
Level 1 certification would involve a simple direct sample of the image (or video stream) reading the exif data and calculating a block chain based on image size, number of pixels, color bit range, brightness and a random sample of pixels in the entire image (or image frame of a video) combined with the time code and GPS coordinates to create a simple block chain. This would minimally certify the image as unaltered in any way and would also be the fastest method.
The second level of certification would involve the fast method of level 1, but would break the image (or image frame of a stream) into sectors and create a block chain of each sector. This would be a sub chain to the primary chain built for the image itself (level 1) It would, by necessity, include more data (including line count, number of pixels per sector, etc.) and could be used to compute the equivalent of a level 1 chain, but vastly more detailed.
Level three would involve doing a line by line hashing of the pixels in the image, using the number of pixels in each line, number of color or brightness changes, bit rate of the image itself and then combine with the GPS and timing info to create a block, each line would be it's own block and would be used to compute a complete chain of the image (or audio/video streaming frame). This method would be the most computationally intensive and complex, but the level of certification would mean that it could be verified as authentic under rules of evidence in a court of law or when used in a news item to certify that the data stream was shot on location or is otherwise authentic in content.
A nice additional feature of this set of methods, especially in a video stream, it can be used to reference specific areas of the stream (much like a built in time code was being used on old video formats for over the air transmission) Also, because each "Frame" in the stream has it's own block chain, a single frame or series of frames can be used independent of the others or as a single stream/picture in multimedia content (such as presentations). Since the data, itself would not be visible to the end user (unless a program designed to read it were used, or that feature were included in the multimedia playback program), the user wouldn't be concerned too much. As
Security and authenticity
a security model, there are benefits that become immediately apparent. The first of these is determining if the stream or frame had been altered. This could be done by computing a new block chain from the existing video/audio or picture content and comparing it to the existing. Any differences would result in a failure and even show where the alterations took place. For evidence certification, this would be a definite requirement to insure authenticity in chain of custody. This very same block chain could also be used to encrypt the data stream (audio/video or picture frame) for secure storage, or also allow recovery should the data be otherwise corrupted. Something like this is already in use today in the guise of bittorrent. The function is similar in that the original file is hashed, broken into chunks, transmitted and then reassembled using a provided key. Only block chain would take this a step further and allow certification of the data as well. I must point out that the differences between torrent and block chain make a 1 to 1 comparison irrelevant and is only used to show simple comparative use.
In conclusion
Block chain shows promise as a method to authenticate, in real time, events being recorded in audio/video or as pictures at a given time and location. It also shows promise as a method for preserving evidence and chain of custody where multimedia content might be used in a court of law or during an investigation leading up to a court proceeding. Block chain Also shows promise as a protection and recovery mechanism that could be used concurrent to other methods for data preservation and recovery. In short, this technology has a wide range of uses that has not yet been explored and shows promise in being used in everyday life. This technology would effectively remove the conundrum of flawed perception being made reality and would effectively nullify some elements of propaganda currently in use.
More information about the PLUG-discuss
mailing list