Поиска по сайту
 

Video content verification using blockchain technology

Скачать в формате pdf
Video content verification using blockchain technology
Viacheslav Voronin
Moscow State University of Technology
«STANKIN»
Moscow, Russian Federation
Evgenii Semenishchev
Moscow State University of Technology
«STANKIN»
Moscow, Russian Federation
Aleksandr Zelensky
Pro-rector for Research Work and
R&D Politics
Moscow State University of Technology
«STANKIN»
Moscow, Russian Federation
Iliya Svirin
CJSC Nordavind, Moscow
Moscow, Russian Federation
Andrey Alepko
Dept. of Radio-Electronics Systems
Don State Technical University
Rostov-on-Don, Russian Federation
In distributed high-performance systems that allow
processing large data sets, the important task is the analysis
and verification of the video data. The problem of confirming
data is relevant for many areas. The need for this arises when
fixing offences, banking, remote management, confirmation of
actions, etc. In this paper, we describe an approach to verifying
video received by a mobile device like mobile phone, tablet or
PC, equipped with a camera and controlled by the operating
system (Windows, Android or iOS). The proposed algorithm
uses the procedure of entering the Swype code using the
movement mobile camera. To improving the accuracy, we use
additional information from different mobile sensors like
accelerometer, gyro, barometer, and GPS. In the process of
data verification, video transmission to the server is not
perform what ensures the privacy of the captured video data.
The server and mobile device stores data about the file size, the
date of its recording, the time, the data device and its position.
Keywords— verification data, Swype code, mobile device,
PROVER, blockchain
I. I
NTRODUCTION
The revolution of smartphones, which started ten years
ago with the advent of the iPhone and tablet computers, had
a significant impact on the ways users communicate in the
network. Everyone suddenly became the creator and
distributor of content, causing, as a consequence, his
exponential growth. Such data has a large dimension. They
can be multi-layered and contain data on the gradient of
temperature, 3D, stickers and explanations, and so on. Big
data requires large calculations, which are most often
impossible on a mobile device. To analyze such data it is
necessary to use Smart Cloud.
This has significantly changed and continues to change
many sectors of the economy, launching the processes of
digitizing the world around us. Photo and video content are
actively used not only for entertainment and educational
purposes but also for other needs, including economic and
legal nature - financial, insurance, judicial, medical and other
services. In this regard, there is a serious need for an
independent, decentralized service that objectively
guarantees the authenticity of the created video content and
protects it from possible forgery and unfair editing.
The authenticity of digital video recordings of events and
facts of commercial and legal value is often questionable
because video files can be edited for forgery purposes, by
using a virtual camera (emulator). Attributes such as the date
of the video could also be artificially altered.
In this connection, the actual task is to operate on
verification data. Verification will guarantee the originality
of video data, confirm the time, location and duration of the
video.
II. SOLUTIONS
FOR
VIDEO
VERIFICATION
The authenticity of the content and the rights to it
(copyright) today are mostly being approved at the level of
platforms for storing and demonstrating video materials. An
example of such a solution is Google content ID, which
allows only to confirm the time of downloading a video on
YouTube, this is the basis for the assumption of its
originality, based on the principle of presumption of
authorship (a person is considered as an author until the real
author disputes this fact).
The drawback of this and similar solutions is that they do
not allow to restore the time of real video recording, its
originality, and integrity. Also, these solutions work within
their platforms "only on YouTube," are provided manually
by service administrators upon application and always
depend on the opinion of the service administrators. That is,
there is always a place for the human factor - subjectivity or
simply error [1, 2].
The authors try to protect the content by watermarks and
logos on the video screen, but this can only help them in the
subsequent contestation of its authorship, but not to prove the
fact of forgery. That means that in legal and financial matters
this approach is completely inapplicable.
At the present, the blockchain community already has
various online electronic notary services, which make
possible to certify the existence of "Proof of existence" and
the authorship of "Proof of ownership" of all kinds of files
and documents and digital content:
Block Notary - a service that helps to create "Proof
of existence" of any content (photos, files, any
media) using the TestNet3 or Bitcoin network. The
frontend system is a mobile application for iOS that
registers a document hash in a blockchain.
Emercoin DPO Anti-fake - technology based on the
Emer platform allows to create the item (product) a
unique digital passport stored in a decentralized
database - blockchain and provides services for
208
2018 IEEE International Conference on Smart Cloud
978-1-5386-8000-1/18/$31.00 ©2018 IEEE
DOI 10.1109/SmartCloud.2018.00042
managing this passport. It is focused mainly on the
offline segment - it helps to register individual
details (VIN, IMEI numbers) in the system to the
shield of real goods from fraud.
Stampery - blockchain technology that can verify e-
mail or any files. This simplifies the process of
verifying letters by simply sending them by e-mail to
a specially created mailing address for each
customer. Law firms use Stampery technology for a
very cost-effective way of document verification.
https://www.ascribe.io/ - service for registration of
copyright, further control and distribution of digital
content. It is positioned for digital works of art. It
offers to register a work, put it on sale in a secure
marketplace, and then monitor its use
(demonstration).
https://letsnotar.me/ - an easy service that
automatically stores to blockchain a hash of files that
were uploaded to it. It could launch on a smartphone,
allows to get access the camera, take photos and
videos and to hash them. However, it cannot
guarantee that the video is recorded from a real, not a
virtual camera - so it does not protect against
forgery.
All these services unite one thing: they can assure a hash
of a file that has already been written before or supposedly
written down from the camera of the device, but they cannot
guarantee its originality, integrity, and authenticity of the
video. They do not protect against forgery and unfair editing
because they do not have the appropriate technology to
verify the video content being created [3].
Everyone can record video at any time. It is fast, simple,
available and convenient. Many companies use video content
in their business processes – it gives them a competitive
advantage. But technologies allow others to edit and fake
videos, which seriously undermines the credibility of that
format.
In this work we use PROVER technology allows
guaranteeing that this content was created at a specific time
and place, from the camera of a device, to ensure that there
are no signs of forgery and editing. The PROVER can
become a functional addition, extending the capabilities of
the services discussed above and the level of trust to them.
Having the opportunity to 100% guarantee the authenticity of
the created video file, these services will, for example,
provide services for notarization of the authenticity of video
statements.
PROVER technology use cases:
Fintech, when it comes to authentication of loaner’s
identity. Banks and financial services providers can
verify clients during the customer onboarding
procedure with lower risks of identity theft. Clients
can perform remote actions within the system.
Auto insurance, dealing with fraud of security
addition in contractual terms. Clients can record
video evidence in case of insurance loss for
insurance on demand and car sharing services.
Insurance companies can receive video evidence of
performed services covered by insurance (medical,
repairs).
Simple video proof of ownership, working for
bloggers as example. Individuals and organizations
can store a timestamped video hash on Blockchain as
a digital proof of ownership for original and
authentic video content.
Public statements, to keep secure what is said before
editing. Public speakers, celebrities and businessmen
preventing reputational damage from montage, CGI
and rapidly growing sophisticated machine learning
algorithms and tools able to edit or generate fake
video statements.
Crowdsourced media platforms, to keep ownership
of actual content makers. Public and crowdsourced
news platforms can validate the authenticity,
exclusivity and timing of video news submitted by
individual contributors.
Video platforms with user-generated content, of
exclusive contributions. Both users and platforms
can prove the authenticity and exclusivity of user
generated video content and share monetization
proceeds.
Online dating, keeping users from fake content.
Users can be sure that they are chatting with a real
person on video dating websites and services.
Outsourced reporting, carrying out remote inspection
of actual performance. Employers and contractors
can exchange authentic and time stamped work
reports.
Accident reports, to proof someone’s position in
court using evidence. Both parties involved in a
traffic accident can rely on a video recording to
prove authenticity of time, date and record of the
accident.
Notary actions, allowing people verify video content
without visiting special parties. Parties can maintain
a Blockchain video database of trusted “handshake”
agreements.
Home education and exams control, allowing remote
authentication of online courses. Video recordings
can be used to confirm authenticity of a person
taking an online exam.
III. THE
TECHNOLOGY
OF
PROVER
The developed service contains the following
components (www.prover.io):
A mobile app that installs on a smartphone and
launches with the camera turned on or initiates the
launch of the camera itself.
A set of algorithms and utilities for integrating
PROVER technology into third-party solutions and
services.
Smart Contract PROOF (only for implementation
based on the Ethereum platform).
The presented technology allows performing the
following actions:
Video footage is produced by a real video camera
integrated into the mobile device, and not emulated
by a virtual video camera;
The video material is complete, not edited, without
gluing and insertions;
The record was made in a certain period.
To verify the data, two approaches are used, based on
changes in sensor-fixed mobile devices and analysis of video
data received by the camera. The first approach uses data
obtained with the help of an accelerometer, gyro, barometer,
and GPS. The second approach is based on preliminary
identification with visual Swype code.
209
The algorithm of verification with Swype code
the classical Swype code, which is being entered by
moving user's finger on the touchscreen, forming a
continuous line connecting the points, shown on the screen,
in PROVER technology, the Swype code is being entered by
moving the smartphone with the camera in a record mode.
On Figure 1, we show an example to enter Swype code
using video stream from the mobile camera.
Figure 1. Graphical representation to enter Swype code.
On Figure 2 shows the algorithm for search verification
using Swype code.
The following steps realize the algorithm presented in
Figure 2:
Step 1: Receiving video data. The video data is
transferred in parallel to the screen of the mobile device and
the input of the data search and verification program. Video
analysis is carried out according to the location of objects in
the frame. The algorithm of search is always looking for a
condition for launching verification. This condition is the
circular movement of the mobile device.
Step 2: Search for frame offsets. This section of the
algorithm is based on the application of the phase correlation
method. This method is described in the paper [4], end
realized by the following steps:
2.1) For two frames
()
111
,
f
xy
and
()
222
,
f
xy
, we find
the Fourier transformation
()
111
,
F
uv
and
()
222
,
F
uv
[4-7].
2.2) The cross-phase spectrum of the two spectral
functions
()
111
,
F
uv
and
()
222
,
F
uv
stands for the ratio:
()( )
()( )
12
111 2 22
111 2 22
,,
,,
FF
Fuv F uv
R
F
uv F u v
=
(1)
The resulting expression is a spectral function with a unit
modulus whose phase is equal to the phase difference of the
functions.
Figure 2. The verification algorithm with Swype code.
2.3) We perform the inverse Fourier transform, which is
the phase correlation. Since on adjacent frames, there are the
same elements up to bias:
() ()
22
2
222 111
,,
size size
ua vb
i
xy
F
uv e Fuv
π
§·
−+
¨¸
©¹
=
(2)
In this expression,
b,a
are the peaks of the delta
function.
2.4) In the case of the similarity of the frames, peaks will
be present. The height of the peak determines the degree of
similarity, and the peak position corresponds to the
displacement of frames relative to each other.
Step 3: Search for circular motion (Fig. 3).
210