general multimedia trust authentication framework...

Research ArticleGeneral Multimedia Trust Authentication Frameworkfor 5G Networks

Ling Xing1 Qiang Ma 2 Honghai Wu 1 and Ping Xie 1

1School of Information Engineering Henan University of Science and Technology Luoyang China2School of Information Engineering Southwest University of Science and Technology Mianyang China

Correspondence should be addressed to Qiang Ma maqiang my163com

Received 12 January 2018 Revised 24 April 2018 Accepted 16 May 2018 Published 28 June 2018

Academic Editor Jinsong Wu

Copyright copy 2018 LingXing et alThis is an open access article distributed under theCreative CommonsAttribution License whichpermits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Due to the varieties of services and the openness of network architectures great challenges for information security of the 5Gsystems are posed Although there exist various and heterogeneous security communicationmechanisms it is imperative to developa more general and more ubiquitous authentication method for data security In this paper we propose for the 5G networks anovel multimedia authentication framework which is based upon the trusted content representation (TCR) The framework isgeneral and suitable for various multimedia contents eg text audio and video The generality of the framework is achieved bythe TCR technique which authenticates the contentsrsquo semantics in both high and low levels Analysis shows that the authenticationframework is able to authenticate multimedia contents effectively in terms of active and passive authenticating ways

1 Introduction

The requirements for higher data transmission rate andbetter user services experiences promote the developmentof the 5G cellular networks Compared with its precedentwireless communication standards (ie 1G 2G 3G and 4Gnetworks) the 5G system is proposed to support wider rangeof connected devices types and various service applications[1 2] It is estimated that the 5Gwould become the ubiquitousinformation infrastructure network in the near future [3]Along with the growth of the users and services in 5Gnetworks come the security problems eg data privacy dataintegrity and data authentication [4ndash6] Ensuring the securewireless communication together with the security of datatransferred is mandatory for the success of the 5G networks

Various types of multimedia content proliferate overthe wireless networks Especially equipped with the socialmedia techniques people find it is quite convenient to com-municate on the mobile devices Although the multimediacontents bringmuch convenience for the information sharingover the 5G networks security concerns of the contentscannot be overlooked [7ndash9] Due to openness feature ofthe underlying network the received multimedia contents

should be carefully examined For example questions shouldbe raised whether the obtained video or image has beenattacked by malicious purposes or whether the video hasbeen deliberately altered to screw its original meaning Thusthe security of multimedia contents must be ensured to keepthe integrity of the contents as intact as the original

As far as the formats of the multimedia are concernedmethods to authenticate the contents are different in termsof purpose efficiency and validity Those methods or algo-rithms for authenticating the multimedia contents can beroughly classified into four groups ie watermarking-basedencryption-based streaming-based and robust-hashing-based methods

The watermarking-based authentication methods formultimedia usually adopt the watermark embedding andextraction approaches For example Md Asikuzzaman et al[10] proposed an imperceptible and robust blind watermark-ing for video in which the watermark was embedded into thelevels of the dual-tree complexwavelet transform coefficientsThere are other multimedia watermarking methods whichemphasize the features of the watermark eg the contentdependent video watermark [11] and wavelet-based multiplewatermarks [12] Although the watermarking methods can

HindawiWireless Communications and Mobile ComputingVolume 2018 Article ID 8974802 9 pageshttpsdoiorg10115520188974802

2 Wireless Communications and Mobile Computing

provide the authentication of contents in some degree theembedding process itself harms the integrity of the content Inaddition one watermarking method can only deal with somespecific attacks which hinders these methods from beingwidely adopted to cope with the various attacks existing inthe 5G networks

The encryption-based multimedia authentication meth-ods usually employ the encryption algorithms eg RSAalgorithm and DES algorithm to encrypt the whole or partsof the contents For example Zafar Shahid et al [13] presenteda selective encryption (ie the truncated rice code and theExp-Golomb code) approach for video of High EfficiencyVideo Coding (HEVC) which shows the characteristics offormat compliant and real-time property Similarly GlennVanWallendael et al [14] proposed the encryption for HEVCby utilizing the differences of intraprediction mode the signof motion vector difference and the residual sign Note thatthe primary goal of the encryption-based methods is toensure the confidentialityThe time overhead incurred by theencryption and decryption process is often a nonnegligibleissue for the social media devices

The aim of the streaming-based multimedia authentica-tion methods is to ensure that the received contents havethe integrity property which is verified by the recipientsCurrently many research works have been focused on thisarea eg Kang et al [15] studied the pollution attackdetection and prevention method by trust for peer-to-peerstreaming Lu et al [16] proposed a privacy protectionmethod based on trust for peer-to-peer data sharing networkand Cheng et al [17] authenticated the live streaming mediaby means of TESLA- (Timed Efficient Stream Loss-tolerantAuthentication-) based protocol One category of streaming-based authentication methods is the graph structure typeeg the chain line approach the tree approach and thebutterfly approach which treats the content packets asindividual ones and exploits various packetsrsquo hash attachingmeans [18] The other category is based on the multimediacoding structure For example Kianoosh Mokhtarian et al[19] proposed the authentication for Scalable Video Coding(SVC) streams whose hash appendence follows the codingstructure of SVC

The robust-hashing-based methods to authenticate themultimedia have the best performance in terms of robustnesswhen the authentication is applied under the network cir-cumstances The robustness of the hashes means that hashesremain the same or change a little after the perceptual contentpreserving operations are being conducted For exampleLokanadham Naidu Vadlamudi et al [20] proposed a robusthash algorithm by using features of histogram for imageauthentication Due to the robustness and sensitivity of thiskind of method it is much superior to the other three meth-ods when it comes to the network packet loss phenomenonin 5G networks Since certain packets of multimedia contentsmay be lost because of traffic jams or network failure contentsintegrity verifying process for the recipients should take intoconsideration the robustness of multimediarsquos representation

Another aspect we should bear in mind is that all thefour methods are for the low level semantics authenticationwhich treats only the integrity either in pixels level or

in frames level This absolutely lacks completeness for theauthentication of multimedia contents of the 5G networkssince high level semantics of the contents are overlooked bythe current methods For instance the contentsrsquo high levelsemantics including but not limited to title name authorand format should also be authenticated Therefore in thispaperwe endeavor to tackle themultimedia authentication byproposing combined authentication method for multimediacontents for 5G networkWe propose a generalized multime-dia trust authentication model based on the trusted contentrepresentation (TCR) method where the generalized termmeans that it can be applied to various multimedia contentformats Also we provide the features needed to authenticatethe integrity of the contents

The rest of the paper is organized as follows We describethe architecture of asymmetric wireless communicationchannel in Section 2 where the asymmetric term means thatone channel is safe and is used for authentication informationtransfer while the other one is open and used for multimediacontents transfer In Section 3 we illustrate the TCR indexingtechnique and show the features of TCR In Section 4we explain the framework of the general multimedia trustauthentication for 5G networks in which the philosophy ofasymmetric wireless channel is adopted Then we concludeour work in Section 5

2 Architecture of Asymmetric WirelessCommunication Channel

21 Security Threats to 5G Networks Since the 5G networksare open and insecure the multimedia contents whichare transmitted over them are prone to various attacksby malicious purposes From the perspectives of networkcommunication system we summarize security threats tomultimedia into three groups

(1) The Originality of the Multimedia May Lack Trust Cur-rently users can upload from the mobile devices the variouskinds of multimedia contents up to the video or audiosharing websites by various means However the effectiveand efficient mechanisms to supervise the legacy of usersrsquobehaviors are still unavailable While people find it quiteeasy to collect edit and distribute the multimedia contentsby modern multimedia processing tools (eg PhotoshopIllustrator) people should think twice about the originalityof the received multimedia In other words the user whopublishes the multimedia and claims the author of themultimediamay likely not be the ldquotruerdquo authorOriginality ofthe contents should be checked in some manner This incursthe problem of intellectual property infringementThereforeit is urgent to find a good way to verify the originality of themultimedia

(2) The Channel to Convey the Multimedia May Be AttackedThe openness of the 5G network protocols poses serioussecurity problems to the contents being transmitted overit The notorious attack ldquoman-in-the-middlerdquo is the mostcommon threat to the contents The attacker maliciously

Wireless Communications and Mobile Computing 3

NetworkInteractivity

AuthenticateInformation for User

Trusted Content AuthenticationCenter

Logical Broadcast Channel

Terminal

Wireless Communication Channel

User Required AuthenticationCenter

MultimediaContent

Information

Information Aggregation

Figure 1 Asymmetric wireless communication channel architecture

hurts the multimedia via many means eg video framesrsquorearrangement dropping inserting and altering The pur-pose of the attack is to destroy the contentsrsquo integrity whichdistorts the recipientrsquos understanding of the contents Notethat there are many free packets sniffer tools on the Internetwhich makes the attacks on the transmission process ofmultimedia much easier We state that the threats to themultimedia on its journey from the sender to the receivershould be taken into consideration and the integrity of thecontents needs to be ensured

(3) The Identity of the Multimedia Recipients May Lack TrustRecall that current authentication for the mobile recipientrsquosidentity is mainly based on the user-password method Thisbrings security problems once the legal recipientrsquos platformis attacked For instance an attacker may break into andobtain the legal recipientrsquos password to certain multimediaserver which results in the illegal copy or broadcasting ofthe contents Another scenario is that the multimedia senderrequires checking the identity of the receiver and the integrityof contents received by the receiver eg video conferencingTherefore both the platform security and the received multi-media integrity are necessary to be authenticated

22 The Asymmetric Wireless Communication Channel Theconcept of asymmetric wireless communication channelfor 5G networks is shown in Figure 1 We use the termasymmetric to describe the unbalanced properties of twochannels existing in the architecture One channel is theordinary wireless communication channel which is reservedfor the multimedia contents sharing The channel is dualand its capacity is large enough to support various kinds ofservices It is not transparent to the users and not attacks-freedue to its openness characteristics The other channel whichis more akin to logical broadcasting communication channelis introduced to transfer the authentication information usedfor multimedia contents This channel is single-way andtransparent It does not exist in current networks and shouldbe created by means of broadcasting techniques Its capacityis much less compared to the former one However it is closed

and safe since the information transmitted over this channelis confidential

The Trusted Content Authentication center (TCA center)in Figure 1 serves as a trusted third party It collects themultimedia content information from the 5G network basestations For instance it analyzes the history records ofusersrsquo behaviors and extracts multimedia content informa-tion Then it generates the trusted content representation foreach content resource The representations of authenticationare the safety benchmarks for the multimedia contents Weapply the logical broadcasting channel to safeguard theserepresentationsThe term logicalmeans that it can be realizedby virtual technologies such as the network tunnels

Suppose the representations are safely transferred to theUser Required Authentication center (URA center) thenthe remaining task is to authentication the multimediacontents according to usersrsquo authentication requirementsTheterminal is thewireless devices and is responsible for commu-nicating with the center It can have an application installedwithin its platform and initiate the authentication requestto the URA center The center returns the authenticationinformation together with the multimedia content profile tothe terminal which are wrapped as information aggregationIf the center is not capable of providing relevant authentica-tion information it sends to the terminal a replay indicatingthat the terminal asks the TCA center through the wirelesscommunication channel for the authentication informationThe TCA center forwards the corresponding authenticationinformation to the UCA center which further sends it to theterminal Then the terminal checks the multimedia contentsrsquointegrity by comparing the representation with the contentitself

According to the proposed asymmetric channels for the5G networks there are two ways for the users or terminals toobtain and authenticate the multimedia content informationThe terminal connects to the wireless base station whichprovides the access point to the Internet The users upload ordownload contents through these points In order to ensurethe security of the terminal the access points apply authen-ticity test to the terminals The TCA center forms the TCR


index for the multimedia content Considering the long-tailfeatures of the popularity distribution of the content we onlygenerate the indexes for those ldquopopularrdquo contents which cansatisfy almost all usersrsquo authentication requirements [21 22]

The other channel for the user to obtain the authen-tication information is the logical broadcasting band Theapplication on the terminals asks theURA center for contentsauthentication information by specifying the contents profileThen the center searches its authentication informationdatabase for the specified terms If it finds the requiredone then it returns back to the terminal Note that inorder to protect the security of wireless channel betweenthe terminal and the center encryption methods are to beadopted We have researched the asymmetric channel inour previous work and its detail is referred to [23 24]The database which stores the authentication benchmarksreceives the authentication information from the TCA centerperiodically The security of this logical channel must beensured in order to provide the basis of trust for terminalauthentication test

3 Multimedia Trusted Content Representation

31 Trusted Content Representation Description We firstclarify what information the recipients can obtain from themultimedia contents for 5G networks As for the contentsrecipients they can get two kinds of information ie the highlevel semantics and the low semanticsThe formermeans thatthe descriptions about the contents while the latter meansthe perceptual understanding of the contentsWe define thesetwo terms in detail as follows

(1) Multimedia High Level Semantics (HLS) This kind ofsemantics is the description or explanation information aboutthe multimedia which is usually generated by the contentauthor and ismainly used for the indexing and searchingThesemantics are in the formof plaintext and can be added beforethe contents Examples ofHLS are title author keywords anddate Note that HLS belongs to the conceptual level regardingrecipients understanding of the contents

(2) Multimedia Low Level Semantics (LLS) The LLSdenotes the multimedia data which provides perceptualinformation for the recipients They often exist in the formof binary data eg the pixel matrix of an image and theframe series of a video Note that information of LLS needsno summary of conceptual levels by contents authors orpublishers The perceptual understandings are formed by thecontents recipients themselves

We believe that the authentication ofmultimedia contentsshould cover both HLS and LLS information Specificallythe HLS of contents should remain the same as the one oforiginal contents and the LLS should also be as intact asthe original one The undergoing operations on the LLSwhich are without malicious purposes should preserve theperceptual understanding as the original one Consideringthe 5G network packets loss to the data of LLS we propose toauthenticate the LLS robustly We believe the robust-hashingis more appropriate for 5G networks and choose the robust-hashing methods to generate the LLS of contents

Note that as for the HLS we suggest no robustness toits description One reason is that the HLS information isquite sensitive and one bit change may completely destroyits meaning The other reason is that we use the HLSto locate the multimedia contents and the HLS serves asthe authentication starting points for contents verificationTherefore we propose the TCRmethod which represents theHLS and LLS robustly The model is described as follows

We denote the multimedia contents space by M and themultimedia content by m Let the HLS of M be denoted byS and we separate the space S into several subspaces whichare mutually independent of each other ie the space S isthe cross product of its subspaces Each subspace representscertain conceptual description of M Then the HLS for m isdepicted by s which is defined as

s = (1199041 1199042 119904119899) (1)

The terms of s are the instances of the subspaces of SNote that the number n is an integer which represents thegranularity of the HLS spaces

Regarding the robustness representation of the LLS formultimedia contents we allow the perception preservingoperations on the contents on condition that these opera-tions do not change the representation too much Howeverthe degree of change is determined by the authenticationrequirements relating to the network circumstances In orderto clearly characterize the LLS feature we here examine thecategories of operational results of contentm

We adopt the robust-hashing authentication method forthe LLS representation For the content m the results ofoperations being conducted on m can be thought of havingthree groups iem119904m119888m119889 which are explained as follows

(1) The set m119904 denotes the multimedia contents whichare produced by the perception preserving operations onthe content m eg the scaling of image and the brightnesschange on the video frame

(2)The setm119888 denotes themultimedia contents which areproduced by the content changing operations on the contentm eg the frames reordering of video and the image blockscovering

(3) The set m119889 denotes the multimedia contents whichare different from and independent of contentm The LLS ofm119889 is completely and fundamentally distinguished fromm

We apply the robust-hashing methods to all those abovesets Let the hashing function be denoted by H(sdot) Thus weobtain the corresponding hashes as follows

H = H (m) (2)

H119904 = H (m119904) (3)

H119888 = H (m119888) (4)

H119889 = H (m119889) (5)

Obviously we have the relationH = HcupH119904 cupH119888 cupH119889 whereH denotes the LLS space for multimedia contents


s

H

High Level Semantics Low Level Semantics

c

d

Figure 2 Model of trusted content representation

We propose the trusted content representation methodto describe the HLS and LLS for content m Therefore thecontentm is represented by T which is defined as

T = (S H) (6)

The TCR model is depicted in Figure 2 in which the HLSspace is composed of S and S119889The S119889 is the HLS for contentswhich is different from and independent of contentm Notethat as for the LLSwe allow the robustness whichmeans thatthe hash of the content being verified againstm is secure if itfalls into the setHcupH119904 The parameter 120576 decides the distancebetween the original hash and the allowed maximum hashwhich also draws the line for the robustness of the LLS forcontentm

In regard to the authentication formultimedia contentmwe apply the TCR technique to verify the received contentWe first check the extracted HLS to see if it equals S If no weconclude that the received content is not safe since the highsemantics are not secure if yes we further check the extractedLLS to see whether it is within the range ofH119904 If yes we statethat the content being verified is authenticated if no we assertthat the low level semantics are being attacked and the contentreceived is not secure

32 Properties of Trusted Content Representation We in thispart summarize the properties of the model when it isemployed in the 5G networks to authenticate multimediacontentsThere are five properties ie collision-free securityrobustness sensitivity and compactness

(1) Collision-Free Property The collision-free property meansthat for every two dissimilar multimedia contents they havedifferent trusted content representations Suppose that m isinM and m119889 isin M and contents m and m119889 are different thenwe have

S = S119889 and 119901119903 (1003817100381710038171003817H minusH119889

1003817100381710038171003817 gt 120591) asymp 1 (7)

where symbol and means logic operation AND and 119901119903 meansthe probability The parameter 120591 is determined theoreticallyor empirically and it defines the robustness of the LLS for thecontents In other words the distance for every two different

multimedia contents is greater than the threshold 120591 withprobability one

(2) Security Property This property ensures that the trustedcontent representation for multimedia contents should beresilient against various malicious attacks and be detectableafter being attacked It also states that the representation Tfor certain content cannot be regenerated by attackers Usu-ally this is achieved through key controlled representationgeneration Specifically the LLS should be hashed by keysThe security property has two aspects ie one is the one-wayhash and the other is the unpredictability

The one-way hash means that it is easy to generate thehash for LLS from the contents while it is extremely difficultand impossible to recover the LLS of contents from the hashvalue Suppose the contents m has the representation (S H)and from T can be inferred the contentmrsquo then we have

119901119903 (H (m1015840) = H) asymp 2minus119871 asymp 0 (8)

where the integer L is the length of the hash and we assumethe one and zero in the hash follow uniform distribution

The unpredictability means that given a multimediacontent m it is empirically impossible to infer its hash forLLS Since the hash for LLS generation is controlled by keysuppose key 1198701 is used for the legal generating LLS hash form and key 1198702 is another key which is illegal then we have

119901119903 (H (m | 1198701) == H (m | 1198702)) asymp 2minus119871 asymp 0 (9)

where H(m | 1198701) means that the hash generation iscontrolled by K1 Note that K1 does not equal K2

(3) Robustness Property The robustness property of thismodel is embodied in the LLS for the contents ie someconception preserving operations on the LLS are permittedSuppose we have a contentm and its similar versionm119904 thenwe have the following relation

S == S119904 and1003817100381710038171003817H minusH119904

1003817100381710038171003817 le 120576 (10)

where threshold 120576draws the boundary between similarity anddifference for contents comparison


S

H

M

T

Channel

SemanticsManagement

HLS Analysis

LLS Analysis Trust content representations

LLS Analysis

HLS Analysis

Trust content representations

Safe ChannelT

EqualResults

T rsquo

S rsquo

H rsquo

MultimediaContent M0

Mrsquo

Figure 3 Framework of multimedia trust authentication based on trusted content representation

(4) Sensitivity Property This one is quite important forthe contents security verification It means that the modelcan detect those attacks which result in the changes oftrusted content representationThere are two kinds of attacksregarding the semantics of multimedia One is attacks on theHLS and the other is on the LLS The latter attacks harmrecipientsrsquo perceptual understanding on the LLS We denotethe altered contents of m by m119889119894119891119891 which belongs to the setm119888 capm119889 Then the following relation holds

S = S119889119894119891119891 or10038171003817100381710038171003817H minusH119889119894119891119891

10038171003817100381710038171003817 gt 120576 (11)

where the symbol⋁ denotes OR logic operation

(5) Compactness Property Compactness property means thatthe shorter the trusted content representation is the betterit is If the length of the representation is shorter it requiresless storage and less transmission overhead for 5G networksHowever we believe that there is a balance between the com-pactness and correctness of the representation For instanceif the contentsrsquo HLS representation S has much larger lengththen S can represent more detailed information It means theHLS space is divided into subspaces ofmore granularities Onthe contrary if the length is smaller the representation S ismuch more compact But more HLS information cannot beincluded Therefore it needs tactics and empirics to decidethe length of the trusted content representationNote that thisbalance also applies to the LLS representation

Although these five properties exist in the TCR modelwe claim that not all these five properties can be met atone time For example if the model is more emphasized onthe security of the multimedia contents then the securityand sensitivity properties should come first However if themodel is more concerned from the perspective of the trans-mission efficiency then the compactness property should beweighedmoreThus in real-life applications of 5G system theproperties of the model should be determined accordingly

4 General Multimedia TrustAuthentication Framework

When we apply the TCRmodel in the 5G networks we breakthis authentication into two processes ie the semanticsanalysis and adding by the sender the semantics extractionand comparison by the recipient which is shown in Figure 3

We denote the raw multimedia content generated bysender by1198720 First we analyze the content to obtain its HLSand LLS semantics which are represented by symbols S andH respectively The HLS is added to1198720 through semanticsmanagement The adding method can be realized by theappendence to the contentrsquos head or tail The multimedia1198720 together with S is conveyed to the recipient across thechannelThe channel is an open network ie the 5Gnetworkwhose security is not ensured On the other hand the HLSand LLS are synthesized into trusted content representationT which is securely transmitted by safe channel The safechannel can be achieved by PKI (Public Key Infrastructure)technique which encrypts the representation T through thecertificates provided by an authentication center

The safe channel is also used to test the recipientrsquos identityNote that the safe channel serves as the logical broadcastingchannel depicted in Figure 1 Before the recipient can accessany contents from the 5G networks its authenticity must beensured by the base stations Some modern identity methodeg trust computing can be applied to check the securityprofile of the terminalsrsquo platform Thus the base station firstinitiates the identity authentication requirement when therecipient asks for connecting to thewireless networks It startsthe challenges and responses mechanism to perform thecredit checking For instance the terminal hardware storescertain credit issued by the base station parties The basestation sends the nonce to the terminal Then the terminalreturns the cipher message of credit and nonce digitallyencrypted by the stationrsquos public key to the station Thebase station decrypts the cipher to compare the recipientrsquos


ltxml version=10encoding=utf-8gtltHighLevelSemanticsgt

lttitlegtThe Mystery of Van Goghs Earlttitlegt lt--title--gtltkeywordsgtVan GoghEarltkeywordsgt lt--keyword--gtltclassificationgtDocumentaryltclassificationgt lt--classification--gtltdescriptiongt

What happened on the December night in 1888 when Vincent vanGogh took a blade to his own ear

ltdescriptiongt lt--description--gtltsourcegtBBCltsourcegt lt--source--gtltlanguagegtEnglishltlanguagegt lt--language--gtltcoveragegt19th centuryltcoveragegt lt--coverage--gt

ltidentifiergthttpwwwbbccoukprogrammesb07nswft

ltidentifiergt lt--identifier--gtltcreatorgtBill Lockeltcreatorgt lt--creator--gtltpublishergtLion Televisionltpublishergt lt--publisher--gtltcontributorgtJack MacInnesltcontributorgt lt--contributor--gtltrightsgtcopyright reservedltrightsgt lt--rights--gtltdategt2016ltdategt lt--date--gtlttypegtVidVideolttypegt lt--type--gt

ltHighLevelSemanticsgt

Figure 4 High level semantics instance based on XML

credit with its stored one and decides whether the recipientis legal or not It also compares the decrypted nonce to itspreviously sent one If they match then it concludes thatthe message is new and free of replay attack Otherwise itbelieves the message is generated by attackers and rejects theaccess requirement We use the base station to actively verifythe authenticity of the recipients in order to prevent somemalicious users from deliberately sabotaging the security ofthe 5G networks

The recipient obtains from the channel the contentdenoted by Mrsquo whose integrity is to be verified by therecipient The recipient extracts from Mrsquo the HLS and LLSsemantics and form the trusted content representation TrsquoThen the recipient compares the representations Trsquo and Tand judges whether the content Mrsquo is secure or not If thosetwo terms are the same the recipient can be ensured thatthe content Mrsquo is of integrity Otherwise it is not secureand it may have been attacked Note that when the recipientcompares the LLS the robustness should be taken intoconsideration

Regarding the semantics management we propose ametadata basedmethod to append theHLS to themultimediacontentThemetadata is organized inXML format It presentsits semantics to the recipient before the content is presentedto the recipient For example the packets of supplementalenhancement information within the H264 stream can beused to store the HLS the label unit of ldquoAPPnrdquo within theJPEG file can be used for the HLS We provided a method todivide the HLS space into 14 subspaces based on our previouswork [25] The instance of XML based HLS representation isshown in Figure 4 in which fourteen concepts are used forcharacterizing the high semantics

We observe from the framework of Figure 3 that thereare both the active and passive authentication modes in ourmodel The HLS information authentication is active since

it adds or appends additional semantics information intothe multimedia contents However the LLS authentication ispassive because we extract the contentrsquos low level conceptualdescription akin to the digest The superiority of our frame-work is that it does not harm the content itself and the trustedcontent representation can be transferred separately fromthe content which can be protected by advanced encryptionalgorithms

5 Security Analysis of the Framework

Suppose that recipient obtains from the 5G networks themultimedia content lacking trust This could be caused byattackers modifying the semantics of the content We denotethe ldquooriginalrdquo content uploaded by an attacker to the networksby TCR Trsquo which generates the HLS and LLS as S1015840 and H1015840respectively The real and secure content relating to the Trsquo isdenoted by T which consists of HLS and LLS respectivelyThe purpose of the attacking is to distort the semantics of thecontent It can be realized from three perspectives The firstone is that the attacker modifies the high semantics of thecontent while keeping the low semantics intact The secondcase is opposite compared to the first attacking the lowsemantics of the content while keeping the high semanticsintact The third is that attacker modifies both the high andlow semantics

Note that in our proposed trust authentication frameworkthe safe channel is to transfer the authentication informationThus we assume this channel is free of attacks and thedestination of the channel can obtain the secure and intactinformation We analyze the security of the framework fromthree aspects according to the types of attacks to multimediaof the 5G networks

For the first scenario the high semantics is attacked onlyThus we have S1015840 = S and H1015840 = H The terminal obtains


the TCR from the URA center in Figure 1 as (S1015840 H) TheH could be valid or void because there is a chance that nocontent exists under the high semantics S1015840 If H is void theterminal can claim that the content it received is not trustedIf H is valid then it should have H1015840 = H according tothe collision-free property Thus terminal also can find theattacks Therefore the authentication is successful under thiskind of attack scenario

For the second type attack only the low semantics isaltered Thus it holds that S1015840 = S and H1015840 = H This can beeasily be detected since the terminal can obtain the TCR forthe received content as (S H) According to the robustnessand sensitivity properties the terminal can be alarmed thatH1015840 = H Note that the aim of this attack is to distort thelow semantics and thus the difference between H1015840 and His large enough to be detected Otherwise the H1015840 falls intothe robustness property of H and the low semantics of theattacked version could be almost the same as the original oneThis attack scenario can also be successfully detected

For the third kind of attack both the high and lowsemantics are changed We have S1015840 = S andH1015840 = H It meansthat the attacked multimedia content is completely differentfrom the original one By comparing the HLS of the contentthe URA center could return TCR (S1015840 H1015840) to the terminalif by chance the attacked content is the same as other legalcontents But this chance could be almost impossible sincewe suppose the attack is to distort the semantics and not tochange the content into another legal one Similar to the firstattack scenarioH could be voidThus the terminal can knowthat the content lacks trust

The above analysis shows that the general multimediatrust authentication framework can detect the attacks whichhave already happened It authenticates the contents in apassive way However it can also authentication the contentactively which means that it can prevent the attacks fromhappening Note that the base station first verifies the creditof the terminal before it can access the resources of the 5Gnetworks If certain attacker pretends legal user and asks forpermission to the network it lacks the credit and will failto be part of the wireless network However we state thatour framework cannot stop the legal terminal from harmingthe 5G network security For instance the legal user coulddownload the legal content deliberately alter the content andreupload the content However we believe the legal user orterminal could behave securely and legally and this scenariois beyond our analysis for the authentication framework

6 Conclusion

The secure and sustainable architecture of the 5G networksis vital for the networks health developments We propose ageneral multimedia content trust authentication frameworkto verify the various categories ofmultimediaThe frameworkis novel in that it adopts two channels to transmit theinformation ie one is for the usual multimedia and theother is for the authentication information The former oneis open and prone to attacks while the latter is closed andfree of attacks The framework adopts the trusted content

representation technique which authenticates the contents inboth high and low level semantics The high level semanticsare conceptual terms generated to locate the contents fromhumansrsquo understanding The low level semantics are robustand perceptual terms to measure the integrity of perceptionWe analyze the security of the framework and show that it canauthenticate the multimedia contents actively and passivelyIn our future work we will look at the implementation of thegeneral multimedia contents trust authentication by simulat-ing upon 5G systems Attention should also be focused on theattacks and security experiments on the proposed model

Data Availability

The data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

The authors declare that they have no conflicts of interest

Acknowledgments

This research is supported by National Natural Science Foun-dation of China (Grant no 61771185 Grant no 61772175)Doctoral Research Foundation of Southwest University ofScience and Technology (Grant no 17zx7158) Science andTechnology Research Project of Henan Province (Grant no182102210044 Grant no 182102210708) and Key ScientificResearch Program of Henan Higher Education (Grant no18A510009 Grant no 17A520005)

References

[1] A TzanakakiM Anastasopoulos I Berberana et al ldquoWireless-optical network convergence Enabling the 5G architecture tosupport operational and end-user servicesrdquo IEEE Communica-tions Magazine vol 55 no 10 pp 184ndash192 2017

[2] P K Agyapong M Iwamura D Staehle W Kiess and A Ben-jebbour ldquoDesign considerations for a 5G network architecturerdquoIEEE CommunicationsMagazine vol 52 no 11 pp 65ndash75 2014

[3] L Yan X Fang and Y Fang ldquoA Novel Network Architecturefor CU-Plane Staggered Handover in 5G Decoupled Het-erogeneous Railway Wireless Systemsrdquo IEEE Transactions onIntelligent Transportation Systems vol 18 no 12 pp 3350ndash33622017

[4] N Yang L Wang G Geraci M Elkashlan J Yuan and MDi Renzo ldquoSafeguarding 5G wireless communication networksusing physical layer securityrdquo IEEE Communications Magazinevol 53 no 4 pp 20ndash27 2015

[5] R Chaudhary N Kumar and S Zeadally ldquoNetwork ServiceChaining in Fog and Cloud Computing for the 5G Envi-ronment Data Management and Security Challengesrdquo IEEECommunications Magazine vol 55 no 11 pp 114ndash122 2017

[6] P Gandotra and R K Jha ldquoA survey on green communicationand security challenges in 5G wireless communication net-worksrdquo Journal of Network and Computer Applications vol 96pp 39ndash61 2017


[7] L Xing Q Ma and L Jiang ldquoMicroblog user recommendationbased on particle swarm optimizationrdquoChina Communicationsvol 14 no 5 pp 134ndash144 2017

[8] Q Ma L Xing and L Zheng ldquoAuthentication of ScalableVideo Coding Streams Based on Topological Sort on DecodingDependency Graphrdquo IEEE Access vol 5 pp 16847ndash16857 2017

[9] L Xing Z Zhang H Lin and F Gao ldquoContent Centric Net-work with Label Aided User Modeling and Cellular PartitionrdquoIEEE Access vol 5 pp 12576ndash12583 2017

[10] M Asikuzzaman M J Alam A J Lambert and M RPickering ldquoImperceptible and robust blind video watermarkingusing chrominance embedding a set of approaches in the DTCWT domainrdquo IEEE Transactions on Information Forensics andSecurity vol 9 no 9 pp 1502ndash1517 2014

[11] I Setyawan and I K Timotius ldquoContent-dependent spatio-Temporal video watermarking using 3-dimensional discretecosine transformrdquo in Proceedings of the 2013 5th InternationalConference on Information Technology and Electrical Engineer-ing ICITEE 2013 pp 79ndash83 Yogyakarta October 2013

[12] B Sridhar and C Arun ldquoAn enhanced approach in videowatermarking withmultiple watermarks using waveletrdquo Journalof Communications Technology and Electronics vol 61 no 2 pp165ndash175 2016

[13] Z Shahid and W Puech ldquoVisual protection of HEVC video byselective encryption of CABAC binstringsrdquo IEEE Transactionson Multimedia vol 16 no 1 pp 24ndash36 2014

[14] G Van Wallendael A Boho J De Cock A Munteanu and RVanDeWalle ldquoEncryption for high efficiency video codingwithvideo adaptation capabilitiesrdquo in Proceedings of the 2013 IEEEInternational Conference on Consumer Electronics ICCE 2013pp 31-32 Las Vegas USA January 2013

[15] X Kang and Y Wu ldquoA trust-based pollution attack preventionscheme in peer-to-peer streaming networksrdquo Computer Net-works vol 72 pp 62ndash73 2014

[16] Y Lu W Wang B Bhargava and D Xu ldquoTrust-based privacypreservation for peer-to-peer data sharingrdquo IEEE Transactionson Systems Man and Cybernetics Systems vol 36 no 3 pp498ndash502 2006

[17] C Cheng T Jiang andQ Zhang ldquoTESLA-based homomorphicMAC for authentication in P2P system for live streaming withnetwork codingrdquo IEEE Journal on Selected Areas in Communi-cations vol 31 pp 291ndash298 2013

[18] A Habib D Xu M Atallah B Bhargava and J Chuang ldquoAtree-based forward digest protocol to verify data integrity indistributed media streamingrdquo IEEE Transactions on Knowledgeand Data Engineering vol 17 no 7 pp 1010ndash1013 2005

[19] K Mokhtarian and M Hefeeda ldquoAuthentication of scalablevideo streams with low communication overheadrdquo IEEE Trans-actions on Multimedia vol 12 no 7 pp 730ndash742 2010

[20] L N Vadlamudi R P V Vaddella and V Devara ldquoRobust hashgeneration technique for content-based image authenticationusing histogramrdquoMultimedia Tools andApplications vol 75 no11 pp 6585ndash6604 2016

[21] Y-J Park ldquoThe adaptive clustering method for the longtail problem of recommender systemsrdquo IEEE Transactions onKnowledge and Data Engineering vol 25 no 8 pp 1904ndash19152013

[22] M OrsquoMahony and M Bray ldquoThe long tail of contentrdquo Commu-nications Engineer vol 4 no 4 pp 20ndash25 2006

[23] L Xing J Ma X-H Sun and Y Li ldquoDual-mode transmissionnetworks for DTVrdquo IEEE Transactions on Consumer Electronicsvol 54 no 2 pp 474ndash480 2008

[24] L Xing L Jiang G Yang and B Wen ldquoA novel trustedcomputing model for network security authenticationrdquo Journalof Networks vol 9 no 2 pp 339ndash343 2014

[25] L Xing Q Ma and M Zhu ldquoTensor semantic model for anaudio classification systemrdquo Science China Information Sciencesvol 56 no 6 pp 1ndash9 2013

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018


Active and Passive Electronic Components

VLSI Design



Shock and Vibration


Civil EngineeringAdvances in

Acoustics and VibrationAdvances in



Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of



Journal ofEngineeringVolume 2018

SensorsJournal of



RotatingMachinery


Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018


Chemical EngineeringInternational Journal of Antennas and

Propagation




Navigation and Observation


Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom


provide the authentication of contents in some degree theembedding process itself harms the integrity of the content Inaddition one watermarking method can only deal with somespecific attacks which hinders these methods from beingwidely adopted to cope with the various attacks existing inthe 5G networks

The encryption-based multimedia authentication meth-ods usually employ the encryption algorithms eg RSAalgorithm and DES algorithm to encrypt the whole or partsof the contents For example Zafar Shahid et al [13] presenteda selective encryption (ie the truncated rice code and theExp-Golomb code) approach for video of High EfficiencyVideo Coding (HEVC) which shows the characteristics offormat compliant and real-time property Similarly GlennVanWallendael et al [14] proposed the encryption for HEVCby utilizing the differences of intraprediction mode the signof motion vector difference and the residual sign Note thatthe primary goal of the encryption-based methods is toensure the confidentialityThe time overhead incurred by theencryption and decryption process is often a nonnegligibleissue for the social media devices

The aim of the streaming-based multimedia authentica-tion methods is to ensure that the received contents havethe integrity property which is verified by the recipientsCurrently many research works have been focused on thisarea eg Kang et al [15] studied the pollution attackdetection and prevention method by trust for peer-to-peerstreaming Lu et al [16] proposed a privacy protectionmethod based on trust for peer-to-peer data sharing networkand Cheng et al [17] authenticated the live streaming mediaby means of TESLA- (Timed Efficient Stream Loss-tolerantAuthentication-) based protocol One category of streaming-based authentication methods is the graph structure typeeg the chain line approach the tree approach and thebutterfly approach which treats the content packets asindividual ones and exploits various packetsrsquo hash attachingmeans [18] The other category is based on the multimediacoding structure For example Kianoosh Mokhtarian et al[19] proposed the authentication for Scalable Video Coding(SVC) streams whose hash appendence follows the codingstructure of SVC

The robust-hashing-based methods to authenticate themultimedia have the best performance in terms of robustnesswhen the authentication is applied under the network cir-cumstances The robustness of the hashes means that hashesremain the same or change a little after the perceptual contentpreserving operations are being conducted For exampleLokanadham Naidu Vadlamudi et al [20] proposed a robusthash algorithm by using features of histogram for imageauthentication Due to the robustness and sensitivity of thiskind of method it is much superior to the other three meth-ods when it comes to the network packet loss phenomenonin 5G networks Since certain packets of multimedia contentsmay be lost because of traffic jams or network failure contentsintegrity verifying process for the recipients should take intoconsideration the robustness of multimediarsquos representation

Another aspect we should bear in mind is that all thefour methods are for the low level semantics authenticationwhich treats only the integrity either in pixels level or

in frames level This absolutely lacks completeness for theauthentication of multimedia contents of the 5G networkssince high level semantics of the contents are overlooked bythe current methods For instance the contentsrsquo high levelsemantics including but not limited to title name authorand format should also be authenticated Therefore in thispaperwe endeavor to tackle themultimedia authentication byproposing combined authentication method for multimediacontents for 5G networkWe propose a generalized multime-dia trust authentication model based on the trusted contentrepresentation (TCR) method where the generalized termmeans that it can be applied to various multimedia contentformats Also we provide the features needed to authenticatethe integrity of the contents

The rest of the paper is organized as follows We describethe architecture of asymmetric wireless communicationchannel in Section 2 where the asymmetric term means thatone channel is safe and is used for authentication informationtransfer while the other one is open and used for multimediacontents transfer In Section 3 we illustrate the TCR indexingtechnique and show the features of TCR In Section 4we explain the framework of the general multimedia trustauthentication for 5G networks in which the philosophy ofasymmetric wireless channel is adopted Then we concludeour work in Section 5

2 Architecture of Asymmetric WirelessCommunication Channel

21 Security Threats to 5G Networks Since the 5G networksare open and insecure the multimedia contents whichare transmitted over them are prone to various attacksby malicious purposes From the perspectives of networkcommunication system we summarize security threats tomultimedia into three groups

(1) The Originality of the Multimedia May Lack Trust Cur-rently users can upload from the mobile devices the variouskinds of multimedia contents up to the video or audiosharing websites by various means However the effectiveand efficient mechanisms to supervise the legacy of usersrsquobehaviors are still unavailable While people find it quiteeasy to collect edit and distribute the multimedia contentsby modern multimedia processing tools (eg PhotoshopIllustrator) people should think twice about the originalityof the received multimedia In other words the user whopublishes the multimedia and claims the author of themultimediamay likely not be the ldquotruerdquo authorOriginality ofthe contents should be checked in some manner This incursthe problem of intellectual property infringementThereforeit is urgent to find a good way to verify the originality of themultimedia

(2) The Channel to Convey the Multimedia May Be AttackedThe openness of the 5G network protocols poses serioussecurity problems to the contents being transmitted overit The notorious attack ldquoman-in-the-middlerdquo is the mostcommon threat to the contents The attacker maliciously






Terminal



MultimediaContent

Information




















s = (1199041 1199042 119904119899) (1)








H = H (m) (2)

H119904 = H (m119904) (3)

H119888 = H (m119888) (4)

H119889 = H (m119889) (5)



s

H


c

d



T = (S H) (6)





S = S119889 and 119901119903 (1003817100381710038171003817H minusH119889

1003817100381710038171003817 gt 120591) asymp 1 (7)











S == S119904 and1003817100381710038171003817H minusH119904

1003817100381710038171003817 le 120576 (10)



S

H

M

T

Channel

SemanticsManagement

HLS Analysis


LLS Analysis

HLS Analysis


Safe ChannelT

EqualResults

T rsquo

S rsquo

H rsquo


Mrsquo



S = S119889119894119891119891 or10038171003817100381710038171003817H minusH119889119894119891119891

10038171003817100381710038171003817 gt 120576 (11)































6 Conclusion



Data Availability




Acknowledgments


References





























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia







Terminal



MultimediaContent

Information




















s = (1199041 1199042 119904119899) (1)








H = H (m) (2)

H119904 = H (m119904) (3)

H119888 = H (m119888) (4)

H119889 = H (m119889) (5)



s

H


c

d



T = (S H) (6)





S = S119889 and 119901119903 (1003817100381710038171003817H minusH119889

1003817100381710038171003817 gt 120591) asymp 1 (7)











S == S119904 and1003817100381710038171003817H minusH119904

1003817100381710038171003817 le 120576 (10)



S

H

M

T

Channel

SemanticsManagement

HLS Analysis


LLS Analysis

HLS Analysis


Safe ChannelT

EqualResults

T rsquo

S rsquo

H rsquo


Mrsquo



S = S119889119894119891119891 or10038171003817100381710038171003817H minusH119889119894119891119891

10038171003817100381710038171003817 gt 120576 (11)































6 Conclusion



Data Availability




Acknowledgments


References





























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia












s = (1199041 1199042 119904119899) (1)








H = H (m) (2)

H119904 = H (m119904) (3)

H119888 = H (m119888) (4)

H119889 = H (m119889) (5)



s

H


c

d



T = (S H) (6)





S = S119889 and 119901119903 (1003817100381710038171003817H minusH119889

1003817100381710038171003817 gt 120591) asymp 1 (7)











S == S119904 and1003817100381710038171003817H minusH119904

1003817100381710038171003817 le 120576 (10)



S

H

M

T

Channel

SemanticsManagement

HLS Analysis


LLS Analysis

HLS Analysis


Safe ChannelT

EqualResults

T rsquo

S rsquo

H rsquo


Mrsquo



S = S119889119894119891119891 or10038171003817100381710038171003817H minusH119889119894119891119891

10038171003817100381710038171003817 gt 120576 (11)































6 Conclusion



Data Availability




Acknowledgments


References





























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia



s

H


c

d



T = (S H) (6)





S = S119889 and 119901119903 (1003817100381710038171003817H minusH119889

1003817100381710038171003817 gt 120591) asymp 1 (7)











S == S119904 and1003817100381710038171003817H minusH119904

1003817100381710038171003817 le 120576 (10)



S

H

M

T

Channel

SemanticsManagement

HLS Analysis


LLS Analysis

HLS Analysis


Safe ChannelT

EqualResults

T rsquo

S rsquo

H rsquo


Mrsquo



S = S119889119894119891119891 or10038171003817100381710038171003817H minusH119889119894119891119891

10038171003817100381710038171003817 gt 120576 (11)































6 Conclusion



Data Availability




Acknowledgments


References





























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia



S

H

M

T

Channel

SemanticsManagement

HLS Analysis


LLS Analysis

HLS Analysis


Safe ChannelT

EqualResults

T rsquo

S rsquo

H rsquo


Mrsquo



S = S119889119894119891119891 or10038171003817100381710038171003817H minusH119889119894119891119891

10038171003817100381710038171003817 gt 120576 (11)































6 Conclusion



Data Availability




Acknowledgments


References





























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia

























6 Conclusion



Data Availability




Acknowledgments


References





























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia
























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia


general multimedia trust authentication framework...

Documents