CoChat
Sign Up
Cited by 12Video Coding and Compression TechnologiesChaos-based Image/Signal EncryptionDigital Rights Management and Security

Effectiveness of crypto-transcoding for H.264/AVC and HEVC video bit-streams

Open access
Shah, Rizwan A., Asghar, Mamoona N., Abdullah, Saima, Fleury, Martin, Gohar, Neelam·2019·Multimedia Tools and Applications
DOI 10.1007/s11042-019-7451-5
Open this paper in CoChat — free accountOpen PDF

Abstract

To avoid delays arising from a need to decrypt a video prior to transcoding and then re-encrypt it afterwards, this paper assesses a selective encryption (SE) content protection scheme. The scheme is suited to both recent standardized codecs, namely H.264/Advanced Video Coding (AVC) and High Efficiency Video Coding (HEVC). Specifically, the paper outlines a joint crypto-transcoding scheme for secure transrating of a video bitstream. That is to say it generates new video bitrates, possibly as part of an HTTP Adaptive Streaming (HAS) content delivery network. The scheme will reduce the bitrate to one or more lower desired bit-rate without consuming time in the encryption/decryption process, which would be the case when full encryption is used. In addition, the decryption key no longer needs to be exposed at intermediate middleboxes, including when transrating is performed in a cloud datacenter. The effectiveness of the scheme is variously evaluated: by examination of the SE generated visual distortion; by the extent of computational and bitrate overheads; and by choice of cipher when encrypting the selected elements within the bitstream. Results indicate that there remains: a content; quantization level (after transrating of an encrypted video); and codec-type dependency to any distortion introduced. A further recommendation is that the Advanced Encryption Standard (AES) is preferred for SE to lightweight XOR encryption, despite it being taken up elsewhere as a real-time encryption method.

AI snapshot

Written from the paper’s full text.

TL;DR

This paper evaluates a decoder-compatible selective encryption (SE) scheme for joint crypto-transcoding/transrating of H.264/AVC and HEVC bitstreams, encrypting CABAC sign elements so transcoding can occur without decryption or re-encryption. It finds content-, QP-, and codec-dependent distortion and bitrate/computation trade-offs, with AES-CFB preferred over XOR for SE.

Key findings

  • The proposed SE encrypts only selected CABAC syntax elements—signs of transform coefficients (TCs) and motion vector differences (MVDs)—allowing format-compatible transcoding without exposing decryption keys at intermediate transcoders.
  • HEVC crypto-transcoding reduced bitrate versus H.264/AVC at all tested QPs: for Stefan, 1670 kb to 1130 kb at QP 12 (−32.34%) and 159 kb to 29 kb at QP 48 (−81.76%); for Mobile, 1755 kb to 1213 kb at QP 12 (−30.88%) and 178 kb to 55 kb at QP 48 (−80.29%).
  • HEVC produced less PSNR_Y distortion than H.264/AVC after SE and transrating, e.g. Stefan QP 12: 10.01 dB vs 16.19 dB (+61.74%); Mobile QP 12: 6.7 dB vs 16.5 dB (+146.27%), but HEVC incurred greater encoding/transcoding time, e.g. +81.79% for Stefan QP 12 and +113% for Mobile QP 12.
  • AES-CFB encryption produced strongly distorted frames and increased histogram entropy/randomness, whereas XOR encryption resulted in insufficient distortion; the paper recommends AES over lightweight XOR for SE.

Methods

  • Implemented a closed-loop crypto-transcoder using H.264/AVC reference software JM 18.6 and HEVC reference software HM-15.0 on an Intel Core i3 Core 2 Duo (2.10 GHz), 6 GB RAM, Windows 8.1 Professional.
  • Tested two CIF sequences, Stefan and Mobile (352 × 288 pixels/frame), and one HD 720 sequence, Four People (1280 × 720 pixels/frame), encoded with I, P, and B frames and GOP of 16; videos were first compressed at QP = 0, selectively encrypted, then transrated to QPs 12, 24, 36, and 48.
  • Evaluated visual distortion with PSNR and SSIM, plus Laplacian edge detection and pixelation; also measured file sizes, bitrate, encoding time, and compared AES-CFB versus XOR using visual results and RGB/luminance histograms.

Limitations

  • Stated: the focus is on encryption overhead rather than transcoder architectures, and alternative transcoder architectures are left for future research; the intention was not to compare the relative compression performance of H.264/AVC and HEVC.
  • Stated: VBR is assumed for simplicity in quality comparisons; future work will consider adaptive syntax-element selection, correlation-coefficient analysis, video encryption quantity, and rate-distortion analysis such as Bjøntegaard-Delta.
  • Evident from the text: experiments cover only two CIF sequences and one HD 720 sequence on a single machine, with no UHD testing despite HEVC’s UHD orientation, and no subjective QoE assessment is reported.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1007/s11042-019-7451-5

Cites

76 of 76 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

7 of 47 cited sources checked · 2 supported · 0 flagged · accuracy not yet measured

  • Premisecitation
    § 1 Introduction · p. 2

    Full encryption (naïve encryption) may result in significant performance overhead, especially if software-only encryption occurs.

    “Consequently, it may result in a significant performance overhead, especially if software-only encryption occurs [7].”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [7] Boho A et al (2013) End-to-end security for video distribution: the combination of encryption, watermarking, and video adaptation. IEEE Signal Process Mag 30(2):97–107 source

      “Oceans of choices for video scrambling” identifies encrypting the whole encoded bitstream as the naïve approach. But the passages do not say that this approach causes significant performance overhead or that software-only encryption makes such overhead worse; the overhead discussed in the Abstract concerns the proposed system, not that comparison. No page numbers are available in our copy. (LLM, conf=0.75)

      Experimental — accuracy not yet measured
  • Premisecitation
    § 1 Introduction · p. 2

    Some forms of selective encryption are decoder format-compatible.

    “but some forms of SE are decoder format-compatible.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [8] Boho A, Van Wallendael G, Dooms A, De Cock J, Schelkens P, Prenel B, Van de Wall R (2013) End-to-end security for video distribution. IEEE Sig Process Mag 30(2):97–107 source
  • Interpretationnone given
    § 2.4 Standardized Video Codecs · p. 6

    The proposed scheme is suitable for both H.264/AVC and HEVC video codec standards.

    “As mentioned in Section 1, the scheme is suitable for both current video codec standards, the H.264/AVC standard [20] and the HEVC codec [21]”
    Quote in the paperSubstantial claim2 of 2 sources checked
    • cites [20] Fan Q, Yin H, Min G, Yang P, Luo Y, Lyu Y, Huang H, Jiao L (2018) Video delivery networks: challenges, solutions and future directions. Comput Elec Eng 66:332–341 source

      The Introduction and Section 3.2 discuss video streaming and bitrate adaptation, but do not address whether a proposed scheme works with H.264/AVC or HEVC. Searches of our copy for H.264, HEVC, and codec compatibility found no passage supporting the claim. (LLM, conf=0.77)

      Experimental — accuracy not yet measured
    • cites [21] Fang J, Chen Z, Liao T, Chang P (2014) A fast PU mode decision algorithm for H.264/AVC to HEVC transcoding. Proc 4th Int Conf Computer Sci Info Technol: 215–225 source

      The Introduction (p. 1) describes a transcoder that takes H.264/AVC input and reuses its parameters to encode HEVC output, so the paper supports a method involving both standards. It does not establish that the citing document’s proposed scheme is suitable for both codecs generally. (LLM, conf=0.82)

      Experimental — accuracy not yet measured
  • Findingnone given
    § 5.3 Comparative analysis · p. 18

    When H.264/AVC is targeted, the computational overhead in terms of encryption with transcoding time appears to be lower than the schemes proposed in [15], [54], and [58].

    “Meanwhile, when H.264/AVC is targeted, the computational overhead in terms of encryption with transcoding time appears to be lower than the schemes proposed in [15], [54], and [58].”
    Quote in the paperSubstantial claim1 of 3 sources checked
    • cites [15] Correa G, Assunção P, Volcan Agostini L, da Silva Cruz L (2015) Fast HEVC encoding decisions using data mining. IEEE Trans Circ Syst Video Technol 25(4):660–673 source
    • cites [54] Rothke B (2007) A look at the advanced encryption standard (AES), 6th edition. Info Sec Manag Handbook: 1151–1158, CRC Press, Boca Raton, FL source
    • cites [58] Seufert M, Egger S, Slanina H, Zinner T, Hoβfeld T, Tran-Gia P (2015) A survey on quality of experience of HTTP adaptive streaming. IEEE Commun Surv Tutor 17(1):469–492 source

      Section 3.3 discusses H.264/AVC and the file-size overhead of scalable video coding, but it does not compare encryption or transcoding times. Searches for encryption, transcoding, and computational overhead found no passage supporting the claimed comparison with schemes [15], [54], and [58]. (LLM, conf=0.78)

      Experimental — accuracy not yet measured
  • Findingown data
    § Effectiveness of Crypto-Transcoding for H.264/AVC and HEVC Video Bit-streams · p. 1

    Visual distortion introduced by the SE/transrating scheme depends on quantization level after transrating of an encrypted video.

    “Results indicate that there remains: a content; quantization level (after transrating of an encrypted video); and codectype dependency to any distortion introduced.”
    Quote in the paperSubstantial claim
  • Classified as probably not a substantial claim. 37
  • May combine two claims. 29

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation
    § 1 Introduction · p. 2

    Selective encryption reduces the amount of data encrypted compared to full encryption.

    “compared to full encryption, with selective encryption (SE) [8] not only is the amount of data encrypted reduced”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [8] Boho A, Van Wallendael G, Dooms A, De Cock J, Schelkens P, Prenel B, Van de Wall R (2013) End-to-end security for video distribution. IEEE Sig Process Mag 30(2):97–107 source

  • May combine two claims.

    Interpretationreasoning
    § 5.3 Comparative analysis · p. 18

    The proposed crypto-transcoder has advantages over previously proposed schemes in [15], [54] and [58], though to some extent these advantages are more to do with the development of the HEVC codec than merits of prior schemes.

    “The crypto-transcoder has advantages over previously proposed schemes in [15], [54] and [58], though to some extent these are more to do with the development of the HEVC codec than merits of prior schemes.”
    Quote in the paperSubstantial claim1 of 3 sources checked
    • cites [15] Correa G, Assunção P, Volcan Agostini L, da Silva Cruz L (2015) Fast HEVC encoding decisions using data mining. IEEE Trans Circ Syst Video Technol 25(4):660–673 source
    • cites [54] Rothke B (2007) A look at the advanced encryption standard (AES), 6th edition. Info Sec Manag Handbook: 1151–1158, CRC Press, Boca Raton, FL source
    • cites [58] Seufert M, Egger S, Slanina H, Zinner T, Hoβfeld T, Tran-Gia P (2015) A survey on quality of experience of HTTP adaptive streaming. IEEE Commun Surv Tutor 17(1):469–492 source

      Section 3.3 (Server-side Actions; page unavailable) describes HEVC as an efficient codec, but does not describe a crypto-transcoding scheme or compare one with the proposed crypto-transcoder. Searches for crypto-transcoding and encrypted-video transcoding found no passage addressing that comparison. (LLM, conf=0.76)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Findingown data
    § 5.1 Crypto-transcoding with AVC and HEVC · p. 15

    Comparing Stefan and Mobile, a similar pattern of results is evident, but there are greater differences in quality between the codecs for Mobile, which has greater spatial complexity than Stefan.

    “Comparing between Stefan and Mobile video sequences, a similar pattern of results is evident. However, there are greater differences in quality between the codecs for Mobile , with greater spatial complexity than Stefan .”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data
    § 6 Conclusions · p. 19

    From experiments, the gains from using SE are substantial in terms of limited bitrate overhead and several seconds saved in encoding alone, even for short video sequences.

    “From experiments, the gains from using SE are substantial in terms of limited bitrate overhead and (according to the machine employed) several seconds saved in encoding alone, even for short video sequences.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The machine employed for the experiments is representative of typical transcoding hardware.

  • May combine two claims.

    Findingown data
    § 5.3 Comparative analysis · p. 19

    The proposed crypto-transcoder scheme has low computational overhead for H.264/AVC and medium for HEVC due to video encoding time.

    “Low for H.264/AVC, medium for HEVC due to video encoding time”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data
    § 6 Conclusions · p. 19

    From comparisons of visual distortion in the paper, there is a content dependency and QP-dependency when transrating to lower bitrates, in the sense that SE may contribute reduced distortion at higher bitrates.

    “From comparisons of visual distortion in the paper, it is apparent that there is a content dependency and QP-dependency when transrating to lower bitrates, in the sense that SE may contribute reduced distortion at higher bitrates.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data
    § 5.2 HEVC encoded Crypto-transrating with XOR and AES ciphers · p. 16

    In crypto-transrating of the HD720 Four People video with HEVC, encryption using AES-CFB produces a strongly distorted frame, whereas applying XOR results in insufficient distortion.

    “The visual results (from frame no. 53) indicate that the encryption performed using AES-CFB produces a strongly distorted frame, whereas applying an XOR operation results in insufficient distortion.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Interpretationreasoning
    § 1 Introduction · p. 3

    In the proposed scheme, the decryption procedure is always performed by a target device and not at any intermediate point where content and/or decryption keys may be exposed.

    “Finally, the decryption procedure is always performed by a target device and not at any intermediate point, where the content and/or decryption keys may be exposed.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning
    § Effectiveness of Crypto-Transcoding for H.264/AVC and HEVC Video Bit-streams · p. 1

    The proposed crypto-transcoding scheme reduces the bitrate to one or more lower desired bitrates without consuming time in encryption/decryption, unlike full encryption.

    “The scheme will reduce the bitrate to one or more lower desired bitrate without consuming time in the encryption/decryption process, which would be the case when full encryption is used.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationreasoning
    § 1 Introduction · p. 1

    Full encryption causes additional latency at intermediate transcoders due to the need for decryption.

    “Unfortunately, content protection through encryption results in additional latency at intermediate transcoders due to the need for decryption.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation
    § 1 Introduction · p. 3

    Closed-loop designs minimize drift but re-introduce extra processing latency.

    “Closed-loop designs seek to approximate the cascaded architecture of Fig. 1 by reintroducing the transform stage; however, though drift is minimized, extra processing latency is re-introduced.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [17] Dorwin D, Smith J, Watson M, Bateman A (2017) Encrypted media extensions W3C recommendation. W3C

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationnone given
    § 1 Introduction · p. 3

    Choosing SE rather than full encryption risks that content may not be sufficiently distorted and may expose video content without decryption.

    “Choosing SE rather than full encryption runs a risk that content may not be sufficiently distorted and, so in some way, expose the video content any way without the need for decryption.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation
    § 1 Introduction · p. 3

    In open-loop transcoder designs, temporal error drift builds up across a GoP because no reference is made to the frame store before re-quantization or transform coefficient pruning, even though processing time is reduced.

    “In open-loop transcoder designs, because no reference is made to the frame store before re-quantization or transform coefficient pruning, temporal error drift builds up across a Group-of-Pictures (GoP), even though processing time is reduced.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [17] Dorwin D, Smith J, Watson M, Bateman A (2017) Encrypted media extensions W3C recommendation. W3C

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationnone given
    § 1 Introduction · p. 2

    In RTMP delivery to a CDNaaS with SE, there is no intermediate storage of the decryption key, which is only held by the client device.

    “For example, in the RTMP delivery of the original video to a CDNaaS, were SE to be employed, there would be no intermediate storage of the decryption key, which is only held by the client device.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationreasoning
    § 1 Introduction · p. 3

    SE is applied at the entropy-coding stage of the encoder as a crypto-compression scheme, so there is limited additional computational overhead from encryption.

    “SE is applied at the entropy-coding stage of the encoder as a crypto-compression scheme, so that there is limited additional computational overhead from the encryption process.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning
    § 1 Introduction · p. 3

    The proposed crypto-transcoder avoids temporal error drift and reduces the time previously taken by encryption and decryption.

    “Instead, the proposed cryptotranscoder, while avoiding temporal error drift, reduces the time previously taken-up by encryption and decryption.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given
    § 1 Introduction · p. 2

    Selectively encrypted components can be encrypted by a standardized cipher such as AES in a streaming mode such as CFB.

    “In addition, the selectively encrypted components can be encrypted by a standardized cipher such as AES operating in a streaming mode such as Cipher Feedback (CFB).”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Interpretationnone given
    § 1 Introduction · p. 3

    The proposed scheme employs a selective-encryption method that works only on uniformly-distributed selected syntax elements of the compressed video stream.

    “The scheme presented in this paper employs a selective-encryption method that works only on uniformly-distributed selected syntax elements of the compressed video stream.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning
    § 1 Introduction · p. 2

    It is possible to ensure video is unwatchable owing to distortions while permitting bitrate transcoding without decryption.

    “Thus, it is possible to ensure the video is unwatchable owing to distortions, while at the same time permit bitrate transcoding without decryption.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given
    § 1 Introduction · p. 2

    If transcoders or intermediate devices need to decrypt the video stream, the decryption key is exposed.

    “Should these transcoders or other intermediate devices need to decrypt the video stream then the decryption key is exposed.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given
    § 1 Introduction · p. 3

    In the proposed scheme, the process of changing the QP works exactly as before.

    “The process of changing the QP otherwise works exactly as before.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation
    § 1 Introduction · p. 2

    RC4 is considered cryptographically insecure by the IETF.

    “As RC4 is now considered cryptographically insecure by the Internet Engineering Task Force (IETF) [6]”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [6] Bjork N, Christopoulos C (1998) Transcoder architectures for video coding. IEEE Trans Consumer Electron 44(1):88–98 source

  • Classified as probably not a substantial claim.

    Interpretationreasoning
    § 1 Introduction · p. 3

    In the proposed scheme, all processing takes place within the same transcoder block.

    “As a result, all processing takes place within the same transcoder block.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given
    § 3 Related Work · p. 8

    SE as developed in the current paper supports intermediate processing, such as the transrating analysed herein.

    “As developed in the current paper, it also supports intermediate processing, such as the transrating analysed herein.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationreasoning
    § 2.2 Ciphers: Advanced Encryption Standard (AES) and real-time XOR · p. 5

    As AES employs a single key with a limited key length, efficiency is increased in computational time and memory consumption compared to asymmetric-key algorithms.

    “As AES employs a single key with a limited key length, the efficiency is increased in terms of computational time, and memory consumption compared to asymmetric-key algorithms.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationnone given
    § 3 Related Work · p. 8

    SE as presented in the current paper remains an effective lightweight encryption method, with advantages of lightweight chaos-based encryption.

    “All the same, SE, as presented in the current paper, remains an effective lightweight encryption method, with advantages of lightweight chaos-based encryption.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data
    § 5.1 Crypto-transcoding with AVC and HEVC · p. 15

    The change in distortion for HEVC from low QP to high QP is the combined effect of lower quality video at the higher QPs and the addition of SE.

    “Distortion increases for HEVC in going from low QP (high quality) to high QP (low quality), though this change is the combined effect of lower quality video at the higher QPs and the addition of SE.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductreasoning
    § 5.1 Crypto-transcoding with AVC and HEVC · p. 15

    Tables 3 and 4's PSNR results are averaged across each of the two sequences compared, and thus subjective impressions from Figs. 6 and 7 may not be confirmed for individual sample frames.

    “Recall that Tables 3 and 4's PSNR results are averaged across each of the two sequences compared and, thus, subjective impressions from Figs. 6 and 7 may not be confirmed for individual sample frames.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning
    § 5.1 Crypto-transcoding with AVC and HEVC · p. 12

    The visibility of static portions is because the method of SE works on changing values of MV signs; if there is little motion, the MVD will be small and the impact of encrypting the signs will be small.

    “This is not surprising as the calendar and tennis court are largely static, whereas the method of SE works on changing values of the MV signs. If there is little motion the MVD will be small and the impact of encrypting the signs will be small.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Study design and conductnone given
    § 4 Methodology · p. 9

    Transcoding by means of the closed-loop architecture removes error drift from the video stream.

    “To remove error drift from the video stream, transcoding was by means of the closed-loop architecture of Table 1.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation
    § 2.4 Standardized Video Codecs · p. 6

    HEVC can increase compression ratio by up to 50% compared to H.264/AVC.

    “Though, the compression ratio can be increased by up to 50% by means of an HEVC codec that advantage comes with an increase in codec complexity, which in turn increases processing latency.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [44] Muhammad K, Hamza R, Ahmad J, Lloret J, Wang H, Baik SW (2018) Secure surveillance network for IoT systems using probabilistic image encryption. IEEE Trans Indust Inform 14(8):3679–3689 source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation
    § 2.3 Transcoding · p. 6

    There may be a bitrate overhead arising from scalable coding, though HEVC variants of SVC have gone some way to address that issue by considering inter-layer prediction.

    “However, there may be a bitrate overhead arising from scalable coding, though HEVC variants of SVC [40] have gone some way to address that issue by considering inter-layer prediction.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [40] Massoudi AA, Lefebvre F, De Vleeschouwer C, Macq B, Quisquater J-J (2008) Overview on selective encryption of image and video: Challenges and perspective. EURASIP J Info Sec (5): 1–18 source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationreasoning
    § 2.2 Ciphers: Advanced Encryption Standard (AES) and real-time XOR · p. 5

    By using a continually changing key generated with a PRNG, confidentiality is greatly enhanced, if the initial seed can be securely distributed and the PRNG is sufficiently random.

    “However, by means of a continually changing key generated with a Pseudo-Random Number Generator (PRNG), the confidentiality is greatly enhanced, if the initial seed can be securely distributed and the PRNG is 'sufficiently random'.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation
    § 2.4 Standardized Video Codecs · p. 6

    The HEVC advantage of increased compression comes with an increase in codec complexity, which increases processing latency.

    “Though, the compression ratio can be increased by up to 50% by means of an HEVC codec that advantage comes with an increase in codec complexity, which in turn increases processing latency.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [44] Muhammad K, Hamza R, Ahmad J, Lloret J, Wang H, Baik SW (2018) Secure surveillance network for IoT systems using probabilistic image encryption. IEEE Trans Indust Inform 14(8):3679–3689 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning
    § 2.2 Ciphers: Advanced Encryption Standard (AES) and real-time XOR · p. 5

    If the same key is repeatedly used, XOR security is weak, vulnerable to a known-plaintext attack by XORing plaintext with ciphertext to output the key.

    “However, especially if the same key is repeatedly used, the security is weak, being vulnerable to a known-plaintext attack by XORing the plaintext with the ciphertext to output the key.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation
    § 2.2 Ciphers: Advanced Encryption Standard (AES) and real-time XOR · p. 5

    Compared to prior standardized symmetric ciphers, AES provides a shield against many attacks.

    “Additionally, compared to prior standardized symmetric ciphers, AES provides a shield against many attacks [34].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [34] Lin X, Ma H, Luo L, Chen Y (2012) No-reference video quality assessment in the compressed domain. IEEE Trans Consum Electron 58(2):505–512 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation
    § 2.3 Transcoding · p. 6

    It is possible to apply a form of transcoding to encrypted scalable video after packetization by employing unencrypted packet headers to guide truncation of encrypted packet payload.

    “It is also possible [41] to apply a form of transcoding to encrypted scalable video after packetization. This works by employing the unencrypted packet headers to guide truncation of the encrypted packet payload.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [41] Merkle P, Wang Y, Müller K, Smolic KA, Wiegand T (2009) Video plus depth format for mobile 3D services. Proc IEEE 3DTV Conf: 1–4 source

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationreasoning
    § 2.1 Selective encryption · p. 4

    In a long-term statistical sense there is no bitrate overhead from this type of SE, even though for particular video streams there may be some overhead.

    “Therefore, in a long-term statistical sense there is no bitrate overhead from this type of SE, even though for particular video streams it may turn out that there is some overhead.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: The signs of transform coefficients and motion vector differences follow a Uniform distribution before and after encryption.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation
    § 2.3 Transcoding · p. 6

    Closed loop architecture approximates the cascaded decoding-encoding architecture by reintroducing a transform stage.

    “This architecture approximates the cascaded decoding-encoding architecture by reintroducing a transform stage.”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites [17] Dorwin D, Smith J, Watson M, Bateman A (2017) Encrypted media extensions W3C recommendation. W3C

      Not checked in the free run.

    • cites [42] Moiron S, Faria S, Assunção P, Silva V, Navarro A (2007) H.264/AVC to MPEG-2 video transcoding architecture. Proc Conf Telecomm: 449–452

      Not checked in the free run.

    • cites [43] Morrison DG, Nilsson ME, Ghanbari M (1994) Reduction of bit-rate of compressed video while in its coded form. Proc 6th Int Workshop on Packet Video: 392–406

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation
    § 2.3 Transcoding · p. 5

    To deliver video streaming data to users connected with different types of networks and terminals may need to be adapted dynamically at intermediate locations within a network.

    “To deliver video streaming data to users connected with different types of networks and having different terminals may need to be adapted dynamically at intermediate locations within a network [39].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [39] Marr D, Hildreth E (1980) Theory of edge detection. Proc R Soc Lond B 207:187–217 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation
    § 2.3 Transcoding · p. 5

    It is possible to convert from one codec standard to another, such as H.264/AVC to MPEG-2 for broadcast over digital TV networks.

    “It is possible to convert from one codec standard to another, such as when a video encoded in a newer format, for example H.264/AVC, commonly employed for video over wireless, is converted to a legacy format such as MPEG-2, so that it can be broadcast over digital TV networks [37].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [37] Liu Y, Cheng C, Gue T et al (2016) A lightweight authenticated communication scheme for smart grid. IEEE Sensors J 26(3):836–842 source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation
    § 2.3 Transcoding · p. 6

    Open loop architecture is the fastest and simplest means of video transcoding.

    “The open loop architecture is the fastest and the simplest means of video trans-coding.”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites [17] Dorwin D, Smith J, Watson M, Bateman A (2017) Encrypted media extensions W3C recommendation. W3C

      Not checked in the free run.

    • cites [42] Moiron S, Faria S, Assunção P, Silva V, Navarro A (2007) H.264/AVC to MPEG-2 video transcoding architecture. Proc Conf Telecomm: 449–452

      Not checked in the free run.

    • cites [43] Morrison DG, Nilsson ME, Ghanbari M (1994) Reduction of bit-rate of compressed video while in its coded form. Proc 6th Int Workshop on Packet Video: 392–406

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation
    § 2.2 Ciphers: Advanced Encryption Standard (AES) and real-time XOR · p. 5

    AES has low memory requirements and has been designed to guard against timing attacks.

    “AES has low memory requirements and has been designed to guard against timing attacks.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [33] Lee J, Kim S, Lim K, Lee S (2015) A fast CU size decision algorithm for HEVC. IEEE Trans Circ Syst Video Technol 25(3):411–421 source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation
    § 5.1 Crypto-transcoding with AVC and HEVC · p. 12

    Edge detection and pixelation serve to check the effectiveness of encryption.

    “Notice that a recent way of testing the effectiveness of encryption is to apply edge detection to the distorted frame [80] and pixelation also serves to check the effectiveness of the encryption.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [80] Zhou J, Liu X, Member OCA, Tang YY (2014) Designing an efficient image encryption-then-compression system via prediction error clustering and random permutation. IEEE Trans Info Foren Sec 9(1):39–50 source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation
    § 2.3 Transcoding · p. 6

    Closed-loop architecture is the best compromise for the video transcoding process, though additional latency occurs due to extra latency.

    “This architecture is the best compromise for the video transcoding process, though additional latency occurs due to extra latency.”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites [17] Dorwin D, Smith J, Watson M, Bateman A (2017) Encrypted media extensions W3C recommendation. W3C

      Not checked in the free run.

    • cites [42] Moiron S, Faria S, Assunção P, Silva V, Navarro A (2007) H.264/AVC to MPEG-2 video transcoding architecture. Proc Conf Telecomm: 449–452

      Not checked in the free run.

    • cites [43] Morrison DG, Nilsson ME, Ghanbari M (1994) Reduction of bit-rate of compressed video while in its coded form. Proc 6th Int Workshop on Packet Video: 392–406

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationnone given
    § 2.2 Ciphers: Advanced Encryption Standard (AES) and real-time XOR · p. 5

    XOR cipher is vulnerable to flipping of ciphertext so that a valid but incorrect message is generated (malleability).

    “The cipher is also vulnerable to flipping of the cipher text so that a valid but incorrect 'message' is generated. This effect is termed malleability, when decryption takes place.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationcitation
    § 2.2 Ciphers: Advanced Encryption Standard (AES) and real-time XOR · p. 5

    A long-term AES key can be established through Diffie-Helman key distribution protocol; thereafter a session key is AES encrypted and included in header of XOR SE file.

    “Another possibility is to establish a long-term AES key through the Diffie-Helman key distribution protocol [35]. Thereafter, a session key, possibly for each video frame depending on the level of protection required, is then AES encrypted and included in the header of the XOR SE file.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [35] Liu F, Koenig H (2010) A survey of video encryption algorithms. Comput Sec 25(1):3–15 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation
    § 5.2 HEVC encoded Crypto-transrating with XOR and AES ciphers · p. 16

    AES in Cipher Feedback mode (CFB) works as a stream-cipher.

    “In the Cipher Feedback mode (CFB), AES works as a stream-cipher [28]”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [28] Huyn-Thu Q, Ghanbari M (2008) Scope of validity of PSNR in image/video quality assessment. Electron Letts 44(13):800–801 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation
    § 5.2 HEVC encoded Crypto-transrating with XOR and AES ciphers · p. 16

    AES-CFB benefits from a self-synchronization mechanism for real-time transmission.

    “and additionally benefits from a self-synchronization mechanism for real-time transmission.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [28] Huyn-Thu Q, Ghanbari M (2008) Scope of validity of PSNR in image/video quality assessment. Electron Letts 44(13):800–801 source

  • May combine two claims.

    Interpretationreasoning
    § 6 Conclusions · p. 19

    For longer video streams, including broadcast TV and films, the savings in bit-rate and computation time will be considerable.

    “For longer video streams, including broadcast TV and films, the savings in bit-rate and computation time will be considerable.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The tested video sequences are representative of general video content, including longer video streams such as broadcast TV and films.
50 of 50

  • The paper relies on this without support.

    The signs of transform coefficients and motion vector differences follow a Uniform distribution before and after encryption.

    “In addition, signs can reasonably be expected to follow a Uniform distribution, before and after encryption.”

  • The paper relies on this without support.

    Much legacy video content remains in H.264/AVC format or MPEG-2 codec format.

    “However, much legacy video content remains in H.264/AVC format or even in MPEG-2 codec format.”

  • The paper relies on this without support.

    Content protection is necessary to ensure the commercial viability of a video streaming service.

    “Content protection is deemed necessary to ensure the commercial viability of a video streaming service because otherwise there would be no monetary incentive to make videos and distribute them.”

  • The paper relies on this without support.

    At higher QPs, more transform coefficients are reduced to zero and do not appear in the input to the CABAC engine.

    “At higher QPs, more of the TCs are reduced to zero and, hence, do not appear in the input to the CABAC engine.”

  • The paper relies on this without support.

    Histogram entropy/randomness and visual distortion are adequate indicators of encryption strength or security.

    “This Fig. 10 (c, e) indicates that the AES-CFB cipher results in an increase in entropy or randomness across the pixel values relative to using the XOR cipher.”

  • The paper relies on this without support.

    The tested video sequences are representative of general video content, including longer video streams such as broadcast TV and films.

    “For longer video streams, including broadcast TV and films, the savings in bit-rate and computation time will be considerable.”

  • The paper relies on this without support.

    The machine employed for the experiments is representative of typical transcoding hardware.

    “From experiments, the gains from using SE are substantial in terms of limited bitrate overhead and (according to the machine employed) several seconds saved in encoding alone, even for short video sequences.”

  • The paper relies on this without support.

    There is a need to reduce processing latency at intermediate transcoders arising from encryption for content protection and possibly privacy protection at the user.

    “The main contribution has been to reduce the processing latency at intermediate transcoders arising from a need to encrypt video for reasons of content protection and possibly privacy protection at the user.”
8 of 8

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • HEVC increases compression efficiency but has less error-resilient features and is more sensitive to channel bit errors because of more encoded redundant content.

    “The HEVC codec increases the compression efficiency but has less error-resilient features [13] and is more sensitive to channel bit errors because of more encoded redundant content [14].”

    From VQProtect: Lightweight Visual Quality Protection for Error-Prone Selectively Encrypted Video Streaming (2022) · cites this paper as [13]

    Section 2.4 (p. 6) says HEVC coding refinements can increase the compression ratio by up to 50%. Our copy does not say HEVC has fewer error-resilience features or is more sensitive to channel bit errors because of encoded redundant content; searches for those terms found no supporting passage. (LLM, conf=0.84)

    Experimental — accuracy not yet measured

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Encryption cipher for the Section 5.1 experiments is inconsistent

    Minor

    The evaluation setup says the selected signs were encrypted with AES, but Section 5.2 identifies the Section 5.1 results and figures as tests of XOR encryption. The cipher used for those results is therefore unclear.

1 of 1

Cites

Flagged first, then by citations

Research works printed but not linked: 5

  1. Bing B (2016) Video over wireless. McGraw-Hill Educ, New York, NY
  2. Boho A, Van Wallendael G, Dooms A, De Cock J, Schelkens P, Prenel B, Van de Wall R (2013) End-to-end security for video distribution. IEEE Sig Process Mag 30(2):97–107 · DOI · Paper page
  3. Morrison DG, Nilsson ME, Ghanbari M (1994) Reduction of bit-rate of compressed video while in its coded form. Proc 6th Int Workshop on Packet Video: 392–406
  4. Saleh MA, Tahir NMD, Hashim AH (2018) Fast selective encryption for video stream of high efficiency video coding standard. J Theor Appl Info Technol 96(20):6806–6816
  5. Tamanna S (2013) Transcoding H.265/HEVC”, M.S. thesis, BTH, Karlskrona, Sweden

Cited by

Sorted by citations

Related

Authors
Shah, R. A., Asghar, M. N., Abdullah, S., Fleury, M., & Gohar, N.
Affiliations
  • Islamia University of Bahawalpur (PK)
  • University of Essex (GB)
  • Shaheed Benazir Bhutto Women University Peshawar (PK)
Corresponding author
Fleury, Martin
ORCID
Venue
Multimedia Tools and Applications
Publisher
Springer Science and Business Media LLC
Published
2019-03-22
Volume
78
Issue
15
Pages
21455-21484
ISSN
  • 1380-7501 (ISSN-L)
  • 1573-7721
DOI
10.1007/s11042-019-7451-5
OpenAlex
W2916264022
Semantic Scholar
e31254b44d1d0ffef8eb2853d40274377fe8a733
Cited by
12
References
73
Citations by year
2025: 3 · 2023: 1 · 2022: 2 · 2021: 3 · 2020: 3
FWCI
1.29
Type
journal article
Language
en
Open access
Licence
www.springer.com/tdm (tdm)
Locations
  • Multimedia Tools and Applications (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • Research@THEA (submittedVersion, cc-by)
  • arXiv (Cornell University)
Funders
  • University of Essex
  • Islamia University of Bahawalpur
  • University of Oxford
Topics
  • Physical Sciences › Computer Science › Signal Processing › Video Coding and Compression Technologies
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Chaos-based Image/Signal Encryption
  • Physical Sciences › Computer Science › Information Systems › Digital Rights Management and Security
Keywords
selective encryption, H.264/AVC, HEVC, HTTP adaptive streaming, content protection, cloud data centers, video distortion, AES, XOR encryption

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-23.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-23.

Cited sources are checked by gpt-6-sol.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review