Embedded Multimedia Security Systems by Amit Pande & Joseph Zambreno

Author:Amit Pande & Joseph Zambreno
Language: eng
Format: epub
Publisher: Springer London, London

5.1 Introduction

The recent emergence of embedded multimedia applications such as mobile-TV, surveillance, video messaging, and tele-medicine have increased the scope of multimedia in our personal lives. These applications increase the concerns regarding privacy and security of the targeted subjects. Another growing concern is the protection and enforcement of intellectual property rights for images and videos. These and other issues such as image authentication, rights validation, identification of illegal copies of a (possibly forged) image are grouped and studied under the label of Digital Rights Management (DRM).

The computer security protocols (e.g. SSL [36], TLS [4]) and cryptographic ciphers (e.g. AES [10], DES [11], IDEA [18]) drive much of the world’s electronic communications, commerce, and storage. These techniques have been used for conventional multimedia encryption and authentication.

In one version of these schemes, some form of private-key encryption algorithm is applied over the full or partial output bit stream from the video compression engine. This naive approach is usually suitable for text, and sometimes for small bitrate audio, image, and video files that are being sent over a fast dedicated channel. Secure Real-time Transport Protocol, or for short SRTP [2], is an application of the naive approach. In SRTP, multimedia data are packetized and each packet is individually encrypted using AES. The naive approach enables the same level of security as that of the used conventional cryptographic cipher.

Consequently, a multimedia compression engine (such as a MPEG or H.264 encoder [34]) has an additional encryption engine to ensure multimedia security. Depending on the scheme used, the encryption operation is performed either at some intermediate level during compression or after the final compression. However, these cryptographic ciphers require a large amount of computational resources and often incur large latencies. Hardware implementations of AES are often pipelined, leading to a significantly large latency for real-time applications (31 cycles for AES [14]). The partial or selective encryption schemes [21] encode only the important parts of multimedia data to reduce these computational overheads.

The large data volumes, interactive operations, real-time responses, and scalability features that are inherent to real-time multimedia delivery restrict the practical application of these naive cryptographic schemes. Selective encryption schemes have been proposed in research literature [3, 9, 19, 21, 26] to reduce the computational requirements of full encryption schemes. Lian et al. [19] present a scheme for encryption of Discrete Cosine Transform (DCT) coefficients’ signs and watermarking of DCT coefficients. Lian et al. [19] use Exp-Goloumb codes for the encryption operation. Cheng and Li [6] propose a DWT-based partial encryption scheme which encrypts only a part of compressed data. Only 13–27 % of the output from quadtree compression algorithms is encrypted for typical images. A good summary of efforts in selective or partial encryption of images can be found in [21].

Furthermore, embedded multimedia systems have constraints on power consumption, available computation power, and performance. Real-time embedded systems face additional constraints on power consumption, hardware size and heat generation in the chip which requires design and mapping of computation-savvy encryption schemes for such architectures. Recently, power-aware designs have been proposed for video coding in embedded scenarios [5].

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.