Conclusion and Future Work
In Chaps. 1 – 5 , we have comprehensively reviewed the fundamental concepts and designing techniques related to digital audio watermarking. Specifically, the general background and fundamental concepts are provided in Chap. 1 . The issues regarding the
- PDF / 141,445 Bytes
- 6 Pages / 439.37 x 666.142 pts Page_size
- 9 Downloads / 187 Views
Conclusion and Future Work
Abstract In Chaps. 1–5, we have comprehensively reviewed the fundamental concepts and designing techniques related to digital audio watermarking. Specifically, the general background and fundamental concepts are provided in Chap. 1. The issues regarding the problem on imperceptibility for audio watermarking are discussed in Chap. 2. Typical audio watermark embedding and extraction techniques, from classical ones to their latest variations and developments, are detailed in Chap. 3. Furthermore, two emerging audio watermarking topics, i.e., reversible audio watermarking and cryptography-aided audio watermarking, are investigated in Chaps. 4 and 5, respectively. In this chapter, we discuss the limitations, challenges, and future work on audio watermarking.
6.1 Limitations and Challenges The limitations of existing audio watermarking systems are closely associated with the performance criteria, i.e., imperceptibility, robustness, security, capacity, and computational complexity. First, we note that it is unnecessary for all audio watermarking systems to simultaneously satisfy all of the performance criteria. In fact, this is currently not possible, and more importantly, none of the existing applications has such a requirement. Therefore, the current limitations commonly exist in those systems which trade off a few criteria. For imperceptibility, the criterion that nearly all audio watermarking systems have to account for, the current state-of-the-art methods are psychoacoustic model based analytical tuning or automated heuristic tuning according to audio quality measurements. The former suffers from watermark removal by lossy compression, and the latter suffers from inefficiency. For fragile or semi-fragile audio watermarking, e.g., reversible or cryptography-aided audio watermarking, the channel between transmitter and receiver is secure, thus robustness does not need to be considered during system design. In this way, limitations mainly lie in the trade-off between embedding capacity and imperceptibility. Generally, if watermark embedding is not performed at bit level (such as reversible embedding methods), the commonly applied frame based embedding methods are limited by low embedding rate at 1 bit per frame. Although recent proposals such © The Author(s) 2017 Y. Xiang et al., Digital Audio Watermarking, SpringerBriefs in Signal Processing, DOI 10.1007/978-981-10-4289-8_6
85
86
6 Conclusion and Future Work
as [1] can offer certain improvements on embedding capacity, the resultant one is still relatively low, especially when compared with image watermarking. In applications such as broadcast monitoring and second screen, computational efficiency becomes one of the major criteria. In addition, security has been an important but less studied criterion for watermarking [2, 3]. Due to the nature of watermarking which indicates the difference between watermarking and cryptography, watermarking systems could not achieve the security level of encryption. To date, most existing systems use a
Data Loading...
