Figure - available from: Advances in Multimedia
This content is subject to copyright. Terms and conditions apply.
Comparison of the first three seconds of signal waveforms. (a) Original signal; (b) encrypted signal.
Source publication
We present a scheme for commutative watermarking-encryption (CWE) of audio data and demonstrate its robustness against an important class of attacks, Time-Scale Modifications (TSM). In addition, we show how the proposed CWE scheme can be integrated into a cryptographic protocol enabling public verification of the embedded mark without disclosing th...
Citations
... In (Schmitz & Gruber, 2017) it has been demonstrated that the basic approach also works for audio data. Here, the permutation cipher is applied to discrete sample values obtained from sampling the analogous audio signal. ...
... In this section two extensions of the original CWE scheme for still image data are reported. The first of these extensions is an adaption of the scheme for audio data, which was first presented in (Schmitz & Gruber, 2017). The second extension is concerned with using the watermark for authenticating the content of an image (Schmitz et al., 2015). ...
Histogram-based watermarking schemes are invariant to pixel permutations and can thus be combined with permutation-based ciphers to form a commutative watermarking-encryption scheme. In this chapter, the authors demonstrate the feasibility of this approach for audio data and still image data. Typical histogram-based watermarking schemes based on comparison of histogram bins are prone to desynchronization attacks, where the whole histogram is shifted by a certain amount. These kind of attacks can be avoided by synchronizing the embedding and detection processes, using the mean of the histogram as a calibration point. The resulting watermarking scheme is resistant to three common types of shifts of the histogram, while the advantages of previous histogram-based schemes, especially commutativity of watermarking and permutation-based encryption, are preserved. The authors also report on the results of testing robustness of the still image watermark against JPEG and JPEG2000 compression and on the possibility of using histogram-based watermarks for authenticating the content of an image.
The commutative encryption-watermarking (CEW), in which the encryption and watermarking are distinct and commutative, has become a promising method for protecting data security. For vector maps, current methods combine encryption with watermarking directly, but cannot satisfy the CEW properties. For example, the watermark embedding and encryption are not commutative and lack flexibility. Considering the essential characteristics of vector maps, this paper proposes a CEW method based on the congruence relationship and geometric feature for vector maps. In the proposed CEW method, the angles and the distance ratios are selected as the geometric features which provide separate spaces for cryptography and watermarking operations. They are encrypted by making the original values and encrypted values congruent. At the same time, the watermark is embedded into their residue. As the residue keeps invariant under congruence operations, the watermark can be extracted from either the plaintext or the ciphertext. Experiments are conducted to verify the commutativity between encryption and watermarking. Besides, the superiority in terms of watermark robustness and capacity of the proposed method has been also compared with other encryption-combining-watermarking methods for vector maps.
Word embeddings are fundamentally a form of word representation that links the human understanding of knowledge meaningfully to the understanding of a machine. The representations can be a set of real numbers (a vector). Word embeddings are scattered depiction of a text in an n-dimensional space, which tries to capture the word meanings. This paper aims to provide an overview of the different types of word embedding techniques. It is found from the review that there exist three dominant word embeddings namely, Traditional word embedding, Static word embedding, and Contextualized word embedding. BERT is a bidirectional transformer-based Contextualized word embedding which is more efficient as it can be pre-trained and fine-tuned. As a future scope, this word embedding along with the neural network models can be used to increase the model accuracy and it excels in sentiment classification, text classification, next sentence prediction, and other Natural Language Processing tasks. Some of the open issues are also discussed and future research scope for the improvement of word representation.
The encoding complexity of an image format is a vigorously updating area of study in the field of two-layer protection with wavelet transform compression. In the proposed method, hybrid 2D-FDCT watermarking and RSA encryption for multispectral images predicted an efficient system. This approach satisfies the encryption security, robustness and classification accuracy retention of an algorithm. The two-layer protection of encrypted and embedded watermark image followed by wavelet transform compression minimizes the file size in the exhaustive process for encoding. An important merit is that encoding time is very much reduced in contrast to other security and compression mechanisms. The enhanced value of PSNR as well as trade-off of MES, normalized cross-correlation, the average difference and structural content improves the storage large file size medical image and improves bandwidth to an acceptable level.
Keratoconus detection and diagnosis has become a crucial step of primary importance in the preoperative evaluation for the refractive surgery. With the ophthalmology knowledge improvement and technological advancement in detection and diagnosis, artificial intelligence (AI) technologies like machine learning (ML) and deep learning (DL) play an important role. Keratoconus being a progressive disease leads to visual acuity and visual quality. The real challenge lies in acquiring unbiased dataset to predict and train the deep learning models. Deep learning plays a very crucial role in upturning ophthalmology division. Detecting early stage keratoconus is a real challenge. Hence, our work aims to primarily focus on detecting an early stage and multiple classes of keratoconus disease using deep learning models. This review paper highlights the comprehensive elucidation of machine learning and deep learning models used in keratoconus detection. The research gaps are also identified from which to obtain the need of the hour for detecting keratoconus in humans even before the symptoms are visible.
Histogram-based watermarking schemes are invariant to pixel permutations and can thus be combined with permutation-based ciphers to form a commutative watermarking-encryption scheme. In this chapter, the authors demonstrate the feasibility of this approach for audio data and still image data. Typical histogram-based watermarking schemes based on comparison of histogram bins are prone to desynchronization attacks, where the whole histogram is shifted by a certain amount. These kind of attacks can be avoided by synchronizing the embedding and detection processes, using the mean of the histogram as a calibration point. The resulting watermarking scheme is resistant to three common types of shifts of the histogram, while the advantages of previous histogram-based schemes, especially commutativity of watermarking and permutation-based encryption, are preserved. The authors also report on the results of testing robustness of the still image watermark against JPEG and JPEG2000 compression and on the possibility of using histogram-based watermarks for authenticating the content of an image.
