Search

Orthogonal polyphase sequence sets with desirable properties have numerous applications in sensing fields, e.g., a monostatic radar transmitting such waveforms can effectively counter the coherent repeater jamming (CRJ). However, due to imperfections in practical systems, the delicately designed sequences may suffer significant distortion and reduction in performance when transmitted as radar waveforms. In this paper, we focus on the problem of designing polyphase sequence sets with good correlation properties under spectral distortion. To address this issue, we first introduce a new metric called distortion correlation properties (DCorr), which quantifies the quality of such sequence sets. Subsequently, a majorization-minimization (MM)-based iterative algorithm is developed to solve the problem of minimizing DCorr. A number of numerical examples are provided to demonstrate the effectiveness of the proposed method in improving the correlation properties of polyphase sequence sets under spectral distortion. The results also validate that our proposed algorithms can effectively mitigate the performance reduction caused by the imperfections in practical radar systems simply. [ABSTRACT FROM AUTHOR]

Noncontact vital sign detection is widely used in finding victims in post-disaster search and rescue, through-wall surveillance (TWS), and medical diagnosis and monitoring. Human subject can be remotely sensed by extracting life activities such as respiration and heartbeat. However, the signal to noise ratio (SNR) is often low, particularly in complex environments, which results in errors in both range and respiration frequency (RF) estimation. To improve the accuracy, an improved system for the vital sign detection is presented which is based on impulse ultra-wideband (UWB) radar. The range is determined based on the short-time Fourier transform (STFT) of the standard deviation of the received signals. Further, the ensemble empirical mode decomposition (EEMD) based frequency accumulation (FA) technique is presented to determine RF. Performance results are presented which show that the proposed method is superior to several well-known techniques. [ABSTRACT FROM AUTHOR]

The end-to-end speech translation (ST) model usually adopts the encoder-decoder structure, which takes the speech of the source language as input and directly outputs its translation result in the target language. Since the model performs cross-modal translation, it needs to extract semantically rich representations from speech. Compared with text, speech is fine-grained and contains more noise, which puts a great burden on the encoder of the ST model. The modal gap between speech and text usually causes the ST model to perform inferior to the corresponding machine translation (MT) model. To bridge the cross-modal gap, this paper proposes to use adversarial training to relieve the burden on the ST encoder by providing internal supervision signals. In this approach, the encoder in the ST model can extract representations that contain rich semantics, which greatly improves performance. Experiments on the datasets of Augmented Librispeech English-French and MuST-C English-German show the effectiveness of our method. Further analysis indicates that our proposed method can perform well in low-resource conditions compared to strong baselines. [ABSTRACT FROM AUTHOR]

This study proposes a novel blind watermarking algorithm based on radial harmonic Fourier moments (RHFMs) as well as discrete wavelet transform and discrete cosine transform (DWT-DCT) to achieve copyright protection. A non-adjacent coupled map lattice chaotic system was used to improve the safety in the watermark embedding process. First, a zero-watermark was obtained by performing DCT on the original image. Meanwhile, the direction information of the origin was calculated and embedded in the RHFM domain. During normal watermark embedding, the image-embedded direction information was transformed by a one-level DWT and the low-frequency sub-band was divided into 8 × 8 non-overlapping blocks. Each block was transformed by DCT, and two low-frequency coefficients were selected to embed the encrypted watermark. Experimental results demonstrated that the proposed algorithm is robust against rotation and other attacks. The analysis also showed that the proposed algorithm satisfies the invisibility of watermarking algorithms and has good security performance. [ABSTRACT FROM AUTHOR]

Zero-watermarking algorithms provide the means of effective, lossless copyright protection, but the existing ones can only resist a limited variety of geometric attacks. To overcome this limitation, a novel zero-watermarking algorithm resistant to an extensive variety of geometric attacks is suggested, in which fractional-order Exponent-Fourier moments (FrEFMs) and discrete Fourier transform (DFT) are used as complementary geometric invariant features to construct zero-watermarks. Relying on the rotation, scaling, aspect ratio changing and cropping invariance of FrEFMs amplitude and the perfect translation invariance of DFT amplitude, our algorithm can effectively resist a wider variety of geometric and common image processing attacks, which has been experimentally illustrated in this work. Therefore, the proposed zero-watermarking algorithm is superior to other cutting-edge ones. [ABSTRACT FROM AUTHOR]

The wide dissemination of images makes its protection problems urgent. Chaos theory is widely applied since it provides better solutions to image protection. This paper introduces a novel chaotic system that has line equilibrium points. Chaotic signals generated by chaotic systems are used in encryption and steganography to enhance security. Besides, the scrambling-diffusion encryption system is combined with the least significant bit (LSB)-pixel value differencing (PVD) steganography algorithm. The proposed algorithm process is reversible and the extracted image is lossless. And the whole system is designed and implemented on a field programmable gate array (FPGA), which improves the embedding efficiency and achieves high throughput while ensuring good image quality. [ABSTRACT FROM AUTHOR]

• A new type of QPCET, namely NoQPCET, is constructed. • A new color image robust zero-watermarking algorithm is designed based on NoQPCET. • This algorithm has very strong security by using chaotic system. • The low-order NoQPCET coefficients are used to extract the image feature. The robust watermarking method embeds the information in the image to protect the copyright. The embedding process will destroy the original image and cannot be effectively applied to the images used in the medicine and military with high integrity demand. A robust zero-watermarking algorithm for lossless copyright protection of color images without changing the original image was proposed in this paper. A novel quaternion polar complex exponential transform (NoQPCET) was firstly constructed with effectively improved performance; then the NoQPCET and chaotic system based robust color image zero-watermarking algorithm was suggested for lossless copyright protection of color images. In the proposed algorithm, the robustness against geometric attacks was enhanced due to the excellent geometric invariance of NoQPCET, and the security was also improved because of the initial value sensitivity of the chaotic system. Experimental results showed that this algorithm had very strong robustness against both geometric attacks and common image processing attacks, and was superior to other algorithms. [ABSTRACT FROM AUTHOR]