In this paper, we propose N-Dimensional (ND) Tensor Supervised Neighborhood Embedding (ND TSNE) for discriminant feature representation, which is used for view-based object recognition. ND TSNE uses a general Nth order tensor discriminant and neighborhood-embedding analysis approach for object representation. The benefits of ND TSNE include: (1) a natural way of representing data without losing structure information, i.e., the information about the relative positions of pixels or regions; (2) a reduction in the small sample size problem, which occurs in conventional supervised learning because the number of training samples is much less than the dimensionality of the feature space; (3) preserving a neighborhood structure in tensor feature space for object recognition and a good convergence property in training procedure. With Tensor-subspace features, the random forests is used as a multi-way classifier for object recognition, which is much easier for training and testing compared with multi-way SVM. We demonstrate the performance advantages of our proposed approach over existing techniques using experiments on the COIL-100 and the ETH-80 datasets.

In the past few years, discriminant analysis and manifold learning have been widely used in feature extraction. Recently, the sparse representation technique has advanced the development of pattern recognition. In this paper, we combine both discriminant analysis and manifold learning with sparse representation technique and propose a novel feature extraction approach named sparsity preserving embedding with manifold learning and discriminant analysis. It seeks an embedded space, where not only the sparse reconstructive relations among original samples are preserved, but also the manifold and discriminant information of both original sample set and the corresponding reconstructed sample set is maintained. Experimental results on the public AR and FERET face databases show that our approach outperforms relevant methods in recognition performance.

Accurate device models are very important for the design of high-frequency circuits. One of the factors degrading the accuracy of device models appears during the de-embedding procedure. Generally, to obtain device characteristics without parasitic elements such as pads, a de-embedding procedure is essential. However, some errors are introduced during this procedure, which degrades the accuracy of device models. In this paper, we demonstrate that such errors due to de-embedding are cancelled in cascade circuit design, meaning that cascade circuits can be designed without knowing the actual characteristics of devices. Because it is difficult to know the actual characteristics of devices at a high frequency, the cancellation of the de-embedding error is expected to improve the accuracy of device models at high frequencies. After giving a theoretical treatment of de-embedding error cancellation, we report the results of simulations and measurements performed for verification.

We present an improvement to the existing steganography-based bandwidth extension scheme. Enhanced WB (wideband) speech quality is achieved by embedding multiple highband spectral gains into a G.711 bitstream. The number of spectral gains is selected by optimizing the quantity of the embedding data with respect to the quality of the extended WB speech. Compared to the existing method, the proposed scheme improves the WB PESQ (Perceptual Evaluation of Speech Quality) score by 0.334 with negligible degradation of the embedded narrowband speech.

Locally linear embedding (LLE) is a well-known method for nonlinear dimensionality reduction. The mathematical proof and experimental results presented in this paper show that the neighborhood sizes in LLE must be smaller than the dimensions of input data spaces, otherwise LLE would degenerate from a nonlinear method for dimensionality reduction into a linear method for dimensionality reduction. Furthermore, when the neighborhood sizes are larger than the dimensions of input data spaces, the solutions to LLE are not unique. In these cases, the addition of some regularization method is often proposed. The experimental results presented in this paper show that the regularization method is not robust. Too large or too small regularization parameters cannot unwrap S-curve. Although a moderate regularization parameters can unwrap S-curve, the relative distance in the input data will be distorted in unwrapping. Therefore, in order to make LLE play fully its advantage in nonlinear dimensionality reduction and avoid multiple solutions happening, the best way is to make sure that the neighborhood sizes are smaller than the dimensions of input data spaces.

A pairing over an elliptic curve E/Fpm to an extension field of Fpmk has begun to be attractive in cryptosystems, from the practical and theoretical point of view. From the practical point of view, many cryptosystems using a pairing, called the pairing-based cryptosystems, have been proposed and, thus, a pairing is a necessary tool for cryptosystems. From the theoretical point of view, the so-called embedding degree k is an indicator of a relationship between the elliptic curve Discrete Logarithm Problem (ECDLP) and the Discrete Logarithm Problem (DLP), where ECDLP over E(Fpm) is reduced to DLP over Fpmk by using the pairing. An elliptic curve is determined by mathematical parameters such as the j-invariant or order of an elliptic curve, however, explicit conditions between these mathematical parameters and an embedding degree have been described only in a few degrees. In this paper, we focus on the theoretical view of a pairing and investigate a new condition of the existence of elliptic curves with pre-determined embedding degrees. We also present some examples of elliptic curves over 160-bit, 192-bit and 224-bit Fpm with embedding degrees k < (log p)2 such as k=10, 12, 14, 20, 22, 24, 28.

Subspace learning based face recognition methods have attracted considerable interest in recent years, including Principal Component Analysis (PCA), Independent Component Analysis (ICA), Linear Discriminant Analysis (LDA), and some extensions for 2D analysis. However, a disadvantage of all these approaches is that they perform subspace analysis directly on the reshaped vector or matrix of pixel-level intensity, which is usually unstable under illumination or pose variance. In this paper, we propose to represent a face image as a local descriptor tensor, which is a combination of the descriptor of local regions (K*K-pixel patch) in the image, and is more efficient than the popular Bag-Of-Feature (BOF) model for local descriptor combination. Furthermore, we propose to use a multilinear subspace learning algorithm (Supervised Neighborhood Embedding-SNE) for discriminant feature extraction from the local descriptor tensor of face images, which can preserve local sample structure in feature space. We validate our proposed algorithm on Benchmark database Yale and PIE, and experimental results show recognition rate with our method can be greatly improved compared conventional subspace analysis methods especially for small training sample number.

This paper extends the Brezing-Weng method by parameterizing the discriminant D by a polynomial D(x). To date, the maximum of CM discriminant can be adequately addressed is about 14-digits. Thus the degree of the square free part of D(x) has to be sufficiently small. By making the square free part of D(x) a linear monomial, the degree of the square free part is small and by substituting x to some quadratic monomial, pairing-friendly curves with various discriminants can be constructed. In order that a square free part of D(x) is of the form ax, ax has to be a square element as a polynomial representation in a number field. Two methods are introduced to apply this construction. For k = 5, 8, 9, 15, 16, 20, 24 and 28, the proposed method gives smaller ρ value than those in previous studies.

This paper proposes a de-embedding method for on-chip S-parameter measurements at mm-wave frequency. The proposed method uses only two transmission lines with different length. In the proposed method, a parasitic-component model extracted from two transmission lines can be used for de-embedding for other-type DUTs like transistor, capacitor, inductor, etc. The experimental results show that the error in characteristic impedance between the different-length transmission lines is less than 0.7% above 40 GHz. The extracted pad model is also shown.

An L-2L through-line de-embedding method has been verified up to millimeter wave frequency. The parasitics of the pad can be modeled from the L-2L through-line. Measurement results of the transmission lines and transistors can be de-embedded by subtracting the parasitic matrix of the pad. Therefore, the de-embedding patterns, which is used for modeling active and passive devices, decrease greatly and the chip area also decreases. A one-stage amplifier is firstly implemented for helping verifying the de-embedding results. After that a four-stage 60 GHz amplifier has been fabricated in CMOS 65 nm process. Experimental results show that the four-stage amplifier realizes an input matching better than -10.5 dB and an output matching better than -13 dB at 61 GHz. A small signal power gain of 16.4 dB and a 1 dB output compression point of 4.6 dBm are obtained with a DC current consumption of 128 mA from a 1.2 V power supply. The chip size is 1.5 mm 0.85 mm.

A secure method for steganography is proposed. Pixel-value differencing (PVD) steganography and bit-plane complexity segmentation (BPCS) steganography have the weakness of generating blocky effects and noise in smooth areas and being detectable with steganalysis. To overcome these weaknesses, a secure bit-plane based steganography method on the spatial domain is presented, which uses a robust measure to select noisy blocks for embedding messages. A matrix embedding technique is also applied to reduce the change of cover images. Given that the statistical property of cover images is well preserved in stego-images, the proposed method is undetectable by steganalysis that uses RS analysis or histogram-based analysis. The proposed method is compared with the PVD and BPCS steganography methods. Experimental results confirm that the proposed method is secure against potential attacks.

A flexible noise de-embedding method for on-wafer microwave measurements of silicon MOSFETs is presented in this study. We use the open, short, and thru dummy structures to subtract the parasitic effects from the probe pads and interconnects of a fixtured MOS transistor. The thru standard are used to extract the interconnect parameters for subtracting the interconnect parasitics in gate, drain, and source terminals of the MOSFET. The parasitics of the dangling leg in the source terminal are also modeled and taken into account in the noise de-embedding procedure. The MOS transistors and de-embedding dummy structures were fabricated in a standard CMOS process and characterized up to 20 GHz. Compared with the conventional de-embedding methods, the proposed technique is accurate and area-efficient.

Image enhancement plays an important role in many machine vision applications on images captured in low contrast and low illumination conditions. In this study, we propose a new method for image enhancement based on analysis on embedded surfaces of images. The proposed method gives an insight into the relationship between the image intensity and image enhancement. In our method, scaled surface area and the surface volume are proposed and used to reconstruct the image iteratively for contrast enhancement, and the illumination of the reconstructed image can also be adjusted simultaneously. On the other hand, the most common methods for measuring the quality of enhanced images are Mean Square Error (MSE) or Peak Signal-to-Noise-Ratio (PSNR) in conventional works. The two measures have been recognized as inadequate ones because they do not evaluate the result in the way that the human vision system does. This paper also presents a new framework for evaluating image enhancement using both objective and subjective measures. This framework can also be used for other image quality evaluations such as denoising evaluation. We compare our enhancement method with some well-known enhancement algorithms, including wavelet and curvelet methods, using the new evaluation framework. The results show that our method can give better performance in most objective and subjective criteria than the conventional methods.

This study has demonstrated that the clock tree construction in an SoC should be expanded to consider the intrinsic delay and skew of each IP's clock sink. A novel algorithm, called GDME, is proposed to combine grey relational clustering and DME approach for solving the problem of clock tree construction. Grey relational analysis can cluster the best pair of clock sinks and that guide a tapping point search for a DME algorithm for constructing a clock tree with zero skew and minimal delay. Experimentally, the proposed algorithm always obtains an RC- or RLC-based clock tree with zero skew and minimal delay for all the test cases and benchmarks. Experimental results demonstrate that the GDME improves up to 3.74% for total average in terms of total wire length compared with other DME algorithms. Furthermore, our results for the zero-skew RLC-based clock trees compared with Hspice are 0.017% and 0.2% lower for absolute average in terms of skew and delay, respectively.

Self embedding watermarking is a technique used for tamper detection, localization and recovery. This letter proposes a novel self embedding scheme, in which the halftone version of the host image is exploited as a watermark, instead of a JPEG-compressed version used in most existing methods. Our scheme employs a pixel-wise permuted and embedded mechanism and thus overcomes some common drawbacks of the previous methods. Experimental results demonstrate our technique is effective and practical.

The major application of digital data hiding techniques is to deliver confidential data secretly via public but unreliable computer networks. Most of the existing data hiding schemes, however, exploit the raw data of cover images to perform secret communications. In this paper, a novel data hiding scheme was presented with the manipulation of images based on the compression of side-match vector quantization (SMVQ). This proposed scheme provided adaptive alternatives for modulating the quantized indices in the compressed domain so that a considerable quantity of secret data could be artfully embedded. As the experimental results demonstrated, the proposed scheme indeed provided a larger payload capacity without making noticeable distortions in comparison with schemes proposed in earlier works. Furthermore, this scheme also presented a satisfactory compression performance.

Steganographic methods usually produce distortions in cover images due to the process of embedding secret bits. These distortions are hard to remove, and thus the cover image cannot be recovered. Although the distortions are always small, they cannot be allowed for some sensitive applications. In this paper, we propose a reversible embedding scheme for VQ-compressed images, which allows the original cover image to be completely recovered after the extraction of the secret bits. The embedded payload in the proposed method comprises the secret bits plus the restoration information. In order to reduce the size of payload, we utilized the spatial correlations in the image as the restoration information and then compressed the correlations by a lossless compression method. In addition, an alternative pairing method for codewords was proposed to improve the stegoed image quality and control the embedding capacity. Experimental results showed that the proposed method has the benefit of high efficiency of the steganographic process, high image quality, and adaptive embedding capacity compared with other schemes.

In this paper, we propose an accurate and scalable S-parameter de-embedding method for RF/microwave on-wafer characterization of silicon MOSFETs. Based on cascade configurations, this method utilizes planar open, short, and thru standards to estimate the effects of surrounding parasitic networks on a MOS transistor. The bulk-shielded open and short standards are used to simulate and de-embed the probe-pad parasitics. The thru standard are used to extract the interconnect parameters for subtracting the interconnect parasitics in gate and drain terminals of the MOSFET. To further eliminate the parasitics of dangling leg in source terminal of the MOSFET, we also introduce the microwave and multi-port network analysis to accomplish the two-port-to-three-port transformation for S-parameters. The MOSFET and its corresponding de-embedding standards were fabricated in a standard CMOS process and characterized up to 40 GHz. The scalability of the open, short, and thru standards is demonstrated and the performance of the proposed de-embedding procedure is validated by comparison with several de-embedding techniques.

A de-embedding technique for the measurement of very small parasitic capacitances of package or small module interconnects is presented. At high frequencies small parasitic capacitances become important, and measurement probes can strongly affect measurement results. The present technique is based on additional measurements with only one tip of the probe touching one conductor, while the second tip is kept floating on the substrate. A necessary condition for its application is that the measured capacitance does not depend on the position of the floating probe tip. Measurements with inverted probe tip polarities are also used. In this way, the capacitances between probe tips and DUT can be estimated together with the parasitic capacitances of interest. Depending on the required accuracy, de-embedding of different orders have been introduced, which consider capacitance configurations of increasing complexity. The technique requires the solution of one or more systems of non-linear equations. In the present example the minimization of the norm of the residual of the system has been treated as a least squares problem, and has been solved numerically with MATLAB. The accuracy of the measurement can be also approximately estimated with the residual. As application example, a small module with power and ground planes has been considered. Two different probes have been used. Even though the stray capacitances of the probes are very different, the values of the extracted parasitic capacitances are in agreement with each other. The accuracy has been verified also with simulation results. To this purpose, a combination of known formulas from the literature, a 2D Finite Element Method (FEM) tool and a 3D Boundary Element Method (BEM) tool have been used. A high accuracy can be obtained, even when a strong capacitive coupling between probe ground and DUT is present. The technique can be applied also when only a subset of measurement results are available.

A lossless data embedding method that inserts data in images in the spatial domain is proposed in this paper. Though a lossless data embedding method once distorts an original image to embed data into the image, the method restores the original image as well as extracts hidden data from the image in which the data are embedded. To guarantee the losslessness of data embedding, all pixel values after embedding must be in the dynamic range of pixels. Because the proposed method modifies some pixels to embed data and leaves other pixels as their original values in the spatial domain, it can easily keep all pixel values after embedding in the dynamic range of pixels. Thus, both the capacity and the image quality of generated images are simultaneously improved. Moreover, the proposed method uses only one parameter based on the statistics of pixel blocks to embed and extract data. By using this parameter, this method does not require any reference images to extract embedded data nor any memorization of the positions of pixels in which data are hidden to extract embedded data. In addition, the proposed method can control the capacity for hidden data and the quality of images conveying hidden data by controlling the only one parameter. Simulation results show the effectiveness of the proposed method; in particular, it offers images with superior image quality to conventional methods.