Supervised Feature Learning for Offline Writer Identification Using VLAD and Double Power Normalization

As an indispensable part of identity authentication,offline writer identification plays a notable role in biology,forensics,and historical document analysis.However,identifying handwriting efficiently,stably,and quickly is still challenging due to the method of extracting and processing handwriting features.In this paper,we propose an efficient system to identify writers through handwritten images,which integrates local and global features from similar handwritten images.The local features are modeled by effective aggregate processing,and global features are extracted through transfer learning.Specifically,the proposed system employs a pre-trained Residual Network to mine the relationship between large image sets and specific handwritten images,while the vector of locally aggregated descriptors with double power normalization is employed in aggregating local and global features.Moreover,handwritten image segmentation,preprocessing,enhancement,optimization of neural network architecture,and normalization for local and global features are exploited,significantly improving system performance.The proposed system is evaluated on Computer Vision Lab(CVL)datasets and the International Conference on Document Analysis and Recognition(ICDAR)2013 datasets.The results show that it represents good generalizability and achieves state-of-the-art performance.Furthermore,the system performs better when training complete handwriting patches with the normalization method.The experimental result indicates that it’s significant to segment handwriting reasonably while dealing with handwriting overlap,which reduces visual burstiness.

A novel algorithm of adaptive IIR lattice notch filter and performance analysis

A novel adaptive algorithm of IIR lattice notch filter realized by all-pass filter is presented. The time-averaged estimation of cross correlation of the present instantaneous input signal and the past output signal is used to update the step-size, leading to a considerably improved convergence rate in a low SNR situation and reduced steady-state bias and MSE. The theoretical expression for steady-state bounds on the step-size is derived, and the influence factors on the stable performance of the algorithm theoretically are analyzed. A normalized power factor is then introduced to control variation of step-size in its steady-state bounds. This technique prevents divergence due to the influence of large power input signal and improves robustness. Numerical experiments are performed to demonstrate superiority of the proposed method.

Dark Current Compensation and Sensitivity Adjustment on Gallium Arsenide Linear Array Detector for X-Ray Imaging

For the last several years, the linear array x-ray detector for x-ray imaging with gallium arsenide direct conversion sensitive elements has been developed and tested at the In-stitute for High Energy Physics. The array consists of 16 sensitive modules. Each module has 128 gallium arsenide (GaAs) sensitive elements with 200 μm pitch. Current article describes two key program procedures of initial dark current compensation of each sensitive element in the linear array, and sensitivity adjustment for alignment of strip pattern in the raw image data. As a part of evaluation process a modular transfer function (MTF) was measured with the slanted sharp-edge object under RQA5 technique as it described in the International Electrotechnical Commission 62220-1 standard (high voltage 70 kVp, additional aluminium filter 21 mm) for images with compensated dark currents and adjusted sensitivity of detector elements. The 10% level of the calculated MTF function has spatial resolution within 2 - 3 pair of lines per mm for both vertical and horizontal orientation.

An Efficient Approach for Segmentation, Feature Extraction and Classification of Audio Signals

Due to the presence of non-stationarities and discontinuities in the audio signal, segmentation and classification of audio signal is a really challenging task. Automatic music classification and annotation is still considered as a challenging task due to the difficulty of extracting and selecting the optimal audio features. Hence, this paper proposes an efficient approach for segmentation, feature extraction and classification of audio signals. Enhanced Mel Frequency Cepstral Coefficient (EMFCC)-Enhanced Power Normalized Cepstral Coefficients (EPNCC) based feature extraction is applied for the extraction of features from the audio signal. Then, multi-level classification is done to classify the audio signal as a musical or non-musical signal. The proposed approach achieves better performance in terms of precision, Normalized Mutual Information (NMI), F-score and entropy. The PNN classifier shows high False Rejection Rate (FRR), False Acceptance Rate (FAR), Genuine Acceptance rate (GAR), sensitivity, specificity and accuracy with respect to the number of classes.

Anti-Jamming with Adaptive Arrays Utilizing Power Inversion Algorithm

The convergence rate of the power inversion （PI） algorithm is quite sensitive to the power of the interference with the used fixed parameters in the PI algorithm leading to degradation of its ability to handle interference. This paper presents a normalized PI algorithm that traces the stochastic characteristics of the interference. The algorithm adaptively adjusts the recursive step size to determine the constrained optimized parameters for the Iowpass filter. Simulations show that the normalized PI algorithm achieves faster convergence and produces deeper nulls. The algorithm makes GPS receivers more robust in environments with large variations in the interference strength.

A subband excitation substitute based scheme for narrowband speech watermarking

We propose a new narrowband speech watermarking scheme by replacing part of the speech with a scaled and spectrally shaped hidden signal. Theoretically, it is proved that if a small amount of host speech is modified, then not only an ideal channel model for hidden communication can be established, but also high imperceptibility and good intelligibility can be achieved. Furthermore, a practical system implementation is proposed. At the embedder, the power normalization criterion is first imposed on a passband watermark signal by forcing its power level to be the same as the original passband excitation of the cover speech, and a synthesis filter is then used to spectrally shape the scaled watermark signal. At the extractor, a bandpass filter is first used to get rid of the out-of-band signal, and an analysis filter is then employed to compensate for the distortion introduced by the synthesis filter. Experimental results show that the data rate is as high as 400 bits/s with better bandwidth efficiency, and good imperceptibility is achieved. Moreover, this method is robust against various attacks existing in real applications.