Minimization of errors in discrete wavelet filtering of signals during ultrasonic measurements and testing

Error minimizing methods for discrete wavelet filtering of ultrasonic meter signals are considered. For this purpose, special model signals containing various measuring pulses are generated. The psi function of the Daubechies 28 wavelet is used to generate the pulses. Noise is added to the generated pulses. A comparative analysis of the two filtering algorithms is performed. The first algorithm is to limit the amount of detail of the wavelet decomposition coefficients in relation to signal interference. The minimum value of the root mean square error of wavelet decomposition signal deviation which is restored at each level from the initial signal without noise is determined. The second algorithm uses a separate threshold for each level of wavelet decomposition to limit the magnitude of the detail coefficients that are proportional to the standard deviation. Like in the first algorithm, the task is to determine the level of wavelet decomposition at which the minimum standard error is achieved. A feature of both algorithms is an expanded base of discrete wavelets ‒ families of Biorthogonal, Coiflet, Daubechies, Discrete Meyer, Haar, Reverse Biorthogonal, Symlets (106 in total) and threshold functions garotte, garrote, greater, hard, less, soft (6 in total). The model function uses random variables in both algorithms, so the averaging base is used to obtain stable results. Given features of algorithm construction allowed to reveal efficiency of ultrasonic signal filtering on the first algorithm presented in the form of oscilloscopic images. The use of a separate threshold for limiting the number of detail coefficients for each level of discrete wavelet decomposition using the given wavelet base and threshold functions has reduced the filtering error.

Multidimensional periodic discrete wavelets

In this paper, the definition of a periodic discrete multiresolution analysis is provided. The characterization of such systems is obtained in terms of properties of scaling sequences. Wavelet systems generated by such multiresolution analyses are defined and described. Decomposition and reconstruction formulas for the associated discrete wavelet transform are provided.

Robust deep convolutional neural network against image distortions

Many approaches have been proposed in the literature to enhance the robustness of Convolutional Neural Network (CNN)-based architectures against image distortions. Attempts to combat various types of distortions can be made by combining multiple expert networks, each trained by a certain type of distorted images, which however lead to a large model with high complexity. In this paper, we propose a CNN-based architecture with a pre-processing unit in which only undistorted data are used for training. The pre-processing unit employs discrete cosine transform (DCT) and discrete wavelets transform (DWT) to remove high-frequency components while capturing prominent high-frequency features in the undistorted data by means of random selection. We further utilize the singular value decomposition (SVD) to extract features before feeding the preprocessed data into the CNN for training. During testing, distorted images directly enter the CNN for classification without having to go through the hybrid module. Five different types of distortions are produced in the SVHN dataset and the CIFAR-10/100 datasets. Experimental results show that the proposed DCT-DWT-SVD module built upon the CNN architecture provides a classifier robust to input image distortions, outperforming the state-of-the-art approaches in terms of accuracy under different types of distortions.

Robust Speaker Diarization Based on Daubechies Wavelet, Nonlinear Energy Operator and Pyknogram

Two common disciplines of speech processing are speaker recognition “identification and verification of speaker”, and speaker diarization, “who spoke when”. Motivated by various applications in automatic speaker recognition, speaker indexing, word counting, and audio transcription, speaker diarization (SD) becomes a significant area of signal processing. The basic designing steps of SD are feature extraction, voice activity detection (VAD), segmentation, and clustering. VAD process is accomplished by Daubechies 40, discrete wavelets transform (DWT). Initially, DWT was used for compression, scaling, and denoising of audio-stream and then partitioned into small frames of size 0.12 seconds. Next, features of each frame were extracted by applying nonlinear energy operator (NEO) based pyknogram. To measure the similarity between frames, a sliding window on delta-BIC distance metric was applied. A negative value of its output represents the same segments and vice-versa. To improve the output of the segmentation process, resegmentation was applied by information change rate method. At last, hierarchical clustering groups the homogeneous segments that correspond to a particular speaker and has been graphically represented by the dendrogram. The performance of SD was evaluated by F-measure and speaker diarization error rate (SER) and their results were compared with the traditional speaker diarization system that uses MFCC and BIC for segmentation and clustering. It reveals a significant reduction of 12.3% of SER in the proposed diarization system.