A new statistic has been developed to quantify the amount of regularity in data. This statistic, ApEn (approximate entropy), appears to have potential application throughout medicine, notably in electrocardiogram and related heart rate data analyses and in the analysis of endocrine hormone release pulsatility. The focus of this article is ApEn. We commence with a simple example of what we are trying to discern. We then discuss exact regularity statistics and practical difficulties of using them in data analysis. The mathematic formula development for ApEn concludes the Solution section. We next discuss the two key input requirements, followed by an account of a pilot study successfully applying ApEn to neonatal heart rate analysis. We conclude with the important topic of ApEn as a relative (not absolute) measure, potential applications, and some caveats about appropriate usage of ApEn. Appendix A provides example ApEn and entropy computations to develop intuition about these measures. Appendix B contains a Fortran program for computing ApEn. This article can be read from at least three viewpoints. The practitioner who wishes to use a "black box" to measure regularity should concentrate on the exact formula, choices for the two input variables, potential applications, and caveats about appropriate usage. The physician who wishes to apply ApEn to heart rate analysis should particularly note the pilot study discussion. The more mathematically inclined reader will benefit from discussions of the relative (comparative) property of ApEn and from Appendix A.

A regularity statistic for medical data analysis

The Karhunen-Loeve transform for a class of signals is proven to be a set of periodic sine functions and this Karhunen-Loeve series expansion can be obtained via an FFT algorithm. This fast algorithm obtained could be useful in data compression and other mean-square signal processing applications.

A fast Karhunen-Loeve transform for a class of random processes

As one of the most important directions in the field of computer vision, facial emotion recognition plays an important role in people's daily work and life. Human emotion recognition based on facial expressions is of great significance in the application of intelligent human-computer interaction. However, in the current research on facial emotion recognition, there are some problems such as poor generalization ability of network model and low robustness of recognition system. In this content, we propose a method of feature extraction using the deep residual network ResNet-50, which combines convolutional neural network for facial emotion recognition. Through the experimental simulation of the specified data set, it can be proved that this model is superior to the current mainstream facial emotion recognition models in the performance of facial emotion detection.

Facial expression recognition via ResNet-50

Device free human gesture recognition using Wi-Fi CSI: A survey

This article presents a new convolutional neural network (CNN) topology using wavelet kernels to detect and discriminate single or multiple partial discharge (PD) locations in high voltage power apparatus with increased accuracy. The method is tested on an electrical equipment model with acoustic PD sensors. A cubical tank has been emulated in the laboratory representing the equipment under test and partial discharge sources have been placed at different positions along with the required data acquisition hardware. The present scheme eliminates the requirement of separate algorithms for feature extraction and classification of the acquired PD signals. Wavelet kernels of the CNN play a crucial role in feature learning, and the proposed CNN architecture as a whole, can classify the features in a supervised manner. The performance of the proposed scheme is compared with other existing methods using the same data set . It is found that an overall accuracy of 97.64% is achieved by the proposed method, outperforming other existing methods by a significant margin of at least 5% in terms of accuracy. The developed module is a generic one and can be adapted for different high voltage electrical apparatus with similar topological structures; hence, it can be used in various ways in power industry.

Wavelet Kernel-Based Convolutional Neural Network for Localization of Partial Discharge Sources Within a Power Apparatus

This letter proposes an automated detection and classification of arrhythmia from the electrocardiogram (ECG) signals to employ deep learning (DL) framework based on long short-term memory (LSTM) network. Instead of using the classical LSTM network, a feature-based bidirectional LSTM (bi-LSTM) is employed, where a unidirectionally processed multifractal detrended fluctuation analysis is used to extract suitable features. The online available ECG signals are examined using multifractal parameters to study its nonlinear, stochastic, and complex fluctuations. A feature set comprising of ten features has been extracted from the segmented ECG beats followed by feeding to a single layer bi-LSTM network. Experimental results reveal that the feature-based bi-LSTM network outperforms the state-of-the-art DL methods compared on the same dataset. The proposed algorithm is a generic one and can be used for any computer-aided diagnosis of cardiovascular diseases.

Automated Detection and Classification of Arrhythmia From ECG Signals Using Feature-Induced Long Short-Term Memory Network

This paper presents an efficient technique for early diagnosis of simultaneous faults in different phases of stator winding of a three-phase induction motor due to turn-to-turn short circuit. A rea...

Identification and Classification of Stator Inter-Turn Faults in Induction Motor Using Wavelet Kernel Based Convolutional Neural Network

Hazy images suffer from poor visibility and possess low-contrast, degrading the visibility of the scene. The performance of single-image haze removal methods is limited by priors or constraints. This paper presents an efficient single image dehazing method by cascading two models. The first model is a new atmospheric scattering model from which atmospheric light and transmission map are estimated and the second model is the proposed sparse haze model from which a haze map is estimated. A least-square optimization is carried out on the haze map to restore the haze-free image. Moreover, the analysis of the haze removal process has been investigated by a membership function to observe how much haze has been removed from the input hazy images containing various degrees of haze. Dehazing performances are evaluated on three types of datasets, i.e., real-world hazy images, synthetic images of various degrees of haze, and our developed dataset containing thick haze. Experimental results demonstrate that the proposed approach generates better performance than the state-of-the-art methods, especially in the sky or objects containing white objects, both qualitatively and quantitatively.

Single Image Haze Removal with Haze Map Optimization for Various Haze Concentrations

Eye melanoma is a rare disease but according to malignancy, it is the most common type of cancer. Just like other types of cancers, it is curable for most of the cases if diagnosed properly but the process of diagnosis is quite challenging and is the most problematic issue in the treatment of eye melanoma. This paper presents an automated eye melanoma detection method using a convolutional neural network (CNN). 170 pre-diagnosed samples are taken from a standard database followed by pre-processing to lower resolution samples and finally fed to the CNN architecture. The proposed work eliminates separate feature extraction as well as the classification for the detection of eye melanoma. Although the proposed method requires a huge computation, a high accuracy rate of 91.76% is achieved outperforming the eye melanoma detection using an artificial neural network (ANN).

Biswarup Ganguly

Papers

Wavelet Kernel-Based Convolutional Neural Network for Localization of Partial Discharge Sources Within a Power Apparatus

Automated Detection and Classification of Arrhythmia From ECG Signals Using Feature-Induced Long Short-Term Memory Network

Identification and Classification of Stator Inter-Turn Faults in Induction Motor Using Wavelet Kernel Based Convolutional Neural Network

Single Image Haze Removal with Haze Map Optimization for Various Haze Concentrations

A Deep learning Framework for Eye Melanoma Detection employing Convolutional Neural Network