Leveraging Transfer Learning for Spatio-Temporal Human Activity Recognition from Video Sequences

Human Activity Recognition (HAR) is an active research area due to its applications in pervasive computing, human-computer interaction, artificial intelligence, health care, and social sciences. Moreover, dynamic environments and anthropometric differences between individuals make it harder to recognize actions. This study focused on human activity in video sequences acquired with an RGB camera because of its vast range of real-world applications. It uses two-stream ConvNet to extract spatial and temporal information and proposes a fine-tuned deep neural network. Moreover, the transfer learning paradigm is adopted to extract varied and fixed frames while reusing object identification information. Six state-of-the-art pre-trained models are exploited to find the best model for spatial feature extraction. For temporal sequence, this study uses dense optical flow following the two-stream ConvNet and Bidirectional Long Short Term Memory (BiLSTM) to capture long-term dependencies. Two state-of-the-art datasets, UCF101 and HMDB51, are used for evaluation purposes. In addition, seven state-of-the-art optimizers are used to fine-tune the proposed network parameters. Furthermore, this study utilizes an ensemble mechanism to aggregate spatial-temporal features using a four-stream Convolutional Neural Network (CNN), where two streams use RGB data. In contrast, the other uses optical flow images. Finally, the proposed ensemble approach using max hard voting outperforms state-of-the-art methods with 96.30% and 90.07% accuracies on the UCF101 and HMDB51 datasets.     

Type 2 Diabetes Risk Prediction Using Deep Convolutional Neural Network Based-Bayesian Optimization

Diabetes mellitus is a long-term condition characterized by hyperglycemia. It could lead to plenty of difficulties. According to rising morbidity in recent years, the world’s diabetic patients will exceed 642 million by 2040, implying that one out of every ten persons will be diabetic. There is no doubt that this startling figure requires immediate attention from industry and academia to promote innovation and growth in diabetes risk prediction to save individuals’ lives. Due to its rapid development, deep learning (DL) was used to predict numerous diseases. However, DL methods still suffer from their limited prediction performance due to the hyperparameters selection and parameters optimization. Therefore, the selection of hyper-parameters is critical in improving classification performance. This study presents Convolutional Neural Network (CNN) that has achieved remarkable results in many medical domains where the Bayesian optimization algorithm (BOA) has been employed for hyperparameters selection and parameters optimization. Two issues have been investigated and solved during the experiment to enhance the results. The first is the dataset class imbalance, which is solved using Synthetic Minority Oversampling Technique (SMOTE) technique. The second issue is the model's poor performance, which has been solved using the Bayesian optimization algorithm. The findings indicate that the Bayesian based-CNN model superbases all the state-of-the-art models in the literature with an accuracy of 89.36%, F1-score of 0.88.6, and Matthews Correlation Coefficient (MCC) of 0.88.6.     

Short-term and long-term memory analysis of learning using 2D and 3D educational contents

ABSTRACT

The effect of 2D and 3D educational content learning on memory has been studied using electroencephalography (EEG) brain signal. A hypothesis is set that the 3D materials are better than the 2D materials for learning and memory recall. To test the hypothesis, we proposed a classification system that will predict true or false recall for short-term memory (STM) and long-term memory (LTM) after learning by either 2D or 3D educational contents. For this purpose, EEG brain signals are recorded during learning and testing; the signals are then analysed in the time domain using different types of features in various frequency bands. The features are then fed into a support vector machine (SVM)-based classifier. The experimental results indicate that the learning and memory recall using 2D and 3D contents do not have significant differences for both the STM and the LTM.     

The production of a cellular graphene array by scanning probe lithography and its ability to store electrical charge

A graphene cellular array on an insulating SiO₂ layer was fabricated by scanning probe lithography. The graphene layer was oxidized by an electric field which was applied between the cantilever tip and Si substrate without any electrode directly connected to the graphene layer. When the bias voltage was applied on a cell of patterned graphene through the cantilever tip, charge was accumulated on the cell and preserved for a long time without decay. The accumulated charge and the surface potential were measured by an electrostatic force microscope. The charge retention was measured as a function of time, and the decay time constant was estimated to be ～70 min.     

Diagnostic Study of an Fe-Ni Alloy Plasma Generated by the Fundamental (1064 nm) and Second (532 nm) Harmonics of an Nd: YAG Laser

We studied the spatial evolution of the Fe-Ni plasma generated by the fundamental (1064 nm) and second (532 nm) harmonics of a Q-switched Nd: YAG laser. The experimentally observed line profiles of the neutral iron (Fe I) have been used to extract the plasma temperature (T e ) using the Boltzmann plot method, whereas the electron number density (N e ) has been deter- mined from the Stark broadening. In addition, we studied the spatial behavior of T e and N e with the variation of laser energy for iron plasma by placing the target material (iron-nickel alloy) in air at atmospheric pressure for both modes of the Nd: YAG laser.     

Boron nitride based polyaniline nanocomposite: Preparation,property, and application

In this study, we report first time the electrical properties and photocatalytic activity of HCl doped polyaniline (Pani) and Pani/boron nitride (Pani/BN) nanocomposite prepared by in situ polymerization of aniline using potassium persulfate (K₂S₂O₈) in the presence of hexagonal boron nitride (h‐BN). The prepared Pani and Pani/BN nanocomposite were characterized by Fourier transform infrared, X‐ray diffraction, Thermogravimetric analysis, Scanning electron microscope, and Transmission electron microscope. The stability of the Pani/BN nanocomposite in comparison of Pani in terms of the DC electrical conductivity retention was investigated under isothermal and cyclic aging conditions. The Pani/BN nanocomposite in terms of DC electrical conductivity was observed to be comparatively more thermally stable than Pani. The degradation of Methylene blue (MB) and Rhodamine B (RhB) under UV‐light irradiation were 50 and 56.4%, respectively, over Pani and 65.7 and 71.6%, respectively, over Pani/BN. The results indicated that the extent of degradation of MB and RhB was greater over nanocomposite material than Pani, which may result due to high electron–hole pairs charge separation under UV light. © 2016 Wiley Periodicals, Inc. J. Appl. Polym. Sci. 2016 , 133, 43989.     

Mathematical model for calculating the blowby rate

A semi-empirical model for the calculation of the flow rate of blowby gases from an I.C. engine has been developed. The effect of geometry of the engine was considered. This model was subsequently applied for the cases of three, four and five rings. The blowby flow rate was found to be a hyperbolic function of the number of rings. Calculated blowby flow rates for a model hypothetical engine were compared with experimentally measured values of a diesel engine reported in the literature and were found to be in agreement. The method also enables one to know the pressure variation in each dwell space between the rings with respect to crank angle. Also, a quick and fairly accurate method has been suggested for easy assesment of blowby rate.     

Heterogeneous crystallization and composition dependence of optical parameters in Sn–Sb–Bi–Se chalcogenides

Bulk samples of Sn₁₀Sb_20−x Bi _x Se₇₀ (0 ≤ x ≤ 8) chalcogenide alloys were prepared by the conventional melt quenching technique. Thin films were prepared on well-cleaned glass substrates by thermal evaporation technique. X-ray diffraction studies revealed that the alloys with x = 0 and 2 at.% of Bi were amorphous, whereas the alloys with x = 4, 6, 8 at.% were crystalline. The crystalline phases are identified as due to the formation of Bi₂Se₃ and BiSe₂ phases. The microstructural and differential scanning calorimetric studies show the presence of these phases. A simple, straight forward procedure suggested by Swanepoel has been used to calculate the optical parameters, refractive index, and extinction coefficient. The optical gap for all the samples has been obtained from the Tauc plots. The variation in optical parameters for different Bi concentration has been explained on the basis of presence of defect states and the change in stoichiometry with the change in Bi concentration.