Recognition of human actions using CNN-GWO: a novel modeling of CNN for enhancement of classification performance

Recognizing human actions from unconstrained videos turns to be a major challenging task in computer visualization approaches due to decreased accuracy in the feature classification performance. Therefore to improve the classification performance it is essential to minimize the ‘classification’ errors. Here, in this work, we propose a hybrid CNN-GWO approach for the recognition of human actions from the unconstrained videos. The weight initializations for the proposed deep Convolutional Neural Network (CNN) classifiers highly depend on the generated solutions of GWO (Grey Wolf Optimization) algorithm, which in turn minimizes the ‘classification’ errors. The action bank and local spatio-temporal features are generated for a video and fed into the ‘CNN’ classifiers. The ‘CNN’ classifiers are trained by a gradient descent algorithm to detect a ‘local minimum’ during the fitness computation of GWO ‘search agents’. The GWO algorithms ‘global search’ capability as well as the gradient descent algorithms ‘local search’ capabilities are subjected for the identification of a solution which is nearer to the global optimum. Finally, the classification performance can be further enhanced by fusing the classifiers evidences produced by the GWO algorithm. The proposed classification frameworks efficiency for the recognition of human actions is evaluated with the help of four achievable action recognition datasets namely HMDB51, UCF50, Olympic Sports and Virat Release 2.0. The experimental validation of our proposed approach shows better achievable results on the recognition of human actions with 99.9% recognition accuracy.     

Event Pattern Analysis and Prediction at Sentence Level using Neuro-Fuzzy Model for Crime Event Detection

Classifying the sentences that describe Events is an important task for many applications. In this paper, Event patterns are identified and extracted at sentence level using term features. The terms that trigger Events along with the sentences are extracted from web documents. The sentence structures are analyzed using POS tags. A hierarchal sentence classification model is proposed by considering specific term features of the sentence and the rules are derived. The rules fail to define a clear boundary between the patterns and create ambiguity and impreciseness. To overcome this, suitable fuzzy rules are derived which gives importance to all term features of the sentence. The fuzzy rules are constructed with more variables and generate sixteen patterns. Artificial Neuro-Fuzzy Inference System (ANFIS) model is proposed for training and classifying the sentence patterns for capturing the knowledge present in sentences. The obtained patterns are assigned linguistic grades based on previous classification knowledge. These grades represent the type and quality of information in the patterns. The membership function is used to evaluate the fuzzy rules. The patterns share the membership values between [0–1] which determines the weights for each pattern. Later, higher weighted patterns are considered to build Event Corpus, which helps in retrieving useful and interested information of Event Instances. The performance of the proposed approach classification is evaluated for ‘Crime’ Event by crawling documents from WWW and also evaluated for benchmark dataset for ‘Die’ Event. It is found that the performance of the proposed approach is encouraging when compared with recently proposed similar approaches.     

Modeling and Design of High Performance Converters for Optimal Utilization of Interconnected Renewable Energy Resources to Micro Grid with GOLRS Controller

This paper proposes a bidirectional Z-source dc-dc converter topology for the optimal utilization of renewable energy sources to the microgrid with the proposed control technique. Compared to the existing dc-dc converter circuits, they can reduce in-rush and harmonic current, provide larger range of output dc voltage and improve reliability. It can operate in voltage-fed and current-fed when the place of the source and load is exchanged each other, and it can be perform buck-boost function in these two conditions. Its power flow can be bidirectional. The bidirectional Z-source dc-dc converter is revealed with the consideration of enhanced converter efficiency, effective utilization of renewable energy sources and reduced switching losses. The proposed control technique is the combination of both the grasshopper optimization algorithm (GOA) and local random search (LRS) and hence it is known as GOLRS controller. Here, the searching behaviour of the GOA is enhanced by LRS with the help of two operators named as crossover and mutation. In the proposed technique, the GOLRS is used to generate the optimal gain dataset based on the minimum error objective function and select the exact gain parameter of the PI controller. Batteries are used as an energy source to balance out and allow the renewable power system units to continue running at a steady and stable output power. The proposed technique is executed in the MATLAB/Simulink working platform and compared with various existing techniques.

     

Automatic segmentation of optic disc in retinal fundus images using semi-supervised deep learning

Diseases of the eye require manual segmentation and examination of the optic disc by ophthalmologists. Though, image segmentation using deep learning techniques is achieving remarkable results, it leverages on large-scale labeled datasets. But, in the field of medical imaging, it is challenging to acquire large labeled datasets. Hence, this article proposes a novel deep learning model to automatically segment the optic disc in retinal fundus images by using the concepts of semi-supervised learning and transfer learning. Initially, a convolutional autoencoder (CAE) is trained to automatically learn features from a large number of unlabeled fundus images available from the Kaggle’s diabetic retinopathy (DR) dataset. The autoencoder (AE) learns the features from the unlabeled images by reconstructing the input images and becomes a pre-trained network (model). After this, the pre-trained autoencoder network is converted into a segmentation network. Later, using transfer learning, the segmentation network is trained with retinal fundus images along with their corresponding optic disc ground truth images from the DRISHTI GS1 and RIM-ONE datasets. The trained segmentation network is then tested on retinal fundus images from the test set of DRISHTI GS1 and RIM-ONE datasets. The experimental results show that the proposed method performs on par with the state-of-the-art methods achieving a 0.967 and 0.902 dice score coefficient on the test set of the DRISHTI GS1 and RIM-ONE datasets respectively. The proposed method also shows that transfer learning and semi-supervised learning overcomes the barrier imposed by the large labeled dataset. The proposed segmentation model can be used in automatic retinal image processing systems for diagnosing diseases of the eye.

     

Robust H∞ performance for discrete time T-S fuzzy switched memristive stochasticneural networks with mixed time-varying delays

ABSTRACT

In this paper, we study the robust H _∞ performance for discrete-time T-S fuzzy switched memristive stochastic neural networks with mixed time-varying delays and switching signal design. The neural network under consideration is subject to time-varying and norm bounded parameter uncertainties. Decomposing of the delay interval approach is employed in both the discrete delays and distributed delays. By constructing a proper Lyapunov-Krasovskii functional (LKF) with triple summation terms and using an improved summation inequality techniques. Sufficient conditions are derived in terms of linear matrix inequalities (LMIs) to guarantee the considered discrete-time neural networks to be exponentially stable. Finally, numerical examples with simulation results are given to illustrate the effectiveness of the developed theoretical results.     

Pulse electrodeposited copper indium sulfide films

Copper indium sulfide (CISu) films were deposited by the pulse galvanostatic deposition technique at different duty cycles. The films are polycrystalline with peaks corresponding to the chalcopyrite phase of CISu. The grain size and surface roughness increased from 10 to 25 nm and 0.85 to 2.50 nm respectively with increase of duty cycle. Optical band gap in the range of 1.30–1.51 eV was observed for the films deposited at different duty cycles. Room temperature resistivity of the films is in the range of 0.1–3.67 Ω cm. Photoconductivity measurements were made at room temperature. Photocurrent spectra exhibited maximum corresponding to the band gap of copper indium sulphide. CdS/CuInS₂ fabricated with CISu films deposited at 50% duty cycle have exhibited a V_oc of 0.62 V, J_sc of 16.30 mA cm^?2, FF of 0.71 and efficiency of 7.16%.     

Total phenolic content,antioxidant and antidiabetic properties of methanolic extract of raw and traditionally processed Kenyan indigenous food ingredients

Certain indigenous foods commonly consumed by Kenyan vulnerable groups (the malnourished; children under 5 years of age; pregnant and lactating women; malnourished adults in cases of vitamin or mineral deficiencies, TB, diabetes, cancer, AIDS; refugees; orphans the elderly and the disabled) are not yet evaluated for phenolic content and health relevant functionality. The present study was therefore designed to analyze the phenolic content, antioxidant and antidiabetic properties of methanolic extract of raw and traditionally processed food ingredients. The total phenolic contents of the cereals, legumes, oil seeds and vegetables were ranged from 0.41 to 3.00 g/100 g DM. Amaranth grain (Amaranthus cruentus) and drumstick leaves (Moringa oleifera) exhibited significantly higher phenolic content than the other samples. The methanolic extract of the investigated samples showed promising levels of DPPH radical scavenging activity (81–89%); ferric reducing/antioxidant power (FRAP, 44–744 mmolL^?1 Fe[II]/g extract DM); α-amylase (10–45%) and α-glucosidase (13–80%) inhibition activities. The food ingredients with high phenolic content exhibited relatively higher antioxidant and antidiabetic activities. The results indicate that soaking + cooking is the mild processing method to preserve the phenolic compounds and their health relevant functionality in the presently investigated cereal, legume and oil grains, while cooking is suitable treatment for vegetables.     

Antioxidant Potential and Type II Diabetes-Related Enzyme Inhibition of Cassia obtusifolia L.: Effect of Indigenous Processing Methods

The methanolic extract of Cassia obtusifolia L. (Sicklepod) seed, an underutilized food legume from India, was analyzed for antioxidant and health relevant functionality. The total free phenolic content of the raw seeds was 13.33?±?1.73?g catechin equivalent/100?g extract. The extract exhibited 1,292?mmol Fe[II] per milligram extract of ferric reducing/antioxidant power, 49.92% inhibition of ?-carotene degradation, 65.79% of scavenging activity against DPPH, and 50.78% of superoxide radicals. The in vitro starch digestion bioassay of the extract showed 79.80% of ??-amylase and 81.04% of ??-glucosidase enzyme inhibition characteristics. Sprouting?+?oil frying caused an apparent increase on the total free phenolic content with significant improvement on the antioxidant and free radical scavenging capacity of C. obtusifolia seeds, while soaking?+?cooking as well as open-pan roasting treatments show diminishing effects. Inhibition of ??-amylase and ??-glucosidase enzyme activity was 23.81% and 42.36%, respectively, following sprouting?+?oil-frying treatment. These enzyme inhibition values were similar to that of synthetic antidiabetic agent acarbose.     

Exceptional photocatalytic activity for g-C3N4 activated by H2O2 and integrated with Bi2S3 and Fe3O4 nanoparticles for removal of organic and inorganic pollutants

Herein, hydrogen peroxide activated graphitic carbon nitride (agCN) was combined with Fe₃O₄ and Bi₂S₃ to fabricate agCN/Fe₃O₄/Bi₂S₃ nanocomposites via facile refluxing method, as visible-light-induced photocatalysts for photodegradations of anionic and cationic dyes such as MO, RhB, MB, and photoreduction of Cr(VI). The fabricated samples were explored by XRD, EDX, XPS, TGA, SEM, TEM, HRTEM, VSM, PL, FT-IR, BET, and UV-vis DRS. Photocatalytic activity of the nanocomposite with 20% of Bi₂S₃ was 16.6, 40.4, 19.5, and 12.5 times more than that of the pristine gCN in removal of RhB, MB, MO, and Cr(VI), respectively. A plausible photocatalytic mechanism on the agCN/Fe₃O₄/Bi₂S₃ nanocomposites was proposed by construction of n-n heterojunction between gCN and Bi₂S₃. Also, stability of the magnetic hybrid was characterized through cyclic photocatalytic tests.     

Novel approach to control CdS morphology by simple microwave-solvothermal method

Nanosize CdS powders with different microstructures are prepared in different solvents by using rapid microwave irradiation. Effect of solvents and Cd²⁺ precursors are to be able to control the particle size, and microstructures of CdS have been investigated by X-ray diffraction and TEM. The different particle size and morphologies are observed using different Cd²⁺ precursors in different solvents. TEM micrographs clearly show multiarmed nanorods and spherical shape morphologies of CdS powders are obtained in polar solvent like water (H₂O), whereas non-polar polyol solvent like ethylene glycol (EG), prickle and cluster like morphologies of CdS are achieved with different Cd²⁺ precursors such as CdSO₄ and Cd (CH₃COO)₂. The spectroscopy studies of nanosize CdS are examined by photo-luminescence spectra. Band gap and the absorption co-efficient for nano CdS is also evaluated from optical absorption studies.