A Hybrid Deep Learning Model for Brain Tumour Classification

A brain tumour is one of the major reasons for death in humans, and it is the tenth most common type of tumour that affects people of all ages. However, if detected early, it is one of the most treatable types of tumours. Brain tumours are classified using biopsy, which is not usually performed before definitive brain surgery. An image classification technique for tumour diseases is important for accelerating the treatment process and avoiding surgery and errors from manual diagnosis by radiologists. The advancement of technology and machine learning (ML) can assist radiologists in tumour diagnostics using magnetic resonance imaging (MRI) images without invasive procedures. This work introduced a new hybrid CNN-based architecture to classify three brain tumour types through MRI images. The method suggested in this paper uses hybrid deep learning classification based on CNN with two methods. The first method combines a pre-trained Google-Net model of the CNN algorithm for feature extraction with SVM for pattern classification. The second method integrates a finely tuned Google-Net with a soft-max classifier. The proposed approach was evaluated using MRI brain images that contain a total of 1426 glioma images, 708 meningioma images, 930 pituitary tumour images, and 396 normal brain images. The reported results showed that an accuracy of 93.1% was achieved from the finely tuned Google-Net model. However, the synergy of Google-Net as a feature extractor with an SVM classifier improved recognition accuracy to 98.1%.     

Robust Variable-Step Perturb-and-Observe Sliding Mode Controller for Grid-Connected Wind-Energy-Conversion Systems

In order to extract efficient power generation, a wind turbine (WT) system requires an accurate maximum power point tracking (MPPT) technique. Therefore, a novel robust variable-step perturb-and-observe (RVS-P&O) algorithm was developed for the machine-side converter (MSC). The control strategy was applied on a WT based permanent-magnet synchronous generator (PMSG) to overcome the downsides of the currently published P&O MPPT methods. Particularly, two main points were involved. Firstly, a systematic step-size selection on the basis of power and speed measurement normalization was proposed; secondly, to obtain acceptable robustness for high and long wind-speed variations, a new correction to calculate the power variation was carried out. The grid-side converter (GSC) was controlled using a second-order sliding mode controller (SOSMC) with an adaptive-gain super-twisting algorithm (STA) to realize the high-quality seamless setting of power injected into the grid, a satisfactory power factor correction, a high harmonic performance of the AC source, and removal of the chatter effect compared to the traditional first-order sliding mode controller (FOSMC). Simulation results showed the superiority of the suggested RVS-P&O over the competing based P&O techniques. The RVS-P&O offered the WT an efficiency of 99.35%, which was an increase of 3.82% over the variable-step P&O algorithm. Indeed, the settling time was remarkably enhanced; it was 0.00794 s, which was better than for LS-P&O (0.0841 s), SS-P&O (0.1617 s), and VS-P&O (0.2224 s). Therefore, in terms of energy efficiency, as well as transient and steady-state response performances under various operating conditions, the RVS-P&O algorithm could be an accurate candidate for MPP online operation tracking.     

Stability Enhancement of Anthocyanins from Blackcurrant (Ribes Nigrum L.) Pomace through Intermolecular Copigmentation

Intermolecular copigmentation denotes the interaction between colored anthocyanins and the colorless copigment, which is not bound covalently to the anthocyanin molecule. This is the first study to investigate the effect of intermolecular copigmentation on the stability of individual anthocyanins from dried blackcurrant pomace (DBP) using four pure phenolic acids as copigments (ferulic, caffeic, chlorogenic, and rosmarinic acid). Studies were performed at pH 3.0 and pH 6.0, with a copigment/anthocyanin extract molar ratio of 5:1, during storage at 20 °C. At both pH 3.0 and 6.0, rosmarinic acid showed the strongest hyperchromic and bathochromic effects (p < 0.05) on day 0. However, rosmarinic acid showed low stability during storage. At pH 3.0, chlorogenic acid and control samples were capable of maintaining very high levels of total anthocyanin stability during storage (p < 0.05). On the other hand, ferulic acid and control samples had the longest estimated half-life during storage at pH 6.0. Intermolecular copigmentation successfully increased the half-life, color retention, and antioxidant activity of the anthocyanin solution, with cyanidin-3-O-glucoside (C3G) exhibiting the highest stability at both pH values. Overall, anthocyanins from DBP, in combination with chlorogenic or ferulic acid, showed potential for use in commercial food applications.     

Anti-Disturbance Compensation-Based Nonlinear Control for a Class of MIMO Uncertain Nonlinear Systems

Multi-Inputs-Multi-Outputs (MIMO) systems are recognized mainly in industrial applications with both input and state couplings, and uncertainties. The essential principle to deal with such difficulties is to eliminate the input couplings, then estimate the remaining issues in real-time, followed by an elimination process from the input channels. These difficulties are resolved in this research paper, where a decentralized control scheme is suggested using an Improved Active Disturbance Rejection Control (IADRC) configuration. A theoretical analysis using a state-space eigenvalue test followed by numerical simulations on a general uncertain nonlinear highly coupled MIMO system validated the effectiveness of the proposed control scheme in controlling such MIMO systems. Time-domain comparisons with the Conventional Active Disturbance Rejection Control (CADRC)-based decentralizing control scheme are also included.     

In Vitro and In Silico Anti-Acetylcholinesterase Activity from Macaranga tanarius and Syzygium jambos

Macaranga tanarius (MT) and Syzygium jambos (SJ) are pharmacologically reported to have anti-oxidant, anti-inflammatory, and anti-diabetic effects, and can be neuroprotective agents. Our previous work revealed that MT and SJ exhibited 76.32% and 93.81% inhibition against acetylcholinesterase (AChE) at 50 μg/mL final concentration in their ethyl acetate and hexane fractions, respectively. This study was aimed to investigate the bioactive constituents of MT and SJ and their molecular mechanism toward AChE inhibition. Bioassay-guided isolation afforded prenylflavonoids 1–3 from MT and anacardic acid derivatives 4 and 5 from SJ that were confirmed by NMR and MS data. Compound 5 exerted the strongest anti-AChE potential (IC₅₀: 0.54 μM), followed by 1, 4, 3, and 2 (IC₅₀: 1.0, 2.4, 6.8, and 33 μM, respectively). In silico molecular docking revealed 5 formed stronger molecular interactions including three H-bonds than its derivative 4 based on the saturation of their alkyl chains. The addition of a five carbon-prenyl chain in 1 increased the number of binding interactions, justifying its greater activity than derivatives 2 and 3. This research reflects the first report of AChE inhibitors from these species, thereby adding pharmacological values to MT and SJ as potential remedies in neuroprotection.     

Separation and preconcentration of mercury in water samples by ionic liquid supported cloud point extraction and fluorimetric determination

We have developed a cloud point extraction procedure based on room temperature ionic liquid for the preconcentration and determination of mercury in water samples. Mercury ion was quantitatively extracted with tetraethyleneglycol-bis(3- methylimidazolium) diiodide in the form of its complex with 5,10,15,20-tetra-(4-phenoxyphenyl)porphyrin. The complex was back extracted from the room temperature ionic liquid phase into an aqueous media prior to its analysis by spectrofluorimetry. An overall preconcentration factor of 45 was accomplished upon preconcentration of a 20?mL sample. The limit of detection obtained under the optimal conditions is 0.08?μg mL^?1, and the relative standard deviation for 10 replicate assays (at 0.5?g mL^?1 of Hg) was 2.4%. The method was successfully applied to the determination of mercury in tap, river and mineral water samples.

Enhancement of Raman scattering by individual dielectric microspheres

Enhanced Raman scattering (ERS) from bulk Si and surface brilliant cresyl blue (Bcb) molecules was investigated by individual polystyrene (PS) microspheres placed on top of the samples. It is revealed that the Raman signal depends both on the microsphere size and the numerical aperture (NA) of the collection lens, and a maximum signal enhancement of ∼11‐ and 40‐fold for Si and Bcb is observed, respectively, showing strong ERS effect. The different ERS behavior was elucidated by electromagnetic simulations using the finite element method. The present work also provides information on individual dielectric sphere for applications in microscopy, spectroscopic imaging and improvement, etc. Copyright © 2010 John Wiley & Sons, Ltd.     

An investigation of decomposition stages of a ruthenium polypridyl complex by non-isothermal methods

Thermal properties of [cis-(dithiocyanato)(4,5-diazafluoren-9-one)(4,4??-dicarboxy-2,2??-bipyridyl)ruthenium(II)], [Ru(L ₁)(L ₂)(NCS)₂] (where the ligands L ₁?=?4,5-diazafluoren-9-one, L ₂?=?4,4??-dicarboxy-2,2??-bipyridyl) have been investigated by DTA/TG/DTG measurements under inert atmosphere in the temperature range of 30?C1155?°C. The mass spectroscopy technique has been used to identify the products during pyrolytic decomposition. The pyrolytic final products have been analyzed by X-ray powder diffraction technique. A decomposition mechanism has been also suggested for the cis-[Ru(L ₁)(L ₂)(NCS)₂] complex based on the results of thermogravimetrical and mass analysis. The values of the activation energy, E* have been obtained by using model-free Kissenger?CAkahira?CSunose and Flyn?CWall?COzawa non-isothermal methods for all decomposition stages. Thirteen kinetic model equations have been tested for selecting the best reaction models. The best model equations have been determined as A2, A3, D1, and D2 which correspond to nucleation and growth mechanism for A2 and A3 and diffusion mechanism for D1 and D2. The optimized average values of E* are 31.35, 58.48, 120.85, and 120.56?kJ?mol^?1 calculated by using the best model equations for four decomposition stages, respectively. Also, the average Arrhenius factor, A, has been obtained as 2.21, 2.61, 2.52, and 2.21?kJ?mol^?1 using the best model equation for four decomposition stages, respectively. The ??H*, ??S*, and ??G* functions have been calculated using the optimized values.     

Catalytic Transesterification of Coconut Oil in Biodiesel Production: A Review

Catalysis Surveys from Asia - Biodiesel is one of the renewable energy (RE) sources that has received much interest due to its promising properties. Recently, the use of coconut oil as biodiesel...     

Confocal white light reflection imaging for characterization of metal nanostructures

A simple far-field confocal white light reflection imaging system was proposed using a xenon (Xe) lamp as the incident light source. The best spatial resolution was determined to be about 410 nm. Localized surface plasmons (LSPs) of isolated single and dimer gold nanospheres were studied and the resonance energy difference between their LSPs was extracted. Polarization dependent reflection imaging and contrast spectrum of single silver nanowires were also investigated, which correlate closely with the polarization dependent excitation of their LSP.