Search

The Panchthar, a hilly district of Eastern Nepal, has several medicinal plants which are used by indegenous people for the treatments of different diseases. In the present work, phytochemical and biological activities of hexane, ethyl acetate, acetone, and methanol extracts of Litsea monopetala leaves from the Pachthar district were analyzed. The phytochemical screening demonstrated the presence of major secondary metabolites like flavonoids, alkaloids, terpenoids, polyphenols, glycosides, and proteins. The antibacterial activity was estimated by the disc diffusion method against Bacillus substilus and Escherichia coli. The result showed that ethyl acetate extract exhibit the highest zone of inhibition (ZOI) value of 18 mm/disc. Acetone extract also showed significant antibacterial activity at the concentration of 100 mg/mL. Antioxidant activity from DPPH radical scavenging assay displayed the inhibitory concentration (IC50) values of dried methanol and acetone extracts were 30.74 μg/mL and 202.86 μg/mL, respectively. The methanol extract has the highest total phenolic content of 360.14 mg GAE/g extract and the highest total flavonoid content of 27.13 mg QE/g extract. The present study revealed that Litsea monopetala plant from the Panchthar district is a source of several bioactive substances and can be used in several drugs formulation.

This paper presents a Mispronunciation Detection and Diagnosis (MDD) system based on a range of Automatic Speech Recognition (ASR) models and feature types. The goals of this research are to assess the ability of speech recognition systems to detect and diagnose the common pronunciation errors seen in non-native speakers (L2) of English and to assess the contribution of the information offered by Electromagnetic Articulography (EMA) data in improving the performance of such MDD systems. To evaluate the ability of the ASR systems to detect and diagnose pronunciation errors, the recognized sequence of phonemes generated by the ASR models were aligned with human-labeled phonetic transcripts as well as with the original phonetic prompts. This three-way alignment determined the MDD related metrics of the ASR system. System architectures included GMM-HMM, DNN, and RNN based ASR engines for the MDD system. Articulatory features derived from the Electromagnetic Articulography corpus of Mandarin-Accented English (EMA-MAE) were utilized along with acoustic features to compare the performance of MDD systems. The best performing system using a combination of acoustic and articulatory features had an accuracy of 82.4%, diagnostic accuracy of 75.8% and a false rejection rate of 17.2%.

We address the task of image classification, when the available spectral bands can vary from image to image. We propose a model that learns to represent uncertainty over latent features in a way that is conditioned on the available bands. We expect that images with fewer bands will generally be more difficult to classify and hence have higher uncertainty. We compare two strategies for training such a model, one which uses explicit hierarchical constraints and one which relies on implicit constraints. We evaluate both using RGB and multispectral imagery from the EuroSat dataset and find that the hierarchical approach improves the compatibility of the resulting distributions without sacrificing accuracy.

Cardiovascular system study using Electrocardiogram (ECG) signals have evolved tremendously in the research domain of medical electronics and signal processing community. The extraction of characteristic points of an ECG signal helps in detection of any irregularities if present. This paper attempts to provide the basic understanding of ECG signal, pre-processing and extraction of fiducial points P, Q, R, S, T and their respective amplitude. Continuous Wavelet Transform (CWT) has been used for the localization of R peaks. Finally, other peaks are localized in time plane, using inter-beat interval dependent search windows. The proposed method was tested using ECG signal data obtained from MIT Physionet-ATM database and Signals derived from medical instrumentation lab of M.S. Ramaiah Institute of Technology. The simulation results were quite promising in terms of accurate localization of characteristic peaks, their respective amplitude in time plane and other features like R-R intervals, QRS width, S-T segment length, etc. The simplified procedure proposed in this study shows clear indication for arrhythmias detection which can be extended for automated ECG signal classification by appropriate selection of pattern classifiers.

This paper presents a new method for thermal characterization of THz Schottky diodes. The method is based on the transient current behavior, and it enables the extraction of thermal resistances, thermal time-constants, and peak junction temperatures of THz Schottky diodes. Many typical challenges in thermal characterization of small-area diode devices, particularly those related to self-heating and electrical transients, are either avoided or mitigated. The method is validated with measurements of commercially available single-anode Schottky varactor diodes. A verification routine is performed to ensure the accuracy of the measurement setup, and the characterization results are compared against an in-house measurement-based method and against simulation results of two commercial 3-D thermal simulators. For example, characterization result for the total thermal resistance of a Schottky diode with an anode area of 9 μm2 is within 10% of the average value of 4020 K/W when using all four approaches. The new method can be used to measure small diode devices with thermal time constants down to about 300 ns with the measurement setup described in the paper.

This paper presents the results from evaluation of suitable materials and ink properties for printing millimeter wave components. Dielectric parameters (relative permittivity and loss tangent) of various potential substrate materials are extracted from S-parameter measurements and simulations. Test samples have been printed over Polyethylene Terephthalate (PET) and Polyethylene Naphthalate (PEN) substrates and the measurement of the ink properties have been carried out.

This paper presents the characterization measurements and simulations performed with ACST InGaAs low-barrier Schottky diodes for millimeter wave mixing applications. While these low-barrier Schottky diodes have been successfully used for millimeter wave detector design, their performance when acting as mixers has not been tested yet. This paper shows the performance for this type of diodes in a fundamental mixing configuration with evaluation of conversion loss and noise temperature. A fundamental mixer test-jig is used as the measurement platform with low-loss E-H impedance tuners for RF and LO signal matching. In addition to the measurements, 3D HFSS and ADS circuit simulations are also performed and results are presented. Conversion loss of less than 5 dB is obtained for 0.1 mW LO power in mixer operation at 181-183 GHz.

In this work, we investigate different printing technologies suitable for mass production of millimeter-wave antennas and other devices, e.g., holograms and frequency selective absorbers, on flexible substrates. We concentrate especially on roll-to-roll reverse offset printing. The driving factors are low cost, high accuracy, high efficiency, and reliable performance. Therefore, we need to find and characterize suitable flexible substrates (permittivity and loss tangent at mm-wavelengths), conducting inks (viscosity, surface resistance of the resulting conducting layer), adhesion of the ink to the substrate, and feature size capable for printing mm-wave antennas and other passive devices in high volumes.