Lightweight deep convolutional neural network for background sound classification in speech signals.

Recognizing background information in human speech signals is a task that is extremely useful in a wide range of practical applications, and many articles on background sound classification have been published. It has not, however, been addressed with background embedded in real-world human speech signals. Thus, this work proposes a lightweight deep convolutional neural network (CNN) in conjunction with spectrograms for an efficient background sound classification with practical human speech signals. The proposed model classifies 11 different background sounds such as airplane, airport, babble, car, drone, exhibition, helicopter, restaurant, station, street, and train sounds embedded in human speech signals. The proposed deep CNN model consists of four convolution layers, four max-pooling layers, and one fully connected layer. The model is tested on human speech signals with varying signal-to-noise ratios (SNRs). Based on the results, the proposed deep CNN model utilizing spectrograms achieves an overall background sound classification accuracy of 95.2% using the human speech signals with a wide range of SNRs. It is also observed that the proposed model outperforms the benchmark models in terms of both accuracy and inference time when evaluated on edge computing devices.

Plasmonic metamaterial-based filtering structures with dynamic tunability.

In this Letter, we report the design, analysis, and characterization of first- and second-order plasmonic metamaterial-based multi-mode filtering structures. Further, electronic adaptivity in filter transfer functions is introduced and characterized. First, the basic operating principle of the engineered multi-mode resonator-based bandpass filter is presented. Then the concept is extended by introducing electronic (dynamic) tuning of the bandwidth using semiconductor varactor diodes. Afterwards, to enhance the selectivity and out-of-band filtering response, second-order multi-mode designs are realized. For experimental verification, the hardware prototype is fabricated and characterized using the Keysight analyzer N9918A. The design filtering structures will pave an important role in tunable plasmonic circuits and systems.

Towards Development of a Non-Intrusive and Label-Free THz Sensor for Rapid Detection of Aqueous Bio-Samples Using Microfluidic Approach.

As most of the bio-molecules sizes are comparable to the terahertz (THz) wavelength, this frequency range has spurred great attention for bio-medical and bio-sensing applications. Utilizing such capabilities of THz electromagnetic wave, this paper presents the design and analysis of a new non-intrusive and label-free THz bio-sensor for aqueous bio-samples using the microfluidic approach with real-time monitoring. The proposed THz sensor unit utilizes the highly confined feature of the localized spoof surface plasmon (LSSP) resonator to get high sensitivity for any minute change in the dielectric value near it's surface. The proposed sensor, which is designed at 1 THz, exploits the reflection behavior (S11) of the LSSP resonator as the sensing response. The proposed sensor has been designed with a high-quality factor of 192 to obtain a high sensitivity of 13.5 MHz/mgml-1. To validate the proposed concept, a similar sensor unit has been designed and implemented at microwave frequency owing to the geometry dependent characteristics of the LSSP. The developed sensor has got a highly sensitive response at microwave frequency with a sensitivity of 1.2771e-4 MHz/mgml-1. A customized read-out circuitry is also designed and developed to get the sensor response in terms of DC-voltage and to provide a proof of concept for the low-cost point of care (PoC) detection solution using the proposed sensor. It is anticipated that the proposed design of highly sensitive sensor will pave a path to develop lab-on-chip systems for bio-sensing applications.

Non-psychiatric Referral among First Encounter Patients Attending the Psychiatry Out Patient Services of a Tertiary Care Hospital: A Descriptive Cross-sectional Study.

INTRODUCTION: Pathway of psychiatric care is defined as the sequence of contacts with individuals and organizations initiated by the distressed person's efforts and his significant others to seek appropriate health care. This study aimed to find the prevalence of non-psychiatric referral as first encounter among patients attending the psychiatry outpatient department of a tertiary care hospital. METHODS: A descriptive cross-sectional study was carried out from 29th March 2015 to 25th April 2015 in the outpatient department of the department of psychiatry of a tertiary via direct interview using the World Health Organization's encounter form. Ethical approval was taken from undergraduate medical research protocol review board (Reference number 105/071/072). Psychiatric diagnoses were made by respective consultants using the International Classification of Diseases-10 Clinical Descriptions and Diagnostic Guidelines criteria. Data was entered in the Microsoft Excel 2007 and analyzed by Stata version 15. Point estimate at 95% Confidence Interval was calculated along with frequency and percentage for binary data. RESULTS: Out of 50 patients, 26 (52%) (38.2-65.8 at 95% Confidence Interval) of new cases in the outpatient department had non-psychiatric referrals. Among them, 13 (26%) referred from faith healers, 7 (14%) from the general hospital and 6 (12%) from medical out patient department. CONCLUSIONS: The prevalence of non-psychiatric referral for the patients seen for the first time in the psychiatry outpatient department was similar to findings from studies done in different parts of South East Asia.