Search

Determination of breast tumors from B-Mode Ultrasound (US) image is a perplexing one. Researches employing statistical modeling such as Nakagami, Normal Inverse Gaussian (NIG) distributed parametric images in this classification task have already explored but experimentation of those statistical models on contourlet transformed coefficient image in breast tumor classification task has not reported yet. The proposed method is established by considering 250 clinical cases from a publicly available database. In this database each clinical case exists as *.bmp format. In the preprocessing step firstly, the ultrasound B-Mode image is binarized to detect the lesion contour. Then contourlet transformation is employed. These contourlet sub band coefficients are shown to be modeled effectively by Nakagami and NIG distributions. These Nakagami and NIG parametric images are obtained by estimating the parameters of those prior statistical distributions locally. Few shape and statistical features are chosen according to their effectiveness on those parametric images. The benign and malignant breast tumors are classified utilizing these features with different classifiers such as the support vector machine, k-nearest neighbors, fitted binary classification decision tree, binary Gaussian kernel classification model, linear classification models for binary learning with high-dimensional etc. It is observed that classification performance of NIG statistical model based parametric version of contourlet coefficient images gained better accuracy than those of Nakagami statistical model. GUB JOURNAL OF SCIENCE AND ENGINEERING, Vol 7, Dec 2020 P 21-26

Accurately segmenting foods from optical images is a challenging task, yet becoming possible with the help of recent advances in Deep Learning based solutions. Automated identification of food items opens up possibilities of useful applications like nutrition intake monitoring. Given large variations in food choices, Deep Learning based solutions still struggle to generate human level accuracy. In this work, we propose a novel Sequential Transfer Learning method using Hierarchical Clustering. This novel approach simulates a step by step problem solving framework based on clustering of similar types of foods. The proposed approach provides up to 6% gain in accuracy compared to traditional network training and generated a robust model performing better in challenging unseen cases. This approach is also tested for segmenting foods in Danish school children meals for dietary intake monitoring as an application.

Reliability and cost are two important driving factors in the development of wind energy. Automation and digitalization of operation and maintenance (O&M) procedures help to increase turbine reliability and reduce the levelized cost of energy (LCOE). Here, we demonstrate a novel method, coined as AQUADA, which may change the current labor-intensive and operation-interfering blade inspection by using thermography and computer vision. We experimentally show that structural damages below the surfaces can be detected and quantified remotely when wind turbine blades are subject to fatigue loads. The data acquisition and analysis are automatically done in one single step, which may shift the current inspection paradigm through more automated O&M procedures. The cost analysis shows that the AQUADA method has a potential to at least half the total inspection cost and reduce the LCOE by 1%–2% when applied to a baseline land-based wind farm consisting of twenty 2.45-MW turbines. All data and source codes are published for researchers to reproduce our results and facilitate further development of AQUADA towards more mature industrial applications.

Smart Grid technology as a platform can encompass several advanced technological features across the spectrum of the power system. Effective Smart Grid system includes smart metering, robust communication infrastructure, deep integration of renewable energy sources, and distributed generation, among other important factors. The presented work in this paper highlights the required adaptation of the potential prime features of Smart Grid technology covering niftier equipment integration and internal cross-platform communication adaptable to the context of Bangladesh. This paper also focuses on the potential cost model for Smart Grid installation in Bangladesh based on the current scenario regarding the power sector. A study on India’s Smart Grid that simulates a quasi-similar circumstance helped better comprehend the essential steps required to implement Smart Grids in Bangladesh.

Real time estimation of nutrition intake from regular food items using mobile-based applications could be a breakthrough in creating public awareness of threats in overeating or faulty food choices. The bottleneck in implementing such systems is to effectively estimate the depths of the food items which is essential to calculate the volumes of foods. Volumes and density of food items can be used to estimate the weights of food eaten and their corresponding nutrition contents. Without specific depth sensors, it is very difficult to estimate the depth of any object from a single camera. Such sensors are equipped only in very advanced and expensive mobile devices. This work investigates the possibilities of using regular cameras to calculate the same using a specific frame structure. We proposed a controlled camera setup to acquire overlapping images of the food from different positions already calibrated to estimate the depths. The results were compared with the Kinect device’s depth measures to show the efficiency of the proposed method. We further investigated the optimum number of camera positions, their corresponding angles, and distances from the object to propose the best configuration for such a controlled system of image acquisition with regular mobile cameras. Overall the proposed method presents a low-cost solution to the depth estimation problem and opens up the possibilities for mobile-based apps for dietary assessment for various health-related problem-solving. GUB JOURNAL OF SCIENCE AND ENGINEERING, Vol 6(1), Dec 2019 P 1-11

In vast Photovoltaic (PV) power plant the output power production decreases significantly due to the fact of non-uniform aging of PV modules. The non-uniform aging of PV modules increases current-voltage (I-V) mismatch among the array modules and causes mismatch power loss (MPL). There are different interconnection topologies of the PV module in an array to minimize MPL and thus maximize the array output power. This paper investigates four different interconnection topologies experimentally on a 4×4 nonuniformly aged PV array. Three different patterns of PV module rearrangement are used to investigate the performance of each interconnection topology in terms of array output power and MPL. The experimental results show that the proposed interconnection topology is yields about 3.28% (average) higher output power than that of the most commonly used series-parallel array topology. GUB JOURNAL OF SCIENCE AND ENGINEERING, Vol 6(1), Dec 2019 P 39-45

This paper presents a comparative analysis of the zero-voltage zero-current switching (ZVZCS) soft switching technique with zero-voltage switching (ZVS) and zero-current switching (ZCS) counterparts. The generalization of the voltage–current crossover or the energy loss factor obtained from simulation of the prototype converter shows that the ZVZCS significantly reduces the losses and helps to improve the efficiency of the converter as compared to the ZVS or the ZCS. On the other hand, it is also found that the soft switching range of operation of the ZVS and the ZCS is largely affected by the maximum switch voltage and switch current, respectively. In contrast, these factors have a negligible effect on the ZVZCS operation which results in an extended range of soft switching operation. Additionally, a detailed PSPICE simulation is performed for selected ZVS, ZCS, and ZVZCS topologies from the recent literature, and the switching losses in the main switches of the converters are measured. It is observed that the energy losses in the ZVZCS mode are reduced on average by approximately 26% at turn on and 20% at turn off as compared to the ZVS and the ZCS. Furthermore, the low standard deviation in this mode confirms a stable low-loss profile which renders an extended soft switching range. An experimental test is also conducted by building the prototype converter to verify the simulation results. It is found that the switching losses are minimum while the converter is operated in the ZVZCS mode. Additionally, the efficiency drop remains consistently low as compared to the ZVS and the ZCS in the whole operating range. Accordingly, the simulation and the experimental results are both found to be consistent.

Enhanced class E inverters (EFn or E/Fn) reduce the high peak switch voltage that is prevalent in class E inverters. Additionally, their stability and load regulation capabilities are improved as compared to class E inverters. This paper proposes an enhanced class E inverter (E/F3) with a piezoelectric transformer (PT) replacing the auxiliary resonant networks. This class E/F3 inverter is designed by adding a tuned auxiliary LC network at the third harmonic of the switching frequency to the class E inverter. Both the primary and auxiliary resonant networks are realized using a piezoelectric transformer (PT). The converter is simulated in LTspice and an experimental prototype is built and tested. It is found that the experimental results concur with the simulation results, with a measured efficiency of 90%. Thus, the theoretical design is verified and the concept of energy extraction is achieved.

Eye temperature and intraocular pressure are two measurable parameters that can be monitored as a health index with aging. Deviations from the normal range of intraocular pressure and temperature lead to the formation of many diseases. This study has been carried out to evaluate the relations between the physiological and anatomical changes of the eye with aging using mathematical modeling. 2D computer-aided design of the human eye has been developed for two major groups: 21 to 30 years and 41 to 50 years. The computer simulation has been carried out to determine the effects of physiological changes of tear evaporation, fluid dynamics, blood flow, and metabolism of eye tissues with aging. The simulation has been carried out in the standing and the supine position of a human body. The rate of temperature change is – 0.0075 K per year in the standing position and – 0.007 K per year in the supine position because of the modeled anatomical and physiological effects. All the three simulation parameters of this study, the temperature of the human eye, the intraocular pressure, and the aqueous humor flow velocity, have been compared with the recent practical and simulation-based experiments to validate our results.

The recognition of coral species based on underwater texture images pose a significant difficulty for machine learning algorithms, due to the three following challenges embedded in the nature of this data: 1) datasets do not include information about the global structure of the coral; 2) several species of coral have very similar characteristics; and 3) defining the spatial borders between classes is difficult as many corals tend to appear together in groups. For this reason, the classification of coral species has always required an aid from a domain expert. The objective of this paper is to develop an accurate classification model for coral texture images. Current datasets contain a large number of imbalanced classes, while the images are subject to inter-class variation. We have analyzed 1) several Convolutional Neural Network (CNN) architectures, 2) data augmentation techniques and 3) transfer learning. We have achieved the state-of-the art accuracies using different variations of ResNet on the two current coral texture datasets, EILAT and RSMAS., Comment: 22 pages, 10 figures

Automated object identification has seen significant progress during the last decade with close to human-level accuracy, aided by deep learning methods. With the rapid rise of obesity and other lifestyle-related diseases worldwide, the availability of fast, automated, and reliable image-based food calorie estimation is becoming a necessity. With the help of a deep learning-based automated object identification system, it is possible to introduce accurate and intelligent solutions in the form of a mobile app. However, for these kind of applications, processing speed is an important concern as the images should be processed in real time. Although plenty of studies have been conducted that focus on food image detection-based calorie estimation, there is still a lack of an image-driven, lightweight, fast, and reliable food calorie estimation system. In this paper, we propose a method based on the parameter-optimized Convolution Neural Networks (CNN) for detecting food images of regular meals using a handheld camera. Once identification process of the food items are complete, the corresponding calories and nutritional facts can be calculated using prior knowledge about the food class. Through our findings, we demonstrate that our proposed approach ensures high accuracy and can significantly simplify the existing manual calorie estimation procedures by converting them into a real-time automated process.

This paper presents an analysis and modeling of the class-E inverter for ZVS/ZVDS execution at any duty ratio. The methodology is to determine the input current to the inverter analytically under the assumption that it always remains positive. The latter is ensured by proper selection of the input inductance such that the inverter always operates either in (1) the border condition mode or in (2) the continuous conduction mode regardless of the input ripple. Using this input current and applying the boundary conditions, the required input capacitance for the ZVS/ZVDS execution is determined at a specified input/output voltage, output power and load. The analysis shows that the ZVS/ZVDS can be achieved while the input capacitance is selected appropriately. A comparison between the analytical and simulation results is also formulated involving the proposed and other existing models. The simulation results that are provided at different duty ratios demonstrate that they are in a better agreement with the proposed analytical model regardless of the input inductance and the state of input ripple current. The analytical modeling is facilitated by using MAPLE®.

This work proposes a novel concept for an intelligent and semi-autonomous human-cyber-physical system (HCPS) to operate future wind turbines in the context of Industry 5.0 technologies. The exponential increase in the complexity of next-generation wind turbines requires artificial intelligence (AI) to operate the machines efficiently and consistently. Evolving the current Industry 4.0 digital twin technology beyond a sole aid for the human decision-making process, the digital twin in the proposed system is used for highly effective training of the AI through machine learning. Human intelligence (HI) is elevated to a supervisory level, in which high-level decisions made through a human–machine interface break the autonomy, when needed. This paper also identifies and elaborates key enabling technologies (KETs) that are essential for realizing the proposed HCPS.

The enhanced class E configurations (EF n and E/F n ) are proposed to counter the high peak switch voltage issue in conventional class E inverter along with an improved stability and augmented load regulation feature. Despite these advantages, the class EF n and E/F n inverters demonstrate a lowered efficiency particularly at higher duty ratios and light loading conditions. The later also affect the stability of the inverter in the entire operating range. To counter the aforementioned issues, this paper proposes a novel family of class EF nm and E/F nm inverters. Two auxiliary resonant networks tuned at nth and mth harmonics of the primary resonant frequency are implemented. The auxiliary networks, configured across the switch and the primary resonant network, delivers the power to the secondary through a multiwinding transformer built in a single core to reduce power loss, space and cost. Consequently, the inverter becomes capable of providing lowered switch voltage and higher efficiency simultaneously. The designed converter is simulated in LTSPICE. A prototype is also built and tested to verify the theoretical design and simulation results. It is observed that the simulation results concur with the theoretical design. Also, the performance parameters are recorded for the designed inverter family.

The electrical characteristics of the photovoltaic (PV) module degrades due to several factors such as non-uniform aging, high atmospheric temperature, insufficient sun lights, shading, dust, soiling, cell damage, etc. In order to investigate the electrical performance of PV modules at such conditions, indoor solar simulators are commonly used for experimental research. In this work, a solar simulator is developed for a university lab to test the electrical characteristics of PV modules and array for low-cost experimental purposes. A microcontroller-based current-voltage (I-V) tracer with a real-time data logger system is used to maintain the reliability of the measuring system. The experimental investigation of the halogen-based solar simulator is carried out using both polycrystalline and monocrystalline PV panels of different power ratings. A prototype (5×8) PV array with 40 PV cells is tested by the solar simulator at a different level of irradiance. The experimental result shows that the average measurement error is ±0.022% which is within the acceptable range of a commercial simulator.