Search

X-ray coronary angiography (XCA) is a principal approach employed for identifying coronary disorders. Deep learning-based networks have recently shown tremendous promise in the diagnosis of coronary disorder from XCA scans. A deep learning-based edge adaptive instance normalization style transfer technique for segmenting the coronary arteries, is presented in this paper. The proposed technique combines adaptive instance normalization style transfer with the dense extreme inception network and convolution block attention module to get the best artery segmentation performance. We tested the proposed method on two publicly available XCA datasets, and achieved a segmentation accuracy of 0.9658 and Dice coefficient of 0.71. We believe that the proposed method shows that the prediction can be completed in the fastest time with training on the natural images, and can be reliably used to diagnose and detect coronary disorders., Comment: Accepted in ICCV Workshop on Computer Vision for Automated Medical Diagnosis 2021

Surveillance cameras are commonly used along US highways by State Departments of Transportation (DOT) to monitor the traffic status. State-of-the-art automatic traffic status monitoring requires that the cameras be preset at a predefined zoom level and viewing direction. However, in deployed systems these cameras have zoom, pan, and tilt features. The DOT operators will change the zoom level and camera viewing direction to gather traffic information for analysis. Therefore, the road and lane locations on the camera image cannot be at preset values. It is desirable to develop an intelligent system to automatically detect road and lanes using highway surveillance video with any zoom level and viewing direction. This paper describes a novel lane identification framework based on the Deep Learning starting point: YOLOv4 vehicle detection. This new technique does not use any painted lane markings. We identify the vehicles on the road and then aggregate the detected vehicles at one horizontal line on the image. This salient horizontal line is where the YOLOv4 gives the highest vehicle detection confidence scores. Most vehicles stay within a lane most of the time and change lanes only occasionally. This assumption is used to find the lane center locations according to the relative number of vehicles passing the salient horizontal line. Our method provides robust results in all weather conditions with a lane detection F1-Score above 0.85 compared to human labelled ground truth.

This paper describes a novel scheme for FPGA based high frequency clock phase difference measurement and correction using systematic subsample accumulation approach. Algorithm uses XOR based Phase detector signal that is subsampled using on-board clock of 375 KHz. Accumulated samples are averaged and a mathematical relation is established between duty cycles of XOR signal and reference signal. Depending on measured phase offset, clock shift bits of DAC device are incremented until known fixed offset is identified between the two clocks. The logic design is implemented on Virtex 5 FPGA for digital sub-system of a space-borne SAR payload. Realized algorithm clearly distinguishes different possible offsets to the order of 100 picoseconds. Implemented algorithm is area efficient consuming 150 LUTs and 130 registers only. This paper discusses algorithm development and implementation thoroughly, followed by flight model hardware results taken over a range of temperatures.

The pandemics like Coronavirus disease 2019 (COVID-19) require Governments and health professionals to make time-sensitive, critical decisions about travel restrictions and resource allocations. This paper identifies various factors that affect the spread of the disease using transaction data and proposes a model to predict the degree of spread of the disease and thus the number of medical resources required in upcoming weeks. We perform a region-wise analysis of these factors to identify the control measures that affect the minimal set of population. Our model also helps in estimating the surges in clinical demand and identifying when the medical resources would be saturated. Using this estimate, we suggest the preventive as well as corrective measures to avoid critical situations.

Around 71% of the Earth’s surface is covered in water. However, a severe scarcity of potable water persists in many countries, as only 2.5% of this water is available as freshwater in groundwater, lakes and rivers. Water is salient for the survival of all living beings and hence, providing access to clean drinking water requires the utmost urgency. Currently, existing technologies for water purification such as Multistage Flash (MSF), Multiple Effect, Vapor Compression (VC), Reverse Osmosis (RO), Ion Exchange, Electrodialysis, and Capacitive Deionization (CDI) require a lot of energy, leading to environmental pollution caused by the consumption of fossil fuels. Desalination of saline water using solar stills is an effective solution to overcome these problems. Solar energy is abundant, everlasting, economical and does not cause emission of greenhouse gases. This paper studies the mathematical modelling of a single slope single basin solar still and a still coupled with evacuated tube collectors. The performances of the two designs are compared to analyse their efficiency and the results thus obtained are presented. This design can also be installed in arid regions and villages where water scarcity is prevalent and access to the electricity grid is not feasible to fulfil the energy requirements.

Depression is the most comprehensive mood ailment that has a notable influence on mental health as well as hindrances in daily life. Machine learning models have contributed to the field of emotion detection in all areas including audio, visual and internet based text data. The idea directs at developing a machine learning based model utilising images and video as an input, to analyze the level of depression among users. Based on the analyzed features the individual will be classified into either of the following depression categories: Minimal, Mild, Moderate, Severe. In the process of depression level detection, the two crucial components are video input and the Beck Depression InventoryII. The solution generates as a result of the correlation between emotion vector and inventory vector represented using visual graphics.

Electric Vehicle (E.V.) relies on its batteries as the power source. One major issue hampering the use of EV is related to its short driving range due to exhaustion of battery. Battery power consumption suffers with many unbounded conditions that result in excess power consumption. Using the battery providently is the basic idea of the work discussed in this paper, and it can definitely contribute in augmenting the driving range of electric vehicle.In this paper, a system is described, by means of which the battery of E.V. can wisely be utilised by the constant loads of headlamps and thus, it becomes Provident Headlamps. By avoiding unnecessary use of battery and using it precisely when and wherever glow is required, battery power can be conserved. A circuitry is designed, which senses the illumination of the surroundings of vehicle and decides the output requirement of headlamps and will regulate the power from battery to the headlamps accordingly, thus conserving battery and slightly increasing the driving range. This technique has been extended to separate controls for left and right lamps, ensuring safety in two way traffic.

Position control of motor is of crucial importance in applications where precise positioning is required. It not only matters how quickly a certain position is obtained, but also how much the positioning error is reduced and how smoothly the position is achieved. Hence, trajectory planning is applied for controlling the path of the actuator, which is expected to obtain the required precise position. In the work presented herein, a brushless DC motor is used for achieving precise positioning, controlled through trajectory planning, using s-curve algorithm. The current work presents a model based approach for control system design for position control, along with experimental results. It presents the effect on performance parameters observed with s-curve algorithm implementation in comparison with the performance parameters obtained without trajectory planning.

Many multi-modal human computer interaction (HCI) systems interact with users in real-time by estimating the user’s pose. Generally, they estimate human poses using depth sensors such as the Microsoft Kinect. For multi-modal HCI interfaces to gain traction in the real world, however, it would be better for pose estimation to be based on data from RGB cameras, which are more common and less expensive than depth sensors. This has motivated research into pose estimation from RGB images. Convolutional Neural Networks (CNNs) represent the state-of-the-art in this literature, for example [1], [2], [9], [13], [14], and [15]. These systems estimate 2D human poses from RGB images. A problem with current CNN-based pose estimators is that they require large amounts of labeled data for training. If the goal is to train an RGB pose estimator for a new domain, the cost of collecting and more importantly labeling data can be prohibitive. A common solution is to train on publicly available pose data sets, but then the trained system is not tailored to the domain. We propose using RGB+D sensors to collect domain-specific data in the lab, and then training the RGB pose estimator using skeletons automatically extracted from the RGB+D data. This paper presents a case study of adapting the RMPE pose estimation network [2] to the domain of the DARPA Communicating with Computers (CWC) program [3], as represented by the EGGNOG data set [8]. We chose RMPE because it predicts both joint locations and Part Affinity Fields (PAFs) in real-time. Our adaptation of RMPE trained on automatically-labeled data outperforms the original RMPE on the EGGNOG data set.

Analyzing enormous amounts of healthcare data to obtain meaningful insights requires efficient and timely solutions. Diabetes is one of the most critical chronic healthcare problems that affect other organs of the human body. Hospital readmission, for patients with diabetes, is a common scenario where a discharged patient is admitted again within a specific time interval. Efficient techniques are needed which can predict the chance of such a readmission, thereby, allowing the possibility of targeted interventions. The aim of this paper is to discuss the performance of different prediction algorithms and associated collaborative paradigms for publically available diabetes data. Apache Spark is used, in the prototype, to decrease the training time. The prototype also addresses underlying challenges such as fault tolerance, scalability, and heterogeneity. The results of various experiments show that the collaborative technique increases the accuracy of a poor performing prediction algorithm by around 22% in one collaborative configuration.

The increasing concerns for health, what individual consumes has certainly become one of the most crucial factors to be measured. The statistics shows that diabetes is amongst the highest health concerns that are found in all age groups, posing a huge risk to form cardiovascular diseases in the long run. Hence, to overcome or probably to commercialize such complications, food industries are targeting the health cautious group of people to make profits. But the question remains, whether these products are really genuine. As we see, market aisles these days are crammed with variety of Anti-Diabetic food-products including wheat-flour, cooking-oil, milk tetra packs, etc. of varied brands claiming that they can manage normal blood glucose levels of a diabetic patient and/ or everyone. This raises a debate as to whether these Anti-Diabetic products are effectual preventive measures or useful for diabetes cure. Thus, in this paper we propose the DMAIC problem solving approach of Six Sigma powered by Threshold Based Incremental-Clustering Algorithm (TBCA) implemented here that takes into account nutritional composition of these Anti-Diabetic food-products to analyze their Sugar-release-controlling capability. To validate considered phenomenon, the association between Diabetes Mellitus (DM) and Anti-Diabetic products data sets, are examined through Principal Component Analysis (PCA) and TBCA-integrated-DMAIC steps of Six Sigma. The outcome of this study concludes that these products are constructive in regulating the blood glucose spikes of a diabetic patient. Extended learning outcome of this study will be, to add TBCA process as a new layer in DMAIC, so as to achieve distributed machine learning system, with sustainability care.

The Internet of Vehicles is an upcoming prototype, one which has introduced cutting edge applications and facilities which require the fulfillment of a certain Quality of Service (QoS). Vehicular Ad Hoc Networks (VANETs) are the backbone for IoV, providing the support for autonomous inter-vehicle communication and road safety management. The communication in VANETS is challenging as they suffer from several issues due to their intrinsic features such as higher vehicle mobility, rapidly changing topology, etc. Hence the strategies adopted in the routing layer play an important role in the overall application performance. Conventionally, in VANETs routing protocol design, each layer of the model is optimized individually, which does not satisfy dynamic system conditions in VANETs and lags behind the acceptable Quality of Service (QoS) standards. In this paper, we propose to design a novel cross layer based reliable vehicular routing model (CL-RVR), to facilitate reliable routing in VANETs, that would cater to QoS requirements for desired applications by combining the parameters from physical and network layers. Extensive experimental simulations are used to analyze the key performance metrics.

Today, waste management from its inception to its disposal is one of the important challenges for the municipal corporations in all over the world. Dust bins placed across cities set at open places are flooding because of increment in the waste each day and making unhygienic condition for the citizens, to maintain a strategic distance from such a circumstance we have proposed wireless solid waste management system for smart cities which allows municipal corporations to monitor status of dustbins remotely over web server and keep cities clean very efficiently by optimizing cost and time required for it. As soon as dustbin has reached its maximum level, waste management department gets alert via SMS via gsm module placed at dustbin so department can send waste collector vehicle to respective location to collect garbage. The objective of the project is to enhance practicality of IoT based solid waste collection and management system for smart city.

Today smartphones are readily available with source of power, faster processing capabilities and connectivity. In this paper, we propose design of circuit to harvest power from widely available standard 3.5mm headphone jack of smartphone itself. Headphone jack can generate tones (sine or square) within the frequency range of 20–20KHz. we have also developed GUI of an android application for tone generation which plays sine wave. The audio output channels of smartphones from different manufacturers were characterized in order to determine the range of design parameters. Proposed design provides few milliwatts to power mHealth peripherals and enhances practicality of smartphone as a medical device. It could be the better option for battery dependent devices. Headphone jack interface could be used to develop modular platforms to monitor vital signs such as Pulse rate, SpO2, EKG.

People communicate through words and gestures,but current voice-based computer interfaces such as Siri exploitonly words. This is a shame: human-computer interfaces wouldbe natural if they incorporated gestures as well as words. Tosupport this goal, we present a new dataset of naturally occurringgestures made by people working collaboratively on blocks worldtasks. The dataset, called EGGNOG, contains over 8 hours ofRGB video, depth video, and Kinect v2 body position data of 40subjects. The data has been semi-automatically segmented into24,503 movements, each of which has been labeled accordingto (1) its physical motion and (2) the intent of the participant.We believe this dataset will stimulate research into natural andgestural human-computer interfaces

This paper presents the results of laboratory work carried out for identifying the outer race bearing fault occurred in an induction motor. The knowledge about fault behavior of an induction motor is extremely important for overall operational life of the machine. The paper refers to spectral analysis of the motor stator current, which includes routine stator current, noise and specific fault current signature. Using separate healthy and faulty bearing on the machine, testing is carried out for obtaining set of healthy and faulty currents for different load conditions. The specific fault signature can be separated using feature extraction in time domain Wavelet and after getting spectral information using classification technique ANN fault is identified.

With progress in the area of computer science, it is achievable to read, process, store and generate information out of the available data. Humongous amount of data is generated, which is of mixed type, including time-series, Boolean, spatial-temporal and alpha-numeric data. This data is generated at a very giant speed and volume, which makes difficult for the traditional clustering algorithms to create and maintain the desired clusters. Thus, the proposed system encourages incremental clustering using a non-probability based similarity measure. The experimental results, of Correlation Based Incremental Clustering Algorithm (CBICA), which are obtained using the Pearson's coefficient of correlation, are compared with the experimental results of the Closeness-Factor Based Algorithm (CFBA), which uses the probability based similarity measures. The threshold computation is done to decide the cluster members in the post clustering phase, to adapt influx of new data. Wherein the new data is accommodated in the available clusters or new clusters are formed, depending upon the threshold values.

In the wake of ever-growing demand for power and energy across US and worldwide, development of energy efficient solutions has become very important. Considering data center applications, cooling power consumption constitutes significant part of the overall energy usage of the system. In the process of optimizing the energy consumed per performance unit, liquid cooling has become one of the key solutions. In this study, 2OU (OpenU; 1OU = 48mm) web servers are tested in a rack level and the effect of higher inlet temperatures in terms of IT and cooling powers, and internal component temperatures are reported. The study serves as a comparison for two different coolant pumping systems i.e. distributed vs. centralized systems. The cooling set up includes a mini rack capable of housing up to eleven liquid cooled web servers and two heat exchangers that exhaust the heat dissipated from the servers to the environment. Each server is equipped with two cold plates cooling the CPUs while rest of the components are air cooled. The configuration that consists of cold plates with integrated pumps is referred as distributed pumping system. Whereas, the configuration with no integrated pumps at cold plates and only has two pumps placed in series with heat exchanger at the rack is referred as centralized pumping system. To study performance characteristics such as device temperatures and power consumptions of server components, synthetic load has been generated on each server using stress-testing tools. The servers are tested for higher inlet temperatures ranging from 25°C to 45°C which falls within the ASHRAE liquid cooled envelope, W4 [1]. This current work is a follow-up study to the analysis conducted comparing centralized and distributed pumping [2].

Traditional electromechanical Energy Meters for electricity are now replaced by digital meters in domestic as well as commercial applications. The trend of automation can be incorporated in presently working digital energy meter to make it smarter. Idea is to design and implement a smart prepaid energy meter prototype having wireless communication facility. The proposed system presents a total electronic three phase four wire energy meter. All power measurements are taken in the digital domain. These readings are transmitted to the mobile of user via wireless GSM technology. User can have the updates of electric energy consumption data on his mobile. Controller is used for controlling all functions of meter. Proposed algorithm for power calculation reduces hardware requirements as power factor is calculated in algorithm which eliminates the need of the zero crossing detector circuit.

Drones are nowadays massively used in environmental applications. This is the device with no pilot. In the list one of the most important applications is environmental monitoring and weather forecasting. The experiment has been carried out at University of Evora, Portugal. In this work sensor unit with wireless module is embedded with drone receiver. In addition controller with programming environment is interfaced with sensor module fixed on top of drone. The environmental parameters proposed in this work are carbon monoxide, smoke, hydrogen, humidity, temperature. The gas sensor needs preheating time of 24 hours for consistent and proper sensing. The digital temperature and humidity sensor is available to get direct digital data instead of analog. The sensing method is effectively evaluated by means of experiments as a hexacopter. The Bluetooth module sends the data to Android Smartphone. The App in Smartphone displays the real time data from the sensor. It forms a loop of sensing, transmitting and receiving. The real time logging may disturb by propellers but its quit normal. The system is compatible for any kind of mobile carrier; honor goes to its lightweight and low form factor. In addition wireless network with artificial intelligence is developed to data log and monitor the environmental parameters.

In recent years due to globalization and increase in competition to capture market share, the customer is having many options of cars at much affordable prices. This has led to tremendous amount of cars on the road. The big cities in India are normally crowded and the city's existing infrastructure falls short to satisfy the growing needs. This problem is also combined with the inefficient use of parking space and uneven distribution of traffic which leads to congestion in vehicular traffic. This paper introduces a novel technique based on Internet of Things platform to improve efficiency of the existing parking systems. It guides driver to find nearest parking slots. The prototype uses cloud based architecture of Internet of things. There is minimum delay in finding an optimal parking slot which helps to ease the traffic congestion. The proposed system also helps users automatically find a free parking space at the smallest distance based on Haversine formula if nearest parking region is full.

A smart city is the global development vision to integrate multiple information and communication technology (ICT) solutions in a secure and simple way to manage a city's assets. Smart transportation is the basic component in future smart cities. Thus we are concentrating on the street light automation by detection of motion of object. Lighting which is not necessary can be neglected by targeted dimming of area of the city, town, roads or individual luminaries. The main objective of street light automation is to reduce the power consumption occurring due to lightening of street light during night when there is no objective motion on the street. During day time, the lights will be turned off and during night it will be turned on. In street light automation we are using Pulse Width Modulation (PWM) technique to control the light intensity in different condition like day, night, no vehicle condition & vehicle condition. In this paper we are working on street light automation by means of self responsive cars with collision avoidance system.

The prediction of Indian summer monsoon (ISM) variability on extended range (3–4 pentads in advance, where pentad refers to 5-day average) is crucial due to its usefulness in agricultural and hydrological planning. Monsoon rainfall occurs in the form of active/break spells which causes large-scale flood/drought condition. Prediction before 3–4 pentad may not stop this condition but can help us to prepare for the same. Hence Climate Forecast System model version 2 (CFSv2) has been adopted to develop a dynamical prediction system for monsoon rainfall on different time scales. The Extended Range Prediction (ERP) group at Indian Institute of Tropical Meteorology has indigenously developed an Ensemble Prediction System based on CFSv2 for the real-time prediction of active/break spells of ISM and has been providing the experimental forecasts up to 4 pentad lead. A multi-model ensemble (MME) framework employing different variants of CFSv2 has shown better skill in the ERP of ISM, compared to the individual variants. Still CFSv2 model shows statistical biases which cause lack of skill in MME and causes false prediction. In this study, it is proposed to employ the artificial neural network based technique to reduce the inherent biases by post-processing the raw model output. The performance is evaluated with mean square error, correlation coordinate and Brier skill score. Result shows that the probabilistic skill of the model is improved with the ANN-based technique.

Stereo vision has usefulness in many applications like 3D scene reconstruction, robot navigation, etc. The disparity between two original stereo images and depth maps are calculated to find the depth levels. An algorithm to generate disparity maps using SIMULINK tool is presented in this paper. The disparity map was first obtained by using MATLAB tool. The objective of the present work is to generate depth map using SIMULINK tool. The SIMULINK model used to obtain depth map dataset images is presented.

A smart antenna is actually a fusion of an array that changes individual antenna elements patterns in response of a dedicated signal to improve performance. Moving one step beyond of beam forming is adaptive algorithm which allows for the computation of continuously amend weights. The process of adaptation must satisfy a specified optimization principle. In this project two Adaptive algorithm Least Mean Square and Recursive Least Mean Square are implemented to revamp beam. Added optimization of beam is accomplished with the help of genetic algorithm which makes beam more optimize. This optimization method is used for the synthesis of antenna array radiation pattern in adaptive beam forming. The optimum value of weights of each antenna element is determined which produces a radiation pattern with minimum side lobe level. Analysis is being carried out with varying different element and space in between element to obtain the minimum weight with the help of genetic algorithm. This improves the performance to achieve the maximum reduction in side lobe level (of the order −20dB) with minimum function calls.

Smart antenna attempt to enhance the received signal, suppress all interfering signals, and increase capacity. Beamforming is one of the mostly used antenna technique. It adjust the radiation beam in one specific direction also reduces multiple access interference. It also reduces common channel interference and multiple path fading, in result it increases capacity of mobile network. In this paper two antenna beam forming algorithms are analyzed. These two algorithms are Least Mean Square and Minimum Variance Distortionless Response. Outputs of these two algorithms are in weights. By using these outputs weights, two parameters i.e. Mean Square Error and Power Beam Width are calculated and two algorithm are compared by this parameters and respective results are discussed.