scholarly journals Detection and tracking people in real-time with YOLO object detector

2020 ◽  
pp. 69-75
Author(s):  
Aziza Srazhdinova ◽  
Asel’ Ahmetova ◽  
Sunvar Anvarov
Keyword(s):  
Neural Network ◽  
Real Time ◽  
Preliminary Analysis ◽  
Model Output ◽  
Detection And Tracking ◽  
Python Programming Language ◽  
Large Program ◽  
The University ◽  
Python Programming ◽  
Tracking Objects

In this article, we wrote not a large program to solve tasks for detection and tracking objects in real-time. The program was written in Python programming language. For object detection, a convolutional neural network was used with YOLOV3 architecture. A preliminary analysis was carried out of several variations of YOLO with CNN models. In the article, we justify why we want to use YOLO, and what it is and how to use and process the model output. We will also present the code in the form of a flowchart and as a result of the program's performance, we will show a picture of the program's operation in real-time, which was launched at one of the live lectures at the University.

Real-Time Multiphase Metering Using Non-Intrusive Microwave Sensor

2001 ◽  
Author(s):  
B. Hogan ◽  
A. I. Al-Shamma’a ◽  
J. Lucas
Keyword(s):  
Neural Network ◽  
Dielectric Properties ◽  
Real Time ◽  
Stream Flow ◽  
Microwave Cavity ◽  
Pump Speed ◽  
Microwave Sensor ◽  
Wide Range ◽  
The University ◽  
Oil Gas

Abstract At the University of Liverpool we have developed a real-time, non-intrusive multiphase dielectric meter capable of measuring the dielectric properties of different mixtures of oil, gas and water in full well stream flow. The design of such a microwave cavity using a range of rf and microwave is described. The experimental results with a wide range of multiphase mixtures from 0 to 100% are reported. In this paper we also present the implementation of a neural network to simulate the response of the meter under all conceivable conditions. Other parameters including pump speed, temperature, salt and sand are also discussed.

scholarly journals APPLICATION OF NEURAL NETWORKS TO DETERMINE THE COORDINATES OF THE SEAT OF FIRE BY MULTIPOINT ELECTRO-OPTICAL SYSTEM

2021 ◽  
pp. 248-251
Author(s):  
Д.Ф. Пирова ◽  
Б.Э. Забержинский ◽  
А.Г. Золин
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
High Performance ◽  
Network Models ◽  
X Rays ◽  
Neural Network Models ◽  
X Ray ◽  
Python Programming Language ◽  
Python Programming

Статья посвящена исследованию методов проектирования интеллектуальных информационных систем и применение моделей искусственных нейронных сетей для диагностического прогнозирования развития пневмонии посредством анализа рентгеновских снимков. В этой работе основное внимание уделяется классификации пневмонии и туберкулеза - двух основных заболеваний грудной клетки - на основе рентгеновских снимков грудной клетки. Данное исследование проводилось при помощи открытой нейросетевой библиотеки Keras и языка программирования Python. Система дает пользователю заключение о том, болен он или нет, тем самым помогая врачам и медицинскому персоналу принять быстрое и информированное решение о наличии заболевания. Разработанная модель, может определить, является ли рентгеновский снимок нормальным или имеет отклонения, которые могут быть пневмонией с точностью 94,87%. Полученные результаты указывают на высокую эффективность применения нейронных сетей при диагностировании пневмонии по рентгеновским снимкам. This paper is devoted to the study of methods of designing intellectual information systems and neural network models application on diagnostic prediction of pneumonia development by X-ray images analysis. This article focuses on the classification of pneumonia and tuberculosis - the two main chest diseases - based on chest x-rays. This study was carried out using the Keras open neural network library and the Python programming language. System returns user a conclusion whether the patient is ill or not helping medical staff to make a quick and informed decision about the presence of the disease. The developed model can determine is the X-ray image normal or has anomalies that can be pneumonia with accuracy up to 94.87%. The results obtained indicate the high performance of the applying neural networks in the diagnosis of pneumonia by X-ray images.

scholarly journals SEMANTIC ANALYSIS OF THE KAZAKH LANGUAGE BASED ON THE APPROACH OF NEURAL NETWORKS

2020 ◽  
Vol 5 (333) ◽  
pp. 68-75
Author(s):  
D. Rakhimova ◽  
◽  
A. Turganbayeva ◽  
◽  
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Semantic Analysis ◽  
Software Implementation ◽  
Graphical Form ◽  
Vector Representation ◽  
Python Programming Language ◽  
Modern Methods ◽  
Python Programming ◽  
Translation Systems

This paper provides an overview of existing modern methods and software approaches for semantic analysis. Based on the research done, it was revealed that, for the semantic analysis of text resources, an approach based on machine learning is most used. This article presents the developed algorithm for the semantic analysis of the text in the Kazakh language. The paper also presents a software solution to this approach implemented in the Python programming language. The vector representation of words was obtained by machine learning based on the corpus, which is 1 million sentences in the Kazakh language. In the software implementation, well-known libraries such as gensim, matplotlib, sklearn, numpy, etc. were used. Based on a set of semantically related pairs of words, an ontology for a specific document is built, which is formed during the operation of a neural network. The paper presents the results of the experiments in the graphical form of a set of words. The novelty of the proposed approach lies in the identification of semantic close words in meaning in texts in the Kazakh language. This work contributes to solving problems in machine translation systems, information retrieval, as well as in analysis and processing systems in the Kazakh language.

scholarly journals Object Detection and Tracking using YOLO v3 Framework for Increased Resolution Video

2020 ◽  
Vol 9 (6) ◽  
pp. 118-125
Keyword(s):  
Operating System ◽  
High Resolution ◽  
Real Time ◽  
The Other ◽  
Video Frame ◽  
Model Output ◽  
Detection And Tracking ◽  
Vehicle Detection And Tracking ◽  
Feature Extractor ◽  
Classified Image

The proposed system is used for vehicle detection and tracking from the high-resolution video. It detects the object (vehicles) and recognizes the object comparing its features with the features of the objects stored in the database. If the features match, then object is tracked. There are two steps of implementation, online and offline process. In offline process the data in the form of images are given to feature extractor and then after to the trained YOLO v3 model and weight files is generated form the pre-trained YOLO v3 model. In online phase, real-time video is applied to feature extractor to extract the features and then applied to the pre-trained YOLO v3 model. The other reference to YOLO v3 model pre-trained is the output of weight file. The YOLO v3 model process on the video frame and weight file extracted features, the model output is classified image. In YOLO v3 Darknet-53 is used along with Keras, some libraries with OpenCV, Tensor Flow, and Numpy. The proposed system is implemented on PC Intel Pentium G500, 8GB and operating system Windows 7 is used for processing our system. The system is tested on PASCAL VOC dataset and the results obtained are accuracy 80%, precision 80%, recall 100%, F1-Score 88%, mAP 76.7%, and 0.018%. The system is implemented using python 3.6.0 software and also tested using real-time video having 1280x720 and 1920x1080 resolutions. The execution time for one frame of video having resolution of 1280x720 (HD) and 1920x1080 (FHD) and 1280x720 (HD) are 1.840 second and 4.414808 seconds respectively with accuracy is 80%.

scholarly journals Development of Computer Vision Applications to Automate the Measurement of the Dimensions of Skin Wounds

Proceedings ◽  
2019 ◽  
Vol 19 (1) ◽  
pp. 18
Author(s):  
David Sánchez-Jiménez ◽  
Fernando Buchón-Moragues ◽  
Begoña Escutia-Muñoz ◽  
Rafael Botella-Estrada
Keyword(s):  
Computer Vision ◽  
Real Time ◽  
Programming Language ◽  
Three Dimensional ◽  
Skin Wounds ◽  
Two Dimensional ◽  
Python Programming Language ◽  
Computer Vision Applications ◽  
Python Programming

This paper shows the progress in the development of two computer vision applications for measuring skin wounds. Both applications have been written in Python programming language and make use of OpenCV and Scipy open source libraries. Their objective is to be part of a software that calculates the dimensions of skin wounds in an objective and reliable way. This could be useful in the clinical follow-up, assessing the evolution of skin wounds, as well as in research, comparing the efficacy of different treatments. Merging these two applications into a single one would allow to generate two-dimensional results in real time, and three-dimensional results after a few hours of processing.

scholarly journals Prediksi Nilai Tukar Dollar (USD) ke Rupiah (IDR) menggunakan Artificial Neural Network

2019 ◽  
Vol 2 (1) ◽  
pp. 1-7
Author(s):  
Ahmad Saparudin ◽  
Tiya Maulidina
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Exchange Rates ◽  
Business Strategy ◽  
Training Dataset ◽  
Python Programming Language ◽  
The Future ◽  
Business Forecasting ◽  
Artificial Neural ◽  
Python Programming

Prediction (forecasting) is the activity of predicting events in the future. In terms of business forecasting has many uses, especially for the leadership of the company one of them i.e. to define its business strategy in the future. In this research, carried out the predictions of exchange rates dollar (USD) to Indonesian rupiah (IDR) on 11/03/2019 - 15/03/2019 using artificial neural networks (ANN) with a training dataset from 01/01/2018 - 08/03/2019. Establishment of ANN in the study formed in the Python programming language. Based on the research conducted, a decrease in the price of the exchange rate of USD to IDR on 11/03/2019 – 15/03/2019.

Data driven methods for real time flood, drought and water quality monitoring: applications for Internet of Water

2020 ◽  
Author(s):  
Brianna Pagán ◽  
Nele Desmet ◽  
Piet Seuntjens ◽  
Erik Bollen ◽  
Bart Kuijpers
Keyword(s):  
Neural Network ◽  
Data Mining ◽  
Water Quality ◽  
Data Assimilation ◽  
Real Time ◽  
Quality Parameters ◽  
Water Quality Parameters ◽  
Weather Data ◽  
Model Estimate ◽  
The University

<p>The Internet of Water (IoW) is a large-scale permanent sensor network with 2500 small, energy-efficient wireless water quality sensors spread across Flanders, Belgium. This intelligent water management system will permanently monitor water quality and quantity in real time. Such a dense network of sensors with high temporal resolution (sub-hourly) will provide unprecedented volumes of data for drought, flood and pollution management, prediction and decisions. While traditional physical hydrological models are obvious choices for utilizing such a dataset, computational costs or limitations must be considered when working in real time decision making.</p><p>In collaboration with the Flemish Institute for Technological Research (VITO) and the University of Hasselt, we present several data mining and machine learning initiatives which support the IoW. Examples include interpolating grab sample measurements to river stretches to monitor salinity intrusion. A shallow feed forward neural network is trained on historical grab samples using physical characteristics of the river stretches (i.e. soil properties, ocean connectivity). Such a system allows for salinity monitoring without complex convection-diffusion modeling, and for estimating salinity in areas with less monitoring stations. Another highlighted project is the coupling of neural network and data assimilation schemes for water quality forecasting. A long short-term memory recurrent neural network is trained on historical water quality parameters and remotely sensed spatially distributed weather data. Using forecasted weather data, a model estimate of water quality parameters are obtained from the neural network. A Newtonian nudging data assimilation scheme further corrects the forecast leveraging previous day observations, which can aid in the correction for non-point or non-weather driven pollution influences. Calculations are supported by an optimized database system developed by the University of Hasselt which further exploits data mining techniques to estimate water movement and timing through the Flanders river network system. As geospatial data increases exponentially in both temporal and spatial resolutions, scientists and water managers must consider the tradeoff between computational resources and physical model accuracy. These type of hybrid approaches allows for near real-time analysis without computational limitations and will further support research to make communities more climate resilient.</p>

Application of Convolutional Neural Network (CNN) for the Prediction of Spam Mail

2021 ◽  
Vol 28 (1) ◽  
Author(s):  
C.I. Ejiofor ◽  
L.C. Ochei
Keyword(s):  
Neural Network ◽  
Natural Language ◽  
Convolutional Neural Network ◽  
Programming Language ◽  
Financial Cost ◽  
Training Time ◽  
Python Programming Language ◽  
Value Loss ◽  
The Rich ◽  
Python Programming

Spam mail has indeed become a global dilemma due to its coevolutionary nature. It has resulted in the loss of organizational resources, possibly financial cost incurred as well as time spent in addressing spam related issues. This has pushed organizations and researchers to the pinnacle of research with the aim of identifying needed solutions. This research paper explores the rich capabilities of Convolutional Neural Network (CNN) for predicting spam mail taking cognizant natural language capabilities. Spam mail prediction was simulated using a simulator built utilizing python programming language to capture the fundamentals of CNN. The CNN training was actualized using 10 epochs. The 1st epoch offers a training time of 4mins, 39s with a loss of 1.7578, accuracy of 0.3508, value loss of 1.2130 and value accuracy 0f 0.5719 while the 10th epoch presents a training time of 4mins, 6s with a loss of 0.5896, accuracy of 0.7936, value loss of 0.8941 and value accuracy of 0.6986.

Sign in / Sign up

Export Citation Format

Share Document