scholarly journals Image Colorization

2019 ◽  
Author(s):  
Дарья Михалина ◽  
Daria Mikhalina ◽  
Александр Кузьменко ◽  
Aleksandr Kuz'menko ◽  
Константин Дергачев ◽  
...  
Keyword(s):  
Neural Network ◽  
Urban Areas ◽  
Black And White ◽  
Number Of Layers ◽  
The Neural Network ◽  
Natural Landscapes ◽  
Video Card ◽  
Trained Neural Network ◽  
Lab Color Model ◽  
Cie Lab

The article discusses one of the latest ways to colorize a black and white image using deep learning methods. For colorization, a convolutional neural network with a large number of layers (Deep convolutional) is used, the architecture of which includes a ResNet model. This model was pre-trained on images of the ImageNet dataset. A neural network receives a black and white image and returns a colorized color. Since, due to the characteristics of ResNet, an input multiple of 255 is received, a program was written that, using frames, enlarges the image for the required size. During the operation of the neural network, the CIE Lab color model is used, which allows to separate the black and white component of the image from the color. For training the neural network, the Place 365 dataset was used, containing 365 different classes, such as animals, landscape elements, people, and so on. The training was carried out on the Nvidia GTX 1080 video card. The result was a trained neural network capable of colorizing images of any size and format. As example we had a speed of 0.08 seconds and an image of 256 by 256 pixels in size. In connection with the concept of the dataset used for training, the resulting model is focused on the recognition of natural landscapes and urban areas.

Study on Neural Networks Usage to Analyse Correlation between Spectrum of Vibration Acceleration Signal from Pin of Ball Mill and its Filling Level

2015 ◽  
Vol 770 ◽  
pp. 540-546 ◽  
Author(s):  
Yuri Eremenko ◽  
Dmitry Poleshchenko ◽  
Anton Glushchenko
Keyword(s):  
Neural Network ◽  
Ball Mill ◽  
Signal Amplitude ◽  
Signal Spectrum ◽  
Laboratory Model ◽  
Training Set ◽  
Vibration Acceleration ◽  
Acceleration Signal ◽  
The Neural Network ◽  
Trained Neural Network

The question about modern intelligent information processing methods usage for a ball mill filling level evaluation is considered. Vibration acceleration signal has been measured on a mill laboratory model for that purpose. It is made with accelerometer attached to a mill pin. The conclusion is made that mill filling level can not be measured with the help of such signal amplitude only. So this signal spectrum processed by a neural network is used. A training set for the neural network is formed with the help of spectral analysis methods. Trained neural network is able to find the correlation between mill pin vibration acceleration signal and mill filling level. Test set is formed from the data which is not included into the training set. This set is used in order to evaluate the network ability to evaluate the mill filling degree. The neural network guarantees no more than 7% error in the evaluation of mill filling level.

scholarly journals Use of a Multilayer Perceptron to Automate Terrain Assessment for the Needs of the Armed Forces

2018 ◽  
Vol 7 (11) ◽  
pp. 430 ◽  
Author(s):  
Krzysztof Pokonieczny
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Multilayer Perceptron ◽  
Learning Algorithm ◽  
Armed Forces ◽  
Shuttle Radar Topography Mission ◽  
Terrain Classification ◽  
The Neural Network ◽  
Data Configuration ◽  
Trained Neural Network

The classification of terrain in terms of passability plays a significant role in the process of military terrain assessment. It involves classifying selected terrain to specific classes (GO, SLOW-GO, NO-GO). In this article, the problem of terrain classification to the respective category of passability was solved by applying artificial neural networks (multilayer perceptron) to generate a continuous Index of Passability (IOP). The neural networks defined this factor for primary fields in two sizes (1000 × 1000 m and 100 × 100 m) based on the land cover elements obtained from Vector Smart Map (VMap) Level 2 and Shuttle Radar Topography Mission (SRTM). The work used a feedforward neural network consisting of three layers. The paper presents a comprehensive analysis of the reliability of the neural network parameters, taking into account the number of neurons, learning algorithm, activation functions and input data configuration. The studies and tests carried out have shown that a well-trained neural network can automate the process of terrain classification in terms of passability conditions.

scholarly journals Axle Temperature Monitoring and Neural Network Prediction Analysis for High-Speed Train under Operation

Symmetry ◽  
2020 ◽  
Vol 12 (10) ◽  
pp. 1662
Author(s):  
Wei Hao ◽  
Feng Liu
Keyword(s):  
Neural Network ◽  
Network Model ◽  
Neural Network Model ◽  
High Speed ◽  
Prediction Method ◽  
High Speed Train ◽  
The Neural Network ◽  
Neural Network Prediction ◽  
Network Prediction ◽  
Trained Neural Network

Predicting the axle temperature states of the high-speed train under operation in advance and evaluating working states of axle bearings is important for improving the safety of train operation and reducing accident risks. The method of monitoring the axle temperature of a train under operation, combined with the neural network prediction method, was applied. A total of 36 sensors were arranged at key positions such as the axle bearings of the train gearbox and the driving end of the traction motor. The positions of the sensors were symmetrical. Axle temperature measurements over 11 days with more than 38,000 km were obtained. The law of the change of the axle temperature in each section was obtained in different environments. The resultant data from the previous 10 days were used to train the neural network model, and a total of 800 samples were randomly selected from eight typical locations for the prediction of axle temperature over the following 3 min. In addition, the results predicted by the neural network method and the GM (1,1) method were compared. The results show that the predicted temperature of the trained neural network model is in good agreement with the experimental temperature, with higher precision than that of the GM (1,1) method, indicating that the proposed method is sufficiently accurate and can be a reliable tool for predicting axle temperature.

scholarly journals Modified Depth-Map Inpainting Method Using the Neural Network

2019 ◽  
Vol 224 ◽  
pp. 04005
Author(s):  
Nikolay Gapon ◽  
Roman Sizyakin ◽  
Marina Zhdanova ◽  
Oksana Balabaeva ◽  
Yigang Cen
Keyword(s):  
Neural Network ◽  
Geometric Model ◽  
Similarity Criterion ◽  
Depth Map ◽  
Experimental Results ◽  
Stereo Pair ◽  
Image Structure ◽  
The Neural Network ◽  
Modern Methods ◽  
Trained Neural Network

This paper proposes a method for reconstructing a depth map obtained using a stereo pair image. The proposed approach is based on a geometric model for the synthesis of patches. The entire image is preliminarily divided into blocks of different size, where large blocks are used to restore homogeneous areas, and small blocks are used to restore details of the image structure. Lost pixels are recovered by copying the pixel values from the source based on the similarity criterion. We used a trained neural network to select the “best like” patch. Experimental results show that the proposed method gives better results than other modern methods, both in subjective and objective measurements for reconstructing a depth map.

Neural networks for modelling the energy consumption of metro trains

2019 ◽  
Vol 234 (7) ◽  
pp. 722-733 ◽  
Author(s):  
Pablo Martínez Fernández ◽  
Pablo Salvador Zuriaga ◽  
Ignacio Villalba Sanchís ◽  
Ricardo Insa Franco
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Energy Consumption ◽  
Mean Square Error ◽  
Mean Square ◽  
The Neural Network ◽  
Train Speed ◽  
Input Variables ◽  
Trained Neural Network ◽  
Traction Power

This paper presents the application of machine learning systems based on neural networks to model the energy consumption of electric metro trains, as a first step in a research project that aims to optimise the energy consumed for traction in the Metro Network of Valencia (Spain). An experimental dataset was gathered and used for training. Four input variables (train speed and acceleration, track slope and curvature) and one output variable (traction power) were considered. The fully trained neural network shows good agreement with the target data, with relative mean square error around 21%. Additional tests with independent datasets also give good results (relative mean square error = 16%). The neural network has been applied to five simple case studies to assess its performance – and has proven to correctly model basic consumption trends (e.g. the influence of the slope) – and to properly reproduce acceleration, holding and braking, although it tends to slightly underestimate the energy regenerated during braking. Overall, the neural network provides a consistent estimation of traction power and the global energy consumption of metro trains, and thus may be used as a modelling tool during further stages of research.

scholarly journals Experiment in Finding Look-Alike European Cities Using Urban Atlas Data

2020 ◽  
Vol 9 (6) ◽  
pp. 406
Author(s):  
Zdena Dobesova
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Data Mining ◽  
Land Use ◽  
Urban Areas ◽  
Nearest Neighbor ◽  
Land Use Patterns ◽  
Use Patterns ◽  
European Cities ◽  
Trained Neural Network

The integration of geography and machine learning can produce novel approaches in addressing a variety of problems occurring in natural and human environments. This article presents an experiment that identifies cities that are similar according to their land use data. The article presents interesting preliminary experiments with screenshots of maps from the Czech map portal. After successfully working with the map samples, the study focuses on identifying cities with similar land use structures. The Copernicus European Urban Atlas 2012 was used as a source dataset (data valid years 2015–2018). The Urban Atlas freely offers land use datasets of nearly 800 functional urban areas in Europe. To search for similar cities, a set of maps detailing land use in European cities was prepared in ArcGIS. A vector of image descriptors for each map was subsequently produced using a pre-trained neural network, known as Painters, in Orange software. As a typical data mining task, the nearest neighbor function analyzes these descriptors according to land use patterns to find look-alike cities. Example city pairs based on land use are also presented in this article. The research question is whether the existing pre-trained neural network outside cartography is applicable for categorization of some thematic maps with data mining tasks such as clustering, similarity, and finding the nearest neighbor. The article’s contribution is a presentation of one possible method to find cities similar to each other according to their land use patterns, structures, and shapes. Some of the findings were surprising, and without machine learning, could not have been evident through human visual investigation alone.

Machine Learning for Web Proxy Analytics

2019 ◽  
Vol 1 (2) ◽  
pp. 30-41
Author(s):  
Mark Maldonado ◽  
Ayad Barsoum
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Decision Making ◽  
Everyday Life ◽  
Small Businesses ◽  
Electronic Device ◽  
Proxy Server ◽  
Proxy Servers ◽  
The Neural Network ◽  
Trained Neural Network

Proxy servers used around the globe are typically graded and built for small businesses to large enterprises. This does not dismiss any of the current efforts to keep the general consumer of an electronic device safe from malicious websites or denying youth of obscene content. With the emergence of machine learning, we can utilize the power to have smart security instantiated around the population's everyday life. In this work, we present a simple solution of providing a web proxy to each user of mobile devices or any networked computer powered by a neural network. The idea is to have a proxy server to handle the functionality to allow safe websites to be rendered per request. When a website request is made and not identified in the pre-determined website database, the proxy server will utilize a trained neural network to determine whether or not to render that website. The neural network will be trained on a vast collection of sampled websites by category. The neural network needs to be trained constantly to improve decision making as new websites are visited.

Damageless Watermark Extraction Using Nonlinear Feature Extraction Scheme Trained on Frequency Domain

2011 ◽  
pp. 117-142
Author(s):  
Kensuke Naoe ◽  
Yoshiyasu Takefuji
Keyword(s):  
Neural Network ◽  
Feature Extraction ◽  
Frequency Domain ◽  
Supervised Learning ◽  
The Neural Network ◽  
New Information ◽  
Extraction Scheme ◽  
Trained Neural Network ◽  
Nonlinear Feature Extraction ◽  
Nonlinear Feature

In this chapter, we propose a new information hiding and extracting method without embedding any information into the target content by using a nonlinear feature extraction scheme trained on frequency domain. The proposed method can detect hidden bit patterns from the content by processing the coefficients of the selected feature subblocks to the trained neural network. The coefficients are taken from the frequency domain of the decomposed target content by frequency transform. The bit patterns are retrieved from the network only with the proper extraction keys provided. The extraction keys, in the proposed method, are the coordinates of the selected feature subblocks and the neural network weights generated by the supervised learning of the neural network. The supervised learning uses the coefficients of the selected feature subblocks as the set of input values, and the hidden bit patterns are used as the teacher signal values of the neural network, which is the watermark signal in the proposed method. With our proposed method, we are able to introduce a watermark scheme with no damage to the target content.

Prediction of Hardness Distribution in Plasma Arc Surface Hardening Using Neural Network

2017 ◽  
pp. 318-333
Author(s):  
Manoj Kumar
Keyword(s):  
Neural Network ◽  
Surface Hardening ◽  
Regression Models ◽  
Network Models ◽  
Hardness Distribution ◽  
Plasma Arc ◽  
Statistical Regression ◽  
Neural Network Models ◽  
The Neural Network ◽  
Trained Neural Network

In this chapter, an attempt has been made to develop neural network models to predict the hardness distribution of hardened zone in plasma arc surface hardening process. The back propagation method with the Levenberg-Marquardt algorithm was used to train the neural network models. Hardness distributions were collected by the experimental setup in the laboratory and the associated data were used to train the neural network models. Furthermore, the prediction of neural network models were compared with those obtained from a statistical regression models. It is confirmed experimentally that the hardness distribution can be accurately predicted by the trained neural network models. The accuracy of hardness distribution prediction using neural network is superior to that using other statistical regression models.

Analysis of Symbol Generation and Integration in a Unified Model Based on a Neural Network

2005 ◽  
Vol 9 (3) ◽  
pp. 297-303
Author(s):  
Yukari Yamauchi ◽  
◽  
Shun'ichi Tano ◽  
Keyword(s):  
Neural Network ◽  
Reinforcement Learning ◽  
Simple Model ◽  
Unified Model ◽  
Symbolic Processing ◽  
Advantages And Disadvantages ◽  
Intelligent Processing ◽  
Knowledge Based ◽  
The Neural Network ◽  
Trained Neural Network

The computational (numerical information) and symbolic (knowledge-based) processing used in intelligent processing has advantages and disadvantages. A simple model integrating symbols into a neural network was proposed as a first step toward fusing computational and symbolic processing. To verify the effectiveness of this model, we first analyze the trained neural network and generate symbols manually. Then we discuss generation methods that are able to discover effective symbols during training of the neural network. We evaluated these through simulations of reinforcement learning in simple football games. Results indicate that the integration of symbols into the neural network improved the performance of player agents.

Sign in / Sign up

Export Citation Format

Share Document