COMPARISON OF CLASSIFIER TRAINING METHODS FOR IMAGE RECOGNITION

2020 ◽  
Author(s):  
Евгений Владимирович Вершинин ◽  
Денис Сергеевич Шахтарин ◽  
Владимир Евгеньевич Сорокин ◽  
Вадим Владимирович Сергеев
Keyword(s):  
Image Recognition ◽  
Training Methods ◽  
Classifier Training ◽  
Selection Of

В работе рассматриваются различные методы распознавания изображений. Представлено их сравнение и выбора оптимального. This paper discusses various methods of image recognition. Their comparison and selection of the optimal one are presented.

KINETIC INDICATORS OF TRAINING METHODS AND SELECTION OF YOUNG ROWERS

sportlogia ◽  
2021 ◽  
Vol 17 (1) ◽  
pp. 13-24
Author(s):  
Željko M. Rajković ◽  
◽  
Darko N. Mitrović ◽  
Vladimir K. Miletić ◽  
Petar M. Spaić ◽  
...  
Keyword(s):  
Average Power ◽  
The Other ◽  
Training Methods ◽  
Average Force ◽  
Other Hand ◽  
The World ◽  
Strict Selection ◽  
Selection Of

Modern diagnostics in rowing enables more and more possibilities for recording, and comparing numerous stroke variables. At the same time, many coaches fall into the trap of strict respect for the prescribed norms, ratios, and temporarily results, which the athlete must achieve if he wants to stay in the world of competitive rowing. On the example of the comparison of rowing schools RC "Danubius" and RC "Partizan", descriptive indicators are on the side of RC "Danubius" at a time of 2000m, average force and average power. No significant differences were found in average force (sig = 0,167) between rowers of RC "Danubius" and RC "Partizan", while statistically significant differences were recorded in time at 2000m (sig = 0,036) and power (sig = 0,02) in favor of rowers of RC “Danubius”. On the other hand, a higher correlation of average force (-0,955) and power (-0,928) with time on 2000m was achieved by RC "Partizan" than RC "Danubius" (-0,931) and (-0,896). The correlation between the average force, and the average power within one team shows a higher correlation for RC “Partizan" (0,95) compared to RC "Danubius" (0,755). The obtained results are not enough for single rower or crew elimination from competition to recreational section in the process of too frequent and strict selection of rowers, considering different possible ways of building rowing techniques and numerous parasitic factors that may affect measured variables, specialy at the age under 14 and novice rowers in general.

scholarly journals Research on Robustness of Related Images for Power Equipment Inspection

2021 ◽  
Vol 2113 (1) ◽  
pp. 012045
Author(s):  
Chunlei Zhou ◽  
Xiangzhou Chen ◽  
Wenli Liu ◽  
Tianyu Dong ◽  
Huang Yun
Keyword(s):  
Deep Learning ◽  
Image Recognition ◽  
Feature Fusion ◽  
Power Equipment ◽  
Actual Measurement ◽  
Practical Applications ◽  
Detection Algorithms ◽  
Traction Substations ◽  
Deep Learning Model ◽  
Selection Of

Abstract With the increase in the number of traction substations year by year, manual inspections are gradually being replaced by unattended inspections. Target detection algorithms based on deep learning are more widely used in intelligent inspections of power equipment. However, in practical applications, it is found that due to the small target to be detected, the accuracy of the deep learning model will decrease when the shooting angle is inclined and the light conditions are poor. This is because the algorithm’s robustness is low, and the detection ability of the model will be seriously affected when the angle or illumination difference with the sample is large. Based on this, the feature fusion part of the YOLOv3 algorithm and the selection of the loss function and the size of the anchor frame are improved, and the improved ASFF fusion method is used to classify various images in the power equipment. Actual measurement and repeated experiments show that the proposed method can be effectively applied to image recognition of various power equipment, optimize robustness, and greatly improve the image recognition efficiency of power equipment.

scholarly journals Factors Affecting Training and Physical Performance in Recreational Endurance Runners

Sports ◽  
2020 ◽  
Vol 8 (3) ◽  
pp. 35 ◽  
Author(s):  
Daniel Boullosa ◽  
Jonathan Esteve-Lanao ◽  
Arturo Casado ◽  
Leonardo A. Peyré-Tartaruga ◽  
Rodrigo Gomes da Rosa ◽  
...  
Keyword(s):  
Physical Performance ◽  
Best Practice ◽  
Scientific Evidence ◽  
Training Methods ◽  
Endurance Running ◽  
Factors Affecting ◽  
Recreational Runners ◽  
Key Aspects ◽  
Monitoring Performance ◽  
Selection Of

Endurance running has become an immensely popular sporting activity, with millions of recreational runners around the world. Despite the great popularity of endurance running as a recreational activity during leisure time, there is no consensus on the best practice for recreational runners to effectively train to reach their individual objectives and improve physical performance in a healthy manner. Moreover, there are lots of anecdotal data without scientific support, while most scientific evidence on endurance running was developed from studies observing both recreational and professional athletes of different levels. Further, the transference of all this information to only recreational runners is difficult due to differences in the genetic predisposition for endurance running, the time available for training, and physical, psychological, and physiological characteristics. Therefore, the aim of this review is to present a selection of scientific evidence regarding endurance running to provide training guidelines to be used by recreational runners and their coaches. The review will focus on some key aspects of the training process, such as periodization, training methods and monitoring, performance prediction, running technique, and prevention and management of injuries associated with endurance running.

The Application of Support Vector Machine in Transformer Fault Diagnosis

2014 ◽  
Vol 926-930 ◽  
pp. 3410-3413
Author(s):  
Xiang Lin Tan ◽  
Chang Shu ◽  
Yun Feng Peng
Keyword(s):  
Neural Network ◽  
Support Vector Machine ◽  
Fault Diagnosis ◽  
Model Selection ◽  
Basic Principle ◽  
Support Vector ◽  
Training Methods ◽  
Structure Selection ◽  
Transformer Fault ◽  
Selection Of

The basic principle, training methods and model selection of support vector machine (SVM) are expounded in this paper, and then we introduce SVM in transformer fault diagnosis which can overcome the problems of the structure selection of neural network.

scholarly journals Neural Network for Forecasting Energy Consumption Load of a Railway Marshalling Yard

2019 ◽  
Vol 17 (3) ◽  
pp. 6-15 ◽  
Author(s):  
V. N. Gridin ◽  
V. V. Doenin ◽  
V. S. Panishchev ◽  
I. D. Bysov
Keyword(s):  
Neural Network ◽  
Energy Consumption ◽  
Training Methods ◽  
Current Period ◽  
Binary Coding ◽  
The Neural Network ◽  
Additional Input ◽  
Daily Energy ◽  
Marshalling Yard ◽  
Selection Of

A multilayer neural network has been designed to forecast average daily energy consumption of a railway marshalling yard. The suggested model comprises a multilayer perceptron using 22 inputs, the n-th number of hidden layers and one output. The number of hidden layers in the neural network and neurons in them was chosen experimentally. A comparative selection of activation functions and training methods has allowed for all other parameters to achieve a minimum average relative error. Two types of loads corresponding to holidays (non-working) and working days were identified. An additional input node with binary coding and two nodes for coding the season were introduced due to a certain repeatability characterizing samples of prediction of loads of energy consumption of the marshalling yard depending on type of a day and on a season. As accounting of the dependence of the forecast on load values in previous days and years (dynamic dependencies) is most important factor, this neural network takes into account the average daily energy consumption during four days of the current period, precedingthe forecasted date, and the average daily power consumption during four days prior to this date during last three years.As a result, considering all factors and experimentally selected parameters of the neural network, the minimum resulting error of MAPE is about 1,4 %, which shows the advantage of the developed neural network in comparison with two other methods of solution of the problem, suggested by other researchers.

scholarly journals Explainable Artificial Intelligence for Human-Machine Interaction in Brain Tumor Localization

2021 ◽  
Vol 11 (11) ◽  
pp. 1213
Author(s):  
Morteza Esmaeili ◽  
Riyas Vettukattil ◽  
Hasan Banitalebi ◽  
Nina R. Krogh ◽  
Jonn Terje Geitung
Keyword(s):  
Artificial Intelligence ◽  
Deep Learning ◽  
Training Methods ◽  
Human Machine Interaction ◽  
Localization Accuracy ◽  
Explainable Ai ◽  
Recent Developments ◽  
High Level ◽  
Machine Interaction ◽  
Selection Of

Primary malignancies in adult brains are globally fatal. Computer vision, especially recent developments in artificial intelligence (AI), have created opportunities to automatically characterize and diagnose tumor lesions in the brain. AI approaches have provided scores of unprecedented accuracy in different image analysis tasks, including differentiating tumor-containing brains from healthy brains. AI models, however, perform as a black box, concealing the rational interpretations that are an essential step towards translating AI imaging tools into clinical routine. An explainable AI approach aims to visualize the high-level features of trained models or integrate into the training process. This study aims to evaluate the performance of selected deep-learning algorithms on localizing tumor lesions and distinguishing the lesion from healthy regions in magnetic resonance imaging contrasts. Despite a significant correlation between classification and lesion localization accuracy (R = 0.46, p = 0.005), the known AI algorithms, examined in this study, classify some tumor brains based on other non-relevant features. The results suggest that explainable AI approaches can develop an intuition for model interpretability and may play an important role in the performance evaluation of deep learning models. Developing explainable AI approaches will be an essential tool to improve human–machine interactions and assist in the selection of optimal training methods.

scholarly journals Selection of training methods in systems dedicated to detection of chemical hazards

2020 ◽  
Author(s):  
Magdalena Dobrowolska-Opała ◽  
Grzegorz Gudzbeler
Keyword(s):  
Chemical Substance ◽  
Training Methods ◽  
Training Module ◽  
Detection Systems ◽  
Hazardous Chemical ◽  
Chemical Hazards ◽  
Depth Analysis ◽  
Selection Of ◽  
The Eu

The paper presents the process of training methods’ selection in systems dedicated to detection of chemical hazards. By examining most common training methods used in detection systems, authors indicates which of them are the most relevant to hazardous chemical substance detection systems, assuming that basic efforts focus on the system as the whole (mainly on detection components) rather than the training module. Furthermore, the indicated methods are characterized in detail and combined with overall utility and economic analysis related to the dimension of the resources involved in the work on the entire project (by the example of the EU-SENSE project). In order to study the training methods and the factors influencing their choice, two research methods were used: (1) in-depth analysis of the literature on the subject of training tools in the field of crisis management and IT solutions, and (2) the case study of the EU-SENSE project (its training module).

scholarly journals Methodologic specificities of developing Russian-speaking students’ monologic skills in Ukrainian

2020 ◽  
Vol 2020 (1 (130)) ◽  
pp. 36-44
Author(s):  
Tatyana Lozan
Keyword(s):  
Scientific Literature ◽  
Mother Tongue ◽  
Necessary Condition ◽  
First Year ◽  
Training Methods ◽  
First Year Students ◽  
Methods And Techniques ◽  
Language Environment ◽  
Methodological Aspects ◽  
Selection Of

The article deals with the problems related to the peculiarities of teaching the Ukrainian language to Russian-speaking students under conditions of Pridnestrovie. The relevance of this study is to create a methodology for the development of the Ukrainian monologic speech of Russian-speaking students as a necessary condition for the development and functioning of the official Ukrainian language. The purpose of the article is to consider the psycholinguistic and methodological aspects of the development of one of the speech competence components – oral monologic speech of Russian-speaking students, to determine the ways, means and methods of developing and improving students’ skills to build monologic statements dedicated to everyday and professional topics. The objectives of the study are as follows: analysis of educational and scientific literature on the problem dealing with the development of the Russian-speaking students’ Ukrainian monologic speech. Taking into account the concretization of the tasks, the study involved a theoretical analysis of scientific and pedagogic literature, which allowed us to find out the modern approaches of scholars to the development of monologic speech within a foreign language environment. These methods were used: analysis, comparison, generalization and systematization of data of educational and scientific literature, pedagogical experience, conceptual provisions of the problem under study, and defining of the main concepts of the study. The article substantiates the importance of optimal selection and expediency of using General didactic and linguo-didactic principles, methods and techniques for effective developing of oral Ukrainian speech of the first-year students who speak Russian as their mother tongue. It is determined that the universal didactic and special principles make it possible to build an optimal system of methods and techniques, means and forms of organization and implementation of the task of monologic skills development. It is found out that the effectiveness of improving the students’ monologic speech depends on the appropriate choice and application of various linguo-didactic methods and techniques, their features and classification approaches. The teaching methods were elaborated based on the principles described in the article which, in its turn, determine the training methods and techniques as well as substantiate the selection of appropriate exercises, tasks, and other means; they constitute the linguo-didactic support for the process related to the developing of Ukrainian oral monologue.

Sign in / Sign up

Export Citation Format

Share Document