Wyniki wyszukiwania - BazTech

1

Unplanned dilution prediction in open stope mining: developing new design charts using Artificial Neural Network classifier

Korigov Sultan, Adoko Amoussou Coffi, Sengani Fhatuwani

Journal of Sustainable Mining

|

2022

|

Vol. 21, iss. 2

157--168

EN

Minimizing dilution is essential in open stope mine design as excessive unplanned dilution can compromise the operation's profitability. One of the main challenges associated with the empirical dilution graph method used to design open stopes is how to determine the boundary of the dilution zones objectively. Hence, this paper explores the implementation of machine learning classifiers to bridge this gap in the conventional dilution graph method. Stope performance data consisting of the stope dilution (unplanned dilution), the modified stability number, and the hydraulic radius were compiled from a mine located in Kazakhstan. First, the conventional dilution graph methods were used to assess the dilution. Next, a Feed-Forward Neural Network (FFNN) classifier was implemented to predict each level of dilution. Overall, the FFNN results indicated that 97% of the stope surfaces were correctly classified, indicating an excellent classification performance, while the conventional dilution graph method did not show a good performance. In addition, the outputs of the FFNN were used to plot new dilution graphs with a probabilistic interpretation illustrating its practicability. It was concluded that the FFNN-based classifier could be a useful tool for open stope design in underground mines.

2

Porównanie skuteczności rozpoznawania obiektów morskich na podstawie obrazów FLIR za pomocą klasyfikatorów typu sieć neuronowa i PCA

Pietkiewicz Tadeusz, Dudek Patryk

Przegląd Elektrotechniczny

|

2021

|

R. 97, nr 11

21--25

PL

W pracy przedstawiono wyniki badań skuteczności metod rozpoznawania obiektów morskich na podstawie obrazów FLIR za pomocą metod opartych na różnych wariantach transformaty PCA oraz konwolucyjnej sieci neuronowej. Jako bazę wzorów wykorzystano zdjęcia dziewięciu typów obiektów zarejestrowanych na Morzu Bałtyckim. Obie metody potwierdziły swoją skuteczność, przy czym sieci neuronowe wykazały wyższą skuteczność na poziomie 96%, natomiast PCA - 80% - 87%..

EN

The paper presents results of the effectiveness research on the of maritime object recognition methods from FLIR images using methods based on different variants of the PCA classifier and a convolutional neural network classifier. Images of nine types of objects recorded in the Baltic Sea have been used as a pattern base. Both methods confirmed their effectiveness, with neural networks showing 96% effectiveness, while PCA 80% - 87%.

3

Monitoring of the crack propagation in welded joint of the tank using multi-class recognition

Bouraou N., Rupich S., Lukianchenko O., Kostina O.

Vibrations in Physical Systems

|

2018

|

Vol. 29

art. no. 2018013

EN

The numerical analysis of the vertical weld-fabricated steel tank is carried out taking into account the defects of the welds in the form of through-the-thickness cracks of different lengths and numbers, which are located in different zones of the object's ring. The influence of defects on the stress of the tank is estimated in the places of sensors installation under the action of vertical load. The usage of multi-class recognition is proposed by classifier based on Probabilistic Neural Network for monitoring of the crack propagation. Multidimensional vectors of diagnostic features are used for multi-class recognition. The training set of vectors is formed for defect-free and defect conditions, the classifier is trained and tested, the analysis of recognition efficiency is carried out by using the probability of correct multi-class recognition.

4

Estimation of severity level of non-proliferative diabetic retinopathy for clinical aid

Kaur J., Mittal D.

Biocybernetics and Biomedical Engineering

|

2018

|

Vol. 38, no. 3

708--732

EN

Diabetic retinopathy, a symptomless complication of diabetes, is one of the significant causes of vision impairment in the world. The early detection and diagnosis can reduce the occurrence of severe vision loss due to diabetic retinopathy. The diagnosis of diabetic retinopathy depends on the reliable detection and classification of bright and dark lesions present in retinal fundus images. Therefore, in this work, reliable segmentation of lesions has been performed using iterative clustering irrespective of associated heterogeneity, bright and faint edges. Afterwards, a computer-aided severity level detection method is proposed to aid ophthalmologists for appropriate treatment and effective planning in the diagnosis of non-proliferative diabetic retinopathy. This work has been performed on a composite database of 5048 retinal fundus images having varying attributes such as position, dimensions, shapes and color to make a reasonable comparison with state-of-the-art methods and to establish generalization capability of the proposed method. Experimental results on per-lesion basis show that the proposed method outperforms state-of-the methods with an average sensitivity/specificity/accuracy of 96.41/96.57/94.96 and 95.19/96.24/96.50 for bright and dark lesions respectively on composite database. Individual per-image based class accuracies delivered by the proposed method: No DR-95.9%, MA-98.3%, HEM-98.4%, EXU-97.4% and CWS-97.9% demonstrate the clinical competence of the method. Major contribution of the proposed method is that it efficiently grades the severity level of diabetic retinopathy in spite of huge variations in retinal images of different databases. Additionally, the substantial combined performance of these experiments on clinical and open source benchmark databases support a strong candidature of the proposed method in the diagnosis of non-proliferative diabetic retinopathy.

5

A classification framework for prediction of breast density using an ensemble of neural network classifiers

Kumar I., Bhadauria H. S., Virmani J., Thakur S.

Biocybernetics and Biomedical Engineering

|

2017

|

Vol. 37, no. 1

217--228

EN

The present work proposes a classification framework for the prediction of breast density using an ensemble of neural network classifiers. Expert radiologists, visualize the textural characteristics of center region of a breast to distinguish between different breast density classes. Accordingly, ROIs of fixed size are cropped from the center location of the breast tissue and GLCM mean features are computed for each ROI by varying interpixel distance 'd' from 1 to 15. The proposed classification framework consists of two stages, (a) first stage: this stage consists of a single 4-class neural network classifier NN0 (B-I/B-II/B-III/B-IV) which yields the output probability vector [PB-I PB-II PB-III PB-IV] indicating the probability values with which a test ROI belongs to a particular breast density class. (b) second stage: this stage consists of an ensemble of six binary neural network classifiers NN1 (B-I/B-II), NN2 (B-I/B-III), NN3 (B-I/B-IV), NN4 (B-II/B-III), NN5 (B-II/B-IV) and NN6 (B-III/B-IV). The output of the first stage of the classification framework, i.e. output on NN0 is used to obtain the two most probable classes for a test ROI. In the second stage this test ROI is passed through one of the binary neural networks, i.e. NN1 to NN6 corresponding to the two most probable classes predicted by NN0. [...]

6

The growing hierarchical self-organizing mapping for large scale power system static security assessment

Boudour Mohamed, Hellal Abdelhafid

Archives of Electrical Engineering

|

2004

|

Vol. 53, nr 1

49--64

EN

The paper presents an artificial neural-net based technique which combines supervised and unsupervised learning for on-line evaluating of power system static security. It automatically scans contingencies of a power system. The proposed approach allows the on-line security evaluation of (N - 1) contingencies by considering the pre-fault state vector. ANN-based pattern recognition is carried out with the growing hierarchical self-organizing feature mapping (GHSOM) in order to provide an adaptive neural net architecture during its unsupervised training process. Numerical tests, carried out on a IEEE 14 buses power system are presented and discussed. The analysis using such method provides accurate results and improves the effectiveness of system security evaluation. It is especially suitable for the static security assessment of large-scale power systems.

PL

W artykule przedstawiono, opartą o technikę sieci neuronowej, metodę ciągłej oceny bezpieczeństwa statycznego systemu, łączącą kontrolowane i niekontrolowane procesy uczenia. Nadzoruje ona automatycznie kontyngencje systemu. Proponowane podejście pozwala na ciągłą ocenę {N - 1) kontyngencji poprzez analizę wektora stanów przedawaryjnych. Rozpoznawanie obrazów, oparte o sieć neuronową, prowadzone jest z zastosowaniem rozwijającego, hierarchicznego, samoorganizującego się odwzorowywania właściwości dla uzyskania adaptacyjnej architektury sieci podczas procesu niekontrolowanego uczenia. Opisano i przedyskutowano badania numeryczne systemu z 14 szynami zbiorczymi (system IEEE 14). Analiza systemu prowadzona za pomocą przedstawionej metody zapewnia dokładność wyników i lepszą efektywność oceny bezpieczeństwa systemu. Metoda jest szczególnie przydatna do oceny bezpieczeństwa statycznego dużych systemów elektroenergetycznych.