scholarly journals A Novel Convolutional Neural Network for Classifying Indian Coins by Denomination

2021 ◽  
Author(s):  
Yash Chauhan ◽  
Prateek Singh
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Academic Research ◽  
High Volume ◽  
Classification Problem ◽  
General Purpose ◽  
Training Data ◽  
Slot Machines ◽  
Validation Data ◽  
Starting Point

Coins recognition systems have humungous applications from vending and slot machines to banking and management firms which directly translate to a high volume of research regarding the development of methods for such classification. In recent years, academic research has shifted towards a computer vision approach for sorting coins due to the advancement in the field of deep learning. However, most of the documented work utilizes what is known as ‘Transfer Learning’ in which we reuse a pre-trained model of a fixed architecture as a starting point for our training. While such an approach saves us a lot of time and effort, the generic nature of the pre-trained model can often become a bottleneck for performance on a specialized problem such as coin classification. This study develops a convolutional neural network (CNN) model from scratch and tests it against a widely-used general-purpose architecture known as Googlenet. We have shown in this study by comparing the performance of our model with that of Googlenet (documented in various previous studies) that a more straightforward and specialized architecture is more optimal than a more complex general architecture for the coin classification problem. The model developed in this study is trained and tested on 720 and 180 images of Indian coins of different denominations, respectively. The final accuracy gained by the model is 91.62% on the training data, while the accuracy is 90.55% on the validation data.

scholarly journals Image classification of different clove (Syzygium aromaticum) quality using deep learning method with convolutional neural network algorithm

2021 ◽  
Vol 905 (1) ◽  
pp. 012018
Author(s):  
I Y Prayogi ◽  
Sandra ◽  
Y Hendrawan
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
Convolutional Neural Network ◽  
Test Data ◽  
Training Data ◽  
Syzygium Aromaticum ◽  
Learning Method ◽  
Validation Data ◽  
Reading Accuracy

Abstract The objective of this study is to classify the quality of dried clove flowers using deep learning method with Convolutional Neural Network (CNN) algorithm, and also to perform the sensitivity analysis of CNN hyperparameters to obtain best model for clove quality classification process. The quality of clove as raw material in this study was determined according to SNI 3392-1994 by PT. Perkebunan Nusantara XII Pancusari Plantation, Malang, East Java, Indonesia. In total 1,600 images of dried clove flower were divided into 4 qualities. Each clove quality has 225 training data, 75 validation data, and 100 test data. The first step of this study is to build CNN model architecture as first model. The result of that model gives 65.25% reading accuracy. The second step is to analyze CNN sensitivity or CNN hyperparameter on the first model. The best value of CNN hyperparameter in each step then to be used in the next stage. Finally, after CNN hyperparameter carried out the reading accuracy of the test data is improved to 87.75%.

scholarly journals Perbandingan Preskrining Lesi Kulit berbasis Convolutional Neural Network: Citra Asli dan Tersegmentasi

2021 ◽  
Vol 8 (4) ◽  
pp. 793
Author(s):  
Agung Wahyu Setiawan
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Skin Lesion ◽  
Convolutional Neural Network ◽  
Training Data ◽  
Validation Data ◽  
Background Removal ◽  
Time To Build ◽  
Significant Difference ◽  
Segmented Images

<p>Seiring dengan bertambahnya prevalensi lesi kulit, maka diperlukan adanya preskrining lesi kulit mandiri yang mudah dan akurat. Pada studi ini, dilakukan perbandingan kinerja preskrining lesi kulit berbasis <em>Convolutional Neural Network</em> antara citra asli dan citra tersegmentasi <em>Grabcut</em> sebagai masukan. Ada dua parameter kinerja yang digunakan sebagai evaluasi, yaitu akurasi serta waktu pembuatan model. Tidak ada perbedaan kinerja akurasi pelatihan dan validasi pembelajaran mesin menggunakan citra asli dengan citra tersegmentasi. Meskipun terdapat proses tambahan berupa penghilangan latar belakang citra menggunakan algortima <em>Grubcut</em>, akurasi pelatihan maupun validasi preskrining lesi kulit tidak mengalami peningkatan yang signifikan. Pada parameter kinerja yang kedua, waktu pembuatan model dipengaruhi oleh jumlah data latih dan validasi. Semakin kecil jumlah data latih yang digunakan, maka waktu pembuatan model akan semakin cepat, dan sebaliknya. Disamping itu, proporsi antara jumlah data latih dengan validasi juga berpengaruh ke akurasi validasi. Pada studi ini, dengan menggunakan jumlah data latih yang lebih kecil dibandingkan data validasi, akurasi validasi mengalami peningkatan dari 0,82% menjadi 0,90%. Studi ini telah memberikan bukti bahwa pada preskrining lesi kulit menggunakan pembelajaran mesin berbasis CNN tidak diperlukan mekanisme adanya penghilangan latar belakang citra. Selain itu, pembuatan model pembelajaran mesin berbasis CNN dapat dilakukan dengan menggunakan data latih sekitar 22,41% dari data total. Diharapkan, hasil studi ini dapat dimanfaatkan untuk pengembangan aplikasi preskrining lesi kulit menggunakan pembelajaran mesin berbasis CNN pada komputer atau gawai dengan sumber daya komputasi yang rendah.</p><p> </p><p><strong><em>Abstract</em></strong></p><p> </p><p class="Abstract"><em>It is necessary to develop a self-prescreening of skin lesion due to the prevalence is increasing every year. This study tries to compare and evaluate the performance of prescreening of a skin lesion in the original and segmented images using Convolutional Neural Network. The Grabcut algorithm is used in the image segmentation process. Two parameters are used to evaluate the performance of the classification, i.e. accuracy and time to build the model. The results show that there is no significant difference in training and validation accuracy between original and segmented images. Even though there is an additional process in removing image background using Grabcut, the accuracy of training and validation do not increase significantly. In the second performance indicator, the time to build the model is influenced by the numbers of training and validation data that are used. The smaller the amount of training data used, the faster the model creation time will be. In addition, the proportion between the amount of training data and validation also affects the accuracy of validation. In this study, using a smaller amount of training data than the validation data, the validation accuracy increased from 0.82 to 0.90. This study has provided evidence that prescreening of skin lesions using machine learning based on CNN does not require image background removal and only about 22.41% of the total data are needed to build the model. One of the contributions of this study is that the results of this study can be used for the development of a skin lesion prescreening application using CNN-based machine learning on computers or devices with low computational resources.</em></p>

scholarly journals Maturity classification of cacao through spectrogram and convolutional neural network

2020 ◽  
Vol 8 (3) ◽  
pp. 228-233
Author(s):  
Gilbert E. Bueno ◽  
Kristine A. Valenzuela ◽  
Edwin R. Arboleda
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Poor Quality ◽  
Training Data ◽  
Training Process ◽  
Validation Data ◽  
Sound File ◽  
Audio Data ◽  
Mel Frequency Cepstral Coefficient

Cacao pod's ideal harvesting time is when it is about to be ripe. Immature harvest would result in hard cacao beans not suitable for fermentation, while overripe cacao pods lead to fungal-infected, defective, and poor-quality yields. The demand for high-quality cacao products is expected to rise due to advancing technology in the present. Pre-harvesting needs to provide optimal identification of which amongst the pods are ripened enough and ready for the next stage of the cacao process. This paper recommends a technique to determine the ripeness of cacao. Nine hundred thirty-three cacao samples were used to collect thumping audio data at five different pod's exocarp locations. Each sound file is 1 second long, creating 4665 cacao sound file datasets at 16kHz sample rate and 16-bit audio bit depth. The process of the Mel-Frequency Cepstral Coefficient Spectogram was then applied to extract recognizable features for the training process. The deep learning method integrated was a convolutional neural network (CNN) to classify the cacao sound successfully. The experimental design model's output exhibits an accuracy of 97.50 % for the training data and 97.13 % for the validation data. While the overall accuracy mean of the classification system is 97.46 %, whether the cacao is unripe or ripe.

Abstract 268: A Convolutional Neural Network for Real-Time Discrimination of Shock-Indicated Rhythms During CPR

Circulation ◽  
2018 ◽  
Vol 138 (Suppl_2) ◽  
Author(s):  
Tetsuo Hatanaka ◽  
Hiroshi Kaneko ◽  
Aki Nagase ◽  
Seishiro Marukawa
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Sensitivity And Specificity ◽  
Binary Classification ◽  
Total Duration ◽  
Training Data ◽  
Validation Data ◽  
Data Set ◽  
The Neural Network ◽  
Fully Connected

Introduction: An interruption of chest compressions during CPR adversely affects patient outcome. Currently, however, periodical interruptions are unavoidable to assess the ECG rhythms and to give shocks for defibrillation if indicated. Evidence suggests a 5-second interruption immediately before shocks may translate into ~15% reduction of the chance of survival. The objective of this study was to build, train and validate a convolutional neural network (artificial intelligence) for detecting shock-indicated rhythms out of ECG signals corrupted with chest compression artifacts during CPR. Methods: Our convolutional neural network consisted of 7 convolutional layers, 3 pooling layers and 3 fully-connected layers for binary classification (shock-indicated vs non-shock-indicated). The input data set was a spectrogram consisting of 56 frequency-bins by 80 time-segments transformed from a 12.16-seconds ECG signal. From AEDs used for 236 patients with out-of-hospital cardiac arrest, 1,223 annotated ECG strips were extracted. Ventricular fibrillation and wide-QRS ventricular tachycardia with HR>180 beats/min were annotated as shock-indicated, and the others as non-shock-indicated. The total length of the strips was 8:49:57 (hr:min:sec) and 8:02:07 respectively for shock-indicated and non-shock-indicated rhythms. Those strips were converted into 465,102 spectrograms allowing partial overlaps and were fed into the neural network for training. The validation data set was obtained from a separate group of 225 patients, from which annotated ECG strips (total duration of 62:11:28) were extracted, yielding 43,800 spectrograms. Results: After the training, both the sensitivity and specificity of detecting shock-indicated rhythms over the training data set were 99.7% - 100% (varying with training instances). The sensitivity and specificity over the validation data set were 99.3% - 99.7% and 99.3% - 99.5%, respectively. Conclusions: The convolutional neural network has accurately and continuously evaluated the ECG rhythms during CPR, potentially obviating the need for rhythm checks for defibrillation during CPR.

Fully automated contrast and non-contrast cardiac view detection in echocardiography a multi-centre, multi-vendor study

2020 ◽  
Vol 41 (Supplement_2) ◽  
Author(s):  
S Gao ◽  
D Stojanovski ◽  
A Parker ◽  
P Marques ◽  
S Heitner ◽  
...  
Keyword(s):  
Neural Network ◽  
Training Data ◽  
Classification Model ◽  
Validation Dataset ◽  
Funding Source ◽  
Private Company ◽  
Validation Data ◽  
Independent Test ◽  
Model Training ◽  
Confusion Matrices

Abstract Background Correctly identifying views acquired in a 2D echocardiographic examination is paramount to post-processing and quantification steps often performed as part of most clinical workflows. In many exams, particularly in stress echocardiography, microbubble contrast is used which greatly affects the appearance of the cardiac views. Here we present a bespoke, fully automated convolutional neural network (CNN) which identifies apical 2, 3, and 4 chamber, and short axis (SAX) views acquired with and without contrast. The CNN was tested in a completely independent, external dataset with the data acquired in a different country than that used to train the neural network. Methods Training data comprised of 2D echocardiograms was taken from 1014 subjects from a prospective multisite, multi-vendor, UK trial with the number of frames in each view greater than 17,500. Prior to view classification model training, images were processed using standard techniques to ensure homogenous and normalised image inputs to the training pipeline. A bespoke CNN was built using the minimum number of convolutional layers required with batch normalisation, and including dropout for reducing overfitting. Before processing, the data was split into 90% for model training (211,958 frames), and 10% used as a validation dataset (23,946 frames). Image frames from different subjects were separated out entirely amongst the training and validation datasets. Further, a separate trial dataset of 240 studies acquired in the USA was used as an independent test dataset (39,401 frames). Results Figure 1 shows the confusion matrices for both validation data (left) and independent test data (right), with an overall accuracy of 96% and 95% for the validation and test datasets respectively. The accuracy for the non-contrast cardiac views of &gt;99% exceeds that seen in other works. The combined datasets included images acquired across ultrasound manufacturers and models from 12 clinical sites. Conclusion We have developed a CNN capable of automatically accurately identifying all relevant cardiac views used in “real world” echo exams, including views acquired with contrast. Use of the CNN in a routine clinical workflow could improve efficiency of quantification steps performed after image acquisition. This was tested on an independent dataset acquired in a different country to that used to train the model and was found to perform similarly thus indicating the generalisability of the model. Figure 1. Confusion matrices Funding Acknowledgement Type of funding source: Private company. Main funding source(s): Ultromics Ltd.

scholarly journals Performance Evaluation of Deep CNN-Based Crack Detection and Localization Techniques for Concrete Structures

Sensors ◽  
2021 ◽  
Vol 21 (5) ◽  
pp. 1688
Author(s):  
Luqman Ali ◽  
Fady Alnajjar ◽  
Hamad Al Jassmi ◽  
Munkhjargal Gochoo ◽  
Wasif Khan ◽  
...  
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Crack Detection ◽  
Concrete Structures ◽  
Model Performance ◽  
Training Data ◽  
Computational Time ◽  
Data Heterogeneity ◽  
Public Datasets ◽  
Detection And Localization

This paper proposes a customized convolutional neural network for crack detection in concrete structures. The proposed method is compared to four existing deep learning methods based on training data size, data heterogeneity, network complexity, and the number of epochs. The performance of the proposed convolutional neural network (CNN) model is evaluated and compared to pretrained networks, i.e., the VGG-16, VGG-19, ResNet-50, and Inception V3 models, on eight datasets of different sizes, created from two public datasets. For each model, the evaluation considered computational time, crack localization results, and classification measures, e.g., accuracy, precision, recall, and F1-score. Experimental results demonstrated that training data size and heterogeneity among data samples significantly affect model performance. All models demonstrated promising performance on a limited number of diverse training data; however, increasing the training data size and reducing diversity reduced generalization performance, and led to overfitting. The proposed customized CNN and VGG-16 models outperformed the other methods in terms of classification, localization, and computational time on a small amount of data, and the results indicate that these two models demonstrate superior crack detection and localization for concrete structures.

scholarly journals Complementary Feature Extractions for Event Identification in Power Systems Using Multi-Channel Convolutional Neural Network

Energies ◽  
2021 ◽  
Vol 14 (15) ◽  
pp. 4446
Author(s):  
Do-In Kim
Keyword(s):  
Neural Network ◽  
Power Systems ◽  
Convolutional Neural Network ◽  
Time Window ◽  
Transient Signal ◽  
Measurement Unit ◽  
Transient Phenomena ◽  
Event Identification ◽  
Starting Point ◽  
Fast Decision

This paper presents an event identification process in complementary feature extractions via convolutional neural network (CNN)-based event classification. The CNN is a suitable deep learning technique for addressing the two-dimensional power system data as it directly derives information from a measurement signal database instead of modeling transient phenomena, where the measured synchrophasor data in the power systems are allocated by time and space domains. The dynamic signatures in phasor measurement unit (PMU) signals are analyzed based on the starting point of the subtransient signals, as well as the fluctuation signature in the transient signal. For fast decision and protective operations, the use of narrow band time window is recommended to reduce the acquisition delay, where a wide time window provides high accuracy due to the use of large amounts of data. In this study, two separate data preprocessing methods and multichannel CNN structures are constructed to provide validation, as well as the fast decision in successive event conditions. The decision result includes information pertaining to various event types and locations based on various time delays for the protective operation. Finally, this work verifies the event identification method through a case study and analyzes the effects of successive events in addition to classification accuracy.

scholarly journals Evaluation of Power Insulator Detection Efficiency with the Use of Limited Training Dataset

2020 ◽  
Vol 10 (6) ◽  
pp. 2104
Author(s):  
Michał Tomaszewski ◽  
Paweł Michalski ◽  
Jakub Osuchowski
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Object Detection ◽  
Convolutional Neural Network ◽  
Deep Neural Networks ◽  
Detection Efficiency ◽  
Training Data ◽  
Training Dataset ◽  
Training Set ◽  
Convolutional Network

This article presents an analysis of the effectiveness of object detection in digital images with the application of a limited quantity of input. The possibility of using a limited set of learning data was achieved by developing a detailed scenario of the task, which strictly defined the conditions of detector operation in the considered case of a convolutional neural network. The described solution utilizes known architectures of deep neural networks in the process of learning and object detection. The article presents comparisons of results from detecting the most popular deep neural networks while maintaining a limited training set composed of a specific number of selected images from diagnostic video. The analyzed input material was recorded during an inspection flight conducted along high-voltage lines. The object detector was built for a power insulator. The main contribution of the presented papier is the evidence that a limited training set (in our case, just 60 training frames) could be used for object detection, assuming an outdoor scenario with low variability of environmental conditions. The decision of which network will generate the best result for such a limited training set is not a trivial task. Conducted research suggests that the deep neural networks will achieve different levels of effectiveness depending on the amount of training data. The most beneficial results were obtained for two convolutional neural networks: the faster region-convolutional neural network (faster R-CNN) and the region-based fully convolutional network (R-FCN). Faster R-CNN reached the highest AP (average precision) at a level of 0.8 for 60 frames. The R-FCN model gained a worse AP result; however, it can be noted that the relationship between the number of input samples and the obtained results has a significantly lower influence than in the case of other CNN models, which, in the authors’ assessment, is a desired feature in the case of a limited training set.

Sign in / Sign up

Export Citation Format

Share Document