Aksara jawa text detection in scene images using convolutional neural network

Afakh, Muhammad Labiyb; Risnumawan, Anhar; Anggraeni, Martianda Erste; Tamara, Mohamad Nasyir; Ningrum, Endah Suryawati

doi:10.1109/kcic.2017.8228567

Cited by 15 publications

(7 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Gaussian filter used is 5x5 with a grayscale image binary image and the variable threshold value. The pixel value will be one of the input pixel values that is more than the threshold value and will be 0 if the input pixel value is below the threshold value (2) The CNN architecture that is built consists of three convolution layers accompanied by max-pooling operations on each layer. Each layer in the convolution layer consists of a filter of a certain size.…”

Section: Cnn Modellingmentioning

confidence: 99%

See 1 more Smart Citation

Convolutional Neural Network Method for Classification of Syllables in Javanese Script

Fauziah

Aprilianta²,

Rustamaji³

2021

Int. J. Artif. Intell. Robot.

View full text Add to dashboard Cite

Javanese script is one of the languages which are a typical Javanese culture. Javanese script is seen in its use in writing the name of a particular agency or location that has historical and tourism value. The use of Javanese script in public places makes the existence of this script seen by many people, not only by the Javanese people. Some of them have difficulty recognizing the Javanese characters they encounter. One method of pattern recognition and image processing is Convolutional Neural Network (CNN). CNN is a method that uses convolution operations in performing feature extraction on images as a basis for classification. The process consists of initial data processing, classification, and syllable formation. The classification consists of 48 classes covering Javanese script types, namely basic letters (Carakan) and voice-modifying scripts (Sandhangan). It is tested with multi-class confusion matrix scenarios to determine the accuracy, precision, and recall of the built CNN model. The CNN architecture consists of three convolution layers with max-pooling operations. The training configuration includes a learning rate of 0.0001, and the number of filters for each convolution layer is 32, 64, and 128 filters. The dropout value used is 0.5, and the number of neurons in the fully-connected layer is 1,024 neurons. The average performance value of accuracy reached 87.65%, the average precision value was 88.01%, and the average recall value was 87.70%.

show abstract

Section: Cnn Modellingmentioning

confidence: 99%

“…Research related to the detection of the presence of Javanese characters was carried out using the Convolutional Neural Network (CNN) method on objects in the form of images [2]. The research carried out the detection process by forming a bounding box around a series of Javanese scripts encountered.…”

Section: Introductionmentioning

confidence: 99%

Convolutional Neural Network Method for Classification of Syllables in Javanese Script

Fauziah

Aprilianta²,

Rustamaji³

2021

Int. J. Artif. Intell. Robot.

View full text Add to dashboard Cite

show abstract

“…Javanese tribble is the largest ethnic group and Javanese is the most widely spoken regional language in Indonesia. The Javanese language has its own traditional letter called Aksara Jawa or Javanese script [1]. Javanese script is a historic Javanese character that has been used by the Mataram Kingdom since the 17th century.…”

Section: Introductionmentioning

confidence: 99%

Hybrid convolutional neural networks-support vector machine classifier with dropout for Javanese character recognition

2023

View full text Add to dashboard Cite

This research paper explores the hybrid models for Javanese character recognition using 15600 characters gathered from digital and handwritten sources. The hybrid model combines the merit of deep learning using convolutional neural networks (CNN) to involve feature extraction and a machine learning classifier using support vector machine (SVM). The dropout layer also manages overfitting problems and enhances training accuracy. For evaluation purposes, we also compared CNN models with three different architectures with multilayer perceptron (MLP) models with one and two hidden layer(s). In this research, we evaluated three variants of CNN architectures and the hybrid CNN-SVM models on both the accuracy of classification and training time. The experimental outcomes showed that the classification performances of all CNN models outperform the classification performances of both MLP models. The highest testing accuracy for basic CNN is 94.2% when using model 3 CNN. The increment of hidden layers to the MLP model just slightly enhances the accuracy. Furthermore, the hybrid model gained the highest accuracy result of 98.35% for classifying the testing data when combining model 3 CNN with the SVM classifier. We get that the hybrid CNN-SVM model can enhance the accuracy results in the Javanese characters recognition.

show abstract

“…Moreover, this approach can simultaneously recognize multiple objects. The detection and recognition of letters are not a new thing; there is research that has been done using different methods, such as in the research [2], which uses the Projection Profile Segmentation and Nearest Centroid Classifier; research [3], which uses Nested Multi-Layer Perceptron network with Modified Direction Feature; and research [4], which uses Convolution Neural Network (CNN).…”

Section: Introductionmentioning

confidence: 99%

Javanese Script Letter Detection Using Faster R-CNN

Faishal,

Sulistiyo,

Ihsan

2023

IJAIDM

View full text Add to dashboard Cite

The Javanese script is now rarely used, and some people no longer recognize it. The construction of a Javanese script recognition system based on digital image processing is one of its preservation efforts. This study proposes a model capable of detecting and recognizing Javanese characters using Faster R-CNN to help people who are not familiar with the Javanese script. Faster R-CNN was chosen because it does not require additional processing compared to the previous method and Faster R-CNN has better accuracy and the ability to detect small objects. Faster R-CNN shows good results in text detection, but the use of Faster R-CNN in detecting Javanese script has not been found which makes its performance unknown, so this study will show how Faster R-CNN performs in detecting Javanese script. In this study, Faster R-CNN was able to show good performance by obtaining mean average precision (mAP) values up to 0.8381, accuracy up to 96.31%, precision up to 96.53%, recall up to 96.38 %, and F1-Score up to 96.41%. These results indicate that Faster R-CNN has better results than the previous method and can detect Javanese characters well.

show abstract

Aksara jawa text detection in scene images using convolutional neural network

Cited by 15 publications

References 17 publications

Convolutional Neural Network Method for Classification of Syllables in Javanese Script

Convolutional Neural Network Method for Classification of Syllables in Javanese Script

Hybrid convolutional neural networks-support vector machine classifier with dropout for Javanese character recognition

Javanese Script Letter Detection Using Faster R-CNN

Contact Info

Product

Resources

About