Sukmawati Nur Endah
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Sukmawati Nur Endah has a total of 12 co-authors. Here are the top co-authors by number of shared publications:
- Retno Kusumaningrum
Universitas Diponegoro
2 shared publications Anwar Petrus F Naiborhu
Diponegoro University, Semarang, Indonesia
1 shared publication- Anwar Petrus F Noiborhu
Universitas Diponegoro
1 shared publication - Linggar Maretva Cendani
Diponegoro University
1 shared publication - Priyo Sidik Sasongko
Universitas Diponegoro
1 shared publication - Rismiyati Rismiyati
Universitas Diponegoro
1 shared publication
Latest Publications
- Open Access
- 6 October 2025
- Published by Engineering, Technology & Applied Science Research in Engineering, Technology & Applied Science Research
- Vol. 15 (5) , 28095-28101
Enhancing Low-Resource Dialectal ASR in Indonesian Using Speech-Transformer Models and Data Augmentation
One of the main challenges faced by researchers in speech recognition is the limitation of data, especially for low-resource languages. A common strategy to improve a model's performance is to expand the data space through data augmentation techniques. Data augmentation has proven effective in increasing the amount of training data and reducing the mismatch between training and testing data. Furthermore, data augmentation is essential for improving the performance of deep neural networks by mitigating overfitting and enhancing the models' generalization capabilities. This study compares the impact of several standard augmentation techniques applied to low-resource dialect speech (time stretching, pitch shifting, noise addition, and gain) on speech recognition performance using a Speech-Transformer architecture. The dataset used consists of Indonesian dialectal speech. The results indicate that the average accuracy improvement in recognition was 57.6%, 57.9%, and 59.3% for Character Error Rate (CER), Word Error Rate (WER), and Sentence Error Rate (SER), respectively, compared to speech recognition without any data augmentation.
- 6 October 2025
- Published by Engineering, Technology & Applied Science Research in Engineering, Technology & Applied Science Research
- Vol. 15 (5) , 28095-28101
Open AccessEnhancing Low-Resource Dialectal ASR in Indonesian Using Speech-Transformer Models and Data Augmentation
One of the main challenges faced by researchers in speech recognition is the limitation of data, especially for low-resource languages. A common strategy to improve a model's performance is to expand the data space through data augmentation techniques. Data augmentation has proven effective in increasing the amount of training data and reducing the mismatch between training and testing data. Furthermore, data augmentation is essential for improving the performance of deep neural networks by mitigating overfitting and enhancing the models' generalization capabilities. This study compares the impact of several standard augmentation techniques applied to low-resource dialect speech (time stretching, pitch shifting, noise addition, and gain) on speech recognition performance using a Speech-Transformer architecture. The dataset used consists of Indonesian dialectal speech. The results indicate that the average accuracy improvement in recognition was 57.6%, 57.9%, and 59.3% for Character Error Rate (CER), Word Error Rate (WER), and Sentence Error Rate (SER), respectively, compared to speech recognition without any data augmentation.
- Open Access
- 1 June 2025
- Published by Universitas Ahmad Dahlan in TELKOMNIKA (Telecommunication Computing Electronics and Control)
- Vol. 23 (3) , 807-815
Indonesian continuous speech recognition optimization with convolution bidirectional long short-term memory architecture
Indonesian continuous speech recognition optimization with convolution bidirectional long short-term memory architecture
- 1 June 2025
- Published by Universitas Ahmad Dahlan in TELKOMNIKA (Telecommunication Computing Electronics and Control)
- Vol. 23 (3) , 807-815
Open AccessIndonesian continuous speech recognition optimization with convolution bidirectional long short-term memory architecture
Indonesian continuous speech recognition optimization with convolution bidirectional long short-term memory architecture
- 1 September 2022
- Published by Springer Nature
- p. 106-122
Aspect-Based Sentiment Analysis of Indonesian-Language Hotel Reviews Using Long Short-Term Memory with an Attention Mechanism
- 1 September 2022
- Published by Springer Nature
- p. 106-122
Aspect-Based Sentiment Analysis of Indonesian-Language Hotel Reviews Using Long Short-Term Memory with an Attention Mechanism
- 24 November 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 122-127
Indonesian Continuous Speech Recognition Using CNN and Bidirectional LSTM
Speech recognition is the machine's ability to recognize words based on speech. Utilization of speech recognition technology into a real form of human-machine interaction. This allows humans to control smart devices more easily. However, the variability of speech such as dialect or accent, vocabulary size, type of recognition, speech speed, environmental noise, and type of microphone will affect the speech recognition rate. In recent years, deep learning approaches such as CNN and BLSTM have been widely used and have provided significant recognition improvements. Inspired by the advantages of CNN in exploiting local interspectral correlations and capturing frequency changes in speech signals and BLSTM in learning the temporal context, this study uses a hybrid CNN and BLSTM models for speech recognition with CTC as decoder. This study uses continuous speech data in Indonesian with five different dialects, namely, Balinese, Bataknese, Javanese, Minangese and Sundanese. There are four test scenarios that carried out sequentially to improve speech recognition performance including layer structure, use of dropouts, number of filters and units, and types of input features. The first three scenarios only use 13 coefficients MFCC without deltas features as the input feature. The results showed that the combination of 2 CNN layers with 64 filters and 2 BLSTM layers with 128 units and the application of a dropout with rate of 0.2 on all hidden layers achieve WER of 37.31%. And the addition of delta and double delta features can reduce the recognition error and achieve WER 10.80%.
- 24 November 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 122-127
Indonesian Continuous Speech Recognition Using CNN and Bidirectional LSTM
Speech recognition is the machine's ability to recognize words based on speech. Utilization of speech recognition technology into a real form of human-machine interaction. This allows humans to control smart devices more easily. However, the variability of speech such as dialect or accent, vocabulary size, type of recognition, speech speed, environmental noise, and type of microphone will affect the speech recognition rate. In recent years, deep learning approaches such as CNN and BLSTM have been widely used and have provided significant recognition improvements. Inspired by the advantages of CNN in exploiting local interspectral correlations and capturing frequency changes in speech signals and BLSTM in learning the temporal context, this study uses a hybrid CNN and BLSTM models for speech recognition with CTC as decoder. This study uses continuous speech data in Indonesian with five different dialects, namely, Balinese, Bataknese, Javanese, Minangese and Sundanese. There are four test scenarios that carried out sequentially to improve speech recognition performance including layer structure, use of dropouts, number of filters and units, and types of input features. The first three scenarios only use 13 coefficients MFCC without deltas features as the input feature. The results showed that the combination of 2 CNN layers with 64 filters and 2 BLSTM layers with 128 units and the application of a dropout with rate of 0.2 on all hidden layers achieve WER of 37.31%. And the addition of delta and double delta features can reduce the recognition error and achieve WER 10.80%.
- 1 October 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Usability Testing for Batik 4.0: A Web Application for Generating 3D Batik Semarangan
Revolution of the industry has now reached the industrial 4.0 era. The development and progression of the industrial revolution have meant that a country needs to have superior products as added value in order to safeguard from the threat of global competition. Indonesia has a product advantage compared to other countries in the form of Indonesian Batik. Indonesian Batik producers are spread across several cities in Indonesia, each having their own unique characteristics, one of which is Semarang. The batik in Semarang is called Batik Semarangan. To keep abreast of the changes brought about by the industrial revolution in Indonesia, a web-based application was developed to produce 3D batik motifs using a triangle mesh. This application is called Batik 4.0. It was important to measure and test the performance and efficiency of the application in order to meet existing objectives and functional requirements of the application. Usability testing was the chosen approach used to facilitate this process and to assess how well the users could learn and use the application. Batik 4.0 was applied on comparison in two platforms in the form of a one-page website and multipage website. This study describes the general description and function of Batik 4.0, and evaluates the usability of this application to produce 3D batik motifs using a triangle mesh in the application of onepage and multipage. The results will be compared to find out which one is better for the application of Batik 4.0, whether on page or multipage. The testing technique was in the form of a survey, of which was questionnaire-based and performance-based. For compare the usability of onepage and multipage, this research use the usability attributes included learnability, satisfaction, effectiveness and efficiency. The results from the testing indicated that the multipage website to generate 3D batik motifs was better than onepage website with learnability value (81.71%), satisfaction (77.43%), effectiveness (94.64%) and efficiency (92.23 %) results.
- 1 October 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Usability Testing for Batik 4.0: A Web Application for Generating 3D Batik Semarangan
Revolution of the industry has now reached the industrial 4.0 era. The development and progression of the industrial revolution have meant that a country needs to have superior products as added value in order to safeguard from the threat of global competition. Indonesia has a product advantage compared to other countries in the form of Indonesian Batik. Indonesian Batik producers are spread across several cities in Indonesia, each having their own unique characteristics, one of which is Semarang. The batik in Semarang is called Batik Semarangan. To keep abreast of the changes brought about by the industrial revolution in Indonesia, a web-based application was developed to produce 3D batik motifs using a triangle mesh. This application is called Batik 4.0. It was important to measure and test the performance and efficiency of the application in order to meet existing objectives and functional requirements of the application. Usability testing was the chosen approach used to facilitate this process and to assess how well the users could learn and use the application. Batik 4.0 was applied on comparison in two platforms in the form of a one-page website and multipage website. This study describes the general description and function of Batik 4.0, and evaluates the usability of this application to produce 3D batik motifs using a triangle mesh in the application of onepage and multipage. The results will be compared to find out which one is better for the application of Batik 4.0, whether on page or multipage. The testing technique was in the form of a survey, of which was questionnaire-based and performance-based. For compare the usability of onepage and multipage, this research use the usability attributes included learnability, satisfaction, effectiveness and efficiency. The results from the testing indicated that the multipage website to generate 3D batik motifs was better than onepage website with learnability value (81.71%), satisfaction (77.43%), effectiveness (94.64%) and efficiency (92.23 %) results.