Sumeth Yuenyong

Dr. Sumeth Yuenyong


Research Focus & Metrics

Academic Career Length

16Years (2011-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
Verified
Journal Article
Open access
PDF file
KWKonlakorn Wongpatikaseree
Konlakorn Wongpatikaseree
Sattaya SingkulSattaya Singkul
Sattaya Singkul
Narit HnoohomNarit Hnoohom
Narit Hnoohom
SYSumeth Yuenyong
Sumeth Yuenyong
Language resources are the main factor in speech-emotion-recognition (SER)-based deep learning models. Thai is a low-resource language that has a smaller data size than high-resource languages such as German. This paper describes the framework of using a pretrained-model-based front-end and back-end network to adapt feature spaces from the speech recognition domain to the speech emotion classification domain. It consists of two parts: a speech recognition front-end network and a speech emotion recognition back-end network. For speech recognition, Wav2Vec2 is the state-of-the-art for high-resource languages, while XLSR is used for low-resource languages. Wav2Vec2 and XLSR have proposed generalized end-to-end learning for speech understanding based on the speech recognition domain as feature space representations from feature encoding. This is one reason why our front-end network was selected as Wav2Vec2 and XLSR for the pretrained model. The pre-trained Wav2Vec2 and XLSR are used for front-end networks and fine-tuned for specific languages using the Common Voice 7.0 dataset. Then, feature vectors of the front-end network are input for back-end networks; this includes convolution time reduction (CTR) and linear mean encoding transformation (LMET). Experiments using two different datasets show that our proposed framework can outperform the baselines in terms of unweighted and weighted accuracies.

16 May 2022 in IEEE Access
Verified
Journal Article
Open access
VSVirach Sornlertlamvanich
Virach Sornlertlamvanich
SYSumeth Yuenyong
Sumeth Yuenyong
The languages spoken in Asia share common morphological analysis errors in word segmentation which normally propagate to higher-level processing, i.e., part-of-speech (POS) tagging, syntactic parsing, word extraction, and named entity recognition (NER), as we discuss in this research. We introduce the Thai character cluster (TCC) to reduce the errors propagated from word segmentation and POS tagging by incorporating it into the character representation layer of bidirectional long short-term memory (BiLSTM) for NER. The initial NER model is created from the original THAI-NEST named-entity (NE) tagged corpus by applying the best performing BiLSTM-CNN-CRF model (the combination of BiLSTM, convolutional neural network (CNN), and conditional random field (CRF)) with the word, POS, and TCC embedding. We determine the errors and improve the consistency of the NE annotation through our holdout method by retraining the model with the corrected training set. After the iteration, the overall result of the annotation F1-score has been improved to reach 89.22%, which improves 16.21% from the model trained on the original corpus. The result of our iterative verification is a promising method for low resource language modeling. As a result, The NE silver standard corpus is newly generated for the Thai NER task, called Bangkok Data NE tagged Corpus (BKD). The consistency of annotation is checked and revised according to the improvement of the scope of NE detection by TCC which can recover the errors in word segmentation.

1 July 2025
Verified
Book Chapter
Closed access
PDF file
SYSumeth Yuenyong
Sumeth Yuenyong
KVKobkrit Viriyayudhakorn
Kobkrit Viriyayudhakorn
APApivadee Piyatumrong
Apivadee Piyatumrong
JJJillaphat Jaroenkantasima
Jillaphat Jaroenkantasima

1 May 2011
Verified
Conference Paper
Closed access
PPPipatthana Phatiwuttipat
Pipatthana Phatiwuttipat
WKWaree Kongprawechon
Waree Kongprawechon
KTKanokvate Tungpimolrut
Kanokvate Tungpimolrut
SYSumeth Yuenyong
Sumeth Yuenyong
The system of this study is aimed to support doctors with an analyzed Cardiac auscultation. Extract the information from the heart sound signal obtained from stethoscope mobbed on the robot arm control is used in the further signal processing. Due to time consuming and accuracy, Support Vector Machine is introduced to replace Neural Network for better performance. With the proposed technique, the high classification performances were achieved 96.4% accuracy for classifying normal and abnormal heart sound while shortening the training time almost 300%. As the result, multi-classifier was introduced for advance development. This paper shows the comparison work between Neural Network and Support Vector Machine.

1 November 2015
Verified
Conference Paper
Closed access
SYSumeth Yuenyong
Sumeth Yuenyong
Echo State Network (ESN) is a special type of neural network with a randomly generated structure called the reservoir. The performance of ESN is sensitive to the reservoir parameters, which have to be tuned for best performance. Tuning of the reservoir parameters using evolutionary algorithms can be slow and produce inconsistent results. In this paper, we present a simple method for generating reservoirs based on templates that makes the reservoir matrices deterministic with respect to the parameters. Compared with the traditional method where the reservoir matrices are random, tuning of the reservoir parameters with an evolutionary algorithm needs less time, less number of cost function evaluations, and produces more reliable results using the proposed method.

11 November 2022 in Big Data and Cognitive Computing
Verified
Journal Article
Open access
PDF file
SYSumeth Yuenyong
Sumeth Yuenyong
KWKonlakorn Wongpatikaseree
Konlakorn Wongpatikaseree
Due to the ubiquitous nature of CCTV cameras that record continuously, there is a large amount of video data that are unstructured. Often, when these recordings have to be reviewed, it is to look for a specific person that fits a certain description. Currently, this is achieved by manual inspection of the videos, which is both time-consuming and labor-intensive. While person description search is not a new topic, in this work, we made two contributions. First, we improve upon the existing state-of-the-art by proposing unsupervised finetuning on the language model that forms a main part of the text branch of person description search models. This led to higher recall values on the standard dataset. The second contribution is that we engineered a complete pipeline from video files to fast searchable objects. Due to the use of an approximate nearest neighbor search and some model optimizations, a person description search can be performed such that the result is available immediately when deployed on a standard PC with no GPU, allowing an interactive search. We demonstrated the effectiveness of the system on new data and showed that most people in the videos can be successfully discovered by the search.