Nawarerk Chalarak

Dr. Nawarerk Chalarak

orcid

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

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

Representative Publications

View All Publications
15 June 2025 in Journal of Current Science and Technology
Verified
Journal Article
Open access
PDF file
TTTanatorn Tanantong
Tanatorn Tanantong
NCNawarerk Chalarak
Nawarerk Chalarak
SJSumet Jirattisak
Sumet Jirattisak
KTKitiya Tanantong
Kitiya Tanantong
KSKrittakom Srijiranon
Krittakom Srijiranon
Psoriasis is a chronic skin disease with significant global and regional impacts, including in Thailand, where its burden is compounded by diagnostic challenges and limited dermatological resources. Psoriasis was selected for this study because it develops in distinct phases, requiring ongoing monitoring and treatment. The distribution of skin lesions plays a crucial role in disease identification and assessment, making it an essential factor for AI-based analysis. The development of AI-based diagnostic tools offers a potential solution. However, there is no publicly available skin disease dataset in Thailand, and image annotation is a challenging and time-consuming task for dermatologists. This scarcity of annotated datasets remains a critical barrier to AI development. This study utilizes the Dermnet dataset and enhances it through the application of image augmentation and style transfer techniques to generate a more diverse and representative dataset, particularly reflecting Thai skin tones. It also evaluates how augmentation techniques affect AI performance in psoriasis classification. The results showed that augmentation significantly enhanced model performance, with EfficientNetB4 achieving the highest accuracy (93.00%) and sensitivity (91.19%). Style transfer emerged as a valuable technique, enabling the creation of skin tone representative datasets that improved model generalizability. These findings align with existing literature. They demonstrate that augmentation techniques can overcome data limitations and enhance model robustness. This study introduces a novel use of style transfer techniques. These are applied to generate augmented datasets that represent Thai skin tones, addressing a critical gap in publicly available dermatology data. By enhancing dataset diversity, style transfer significantly improves the generalizability and accuracy of AI-based psoriasis classification models. These advancements have important implications for clinical practice. They are especially relevant in Thailand and other resource-limited regions, where AI-assisted diagnostics can improve dermatological care access and effectiveness.

28 September 2021 in International Journal of Innovation and Technology Management
Verified
Journal Article
Closed access
PDF file
NCNawarerk Chalarak
Nawarerk Chalarak
YSYasuo Sasaki
Yasuo Sasaki
NUNaoshi Uchihira
Naoshi Uchihira

12 August 2022 in International Journal of Innovation and Technology Management
Verified
Journal Article
Closed access
PDF file
NCNawarerk Chalarak
Nawarerk Chalarak
NUNaoshi Uchihira
Naoshi Uchihira
NSNao Sato
Nao Sato

27 July 2026 in BioMedInformatics
Verified
Journal Article
Open access
PDF file
TTTanatorn Tanantong
Tanatorn Tanantong
KKKanyarat Kanchanaphayak
Kanyarat Kanchanaphayak
NCNittaya Chemkomnerd
Nittaya Chemkomnerd
NCNawarerk Chalarak
Nawarerk Chalarak
KSKrittakom Srijiranon
Krittakom Srijiranon
PSPokpong Songmuang
Pokpong Songmuang
Accurate classification of cells from body fluid specimens can support cytological analysis, but microscopic images often present challenges such as low contrast, unclear boundaries, staining variation, class imbalance, and overlapping morphology. This study evaluated the impact of image enhancement techniques on deep learning-based multi-class classification of body fluid cells. The dataset comprised 7071 microscopic images from Srinagarind Hospital, Thailand, annotated by expert technologists. After preprocessing and cell extraction, 22,062 single-cell images across 13 cell types were obtained. A 70:30 train–test split was used, with augmentation and random undersampling applied only to the training set. Nine individual enhancement filters and five combined settings were tested using MobileNetV3, DenseNet121, ResNet50, and EfficientNetB3. Performance was measured using weighted accuracy, precision, recall, and F1-score. DenseNet121 with Edge Enhance achieved the best individual-filter performance (accuracy 83.36%, F1-score 83.08%). For combined settings, DenseNet121 with CLAHE followed by Detail performed best (accuracy 83.01%, F1-score 82.54%), though it did not surpass the top individual filter. Multi-step enhancement provided limited additional benefit. Class-wise results showed strong performance for distinct cell types, while macrophages, monocytes, and lymphocytes remained challenging. Grad-CAM visualization further indicated that model attention was generally concentrated on relevant cellular regions, including nuclei, cytoplasm, and cell boundaries. Overall, selected enhancement techniques offer modest improvements, but effectiveness depends on model architecture and data characteristics. As this study used a single-institution dataset without external validation, the approach should be considered an assistive framework rather than a clinically validated diagnostic system.


18 August 2026 in Digital
Verified
Journal Article
Open access
PDF file
KSKrittakom Srijiranon
Krittakom Srijiranon
TTTanatorn Tanantong
Tanatorn Tanantong
NKNattanon Keeratiwattapong
Nattanon Keeratiwattapong
NCNawarerk Chalarak
Nawarerk Chalarak
USUsanut Sangtongdee
Usanut Sangtongdee
Crime offense descriptions are often recorded as unstructured text, making large-scale analysis and categorization difficult. This study proposes a semantic clustering framework for Thai crime offense descriptions using sentence embeddings, dimensionality reduction, and unsupervised clustering. Two datasets were obtained from Thonglor Metropolitan Police Station and Mueang Nonthaburi Police Station, Thailand. After preprocessing, the datasets contained 962 and 902 unique offense descriptions, respectively. Each description was transformed into a 768-dimensional embedding using SimCSE-PhayaThaiBERT. The embeddings were represented in Principal Component Analysis (PCA) Space and Uniform Manifold Approximation and Projection (UMAP) Space and clustered using K-Means, DBSCAN, HDBSCAN, and OPTICS. The results showed that UMAP Space generally provided more useful clustering results than PCA Space. Although DBSCAN achieved the highest internal clustering scores, it classified most records as noise. In contrast, HDBSCAN provided a more balanced result by maintaining strong clustering quality while retaining more records for interpretation. Qualitative analysis showed that the discovered clusters corresponded to meaningful offense categories. The proposed framework can support exploratory analysis of Thai crime records without requiring manually labeled data.