Medical & Clinical Vision
Medical image classification, segmentation, and decision-support research across X-ray, MRI, CT, ultrasound, retinal imaging, and dermatology.
- Medical imaging
- Transfer learning
- Segmentation
- Clinical evaluation
RESEARCH AGENDA
Reliable visual intelligence under real-world constraints.
My research asks how computer vision can remain accurate while becoming more efficient, interpretable, and useful in settings where data, computation, and trust are limited.
Research constellation
Four connected research pillars
Hover or focus on a node to see how each pillar connects to projects, publications, and educational practice.
01
Medical & Clinical Vision
Medical image classification, segmentation, and decision-support research across X-ray, MRI, CT, ultrasound, retinal imaging, and dermatology.
Efficient Vision Models
Feature-space efficiency, lightweight architectures, dimensionality reduction, and model design for constrained computational settings.
- Efficient CNNs
- Feature selection
- Multi-scale learning
- Model compression
Explainable & Reliable AI
Attention mechanisms, visual explanations, evaluation beyond accuracy, and responsible framing for high-stakes use.
- Attention
- Interpretability
- Robust evaluation
- Responsible use
AI Education & Human Capability
Interactive demonstrations, teacher development, student mentoring, and ways to make AI concepts observable and teachable.
- AI literacy
- Learning tools
- Teacher development
- Student mentoring
From comparative modelling to a coherent research horizon
From comparative modelling to a coherent research horizon
Foundation
Comparative machine learning and transfer-learning experiments across health, agriculture, and visual-recognition tasks.
Concentration
Medical imaging, visual feature extraction, attention mechanisms, and efficiency-oriented architecture studies.
Current line
Funded projects on paediatric pneumonia, skin-image feature spaces, and collaborative health AI.
Research horizon
Reliable, efficient, and explainable computer vision under limited and heterogeneous data.
Research horizon
Research horizon
Reliable and efficient computer vision for limited, heterogeneous, and high-stakes data, connecting lightweight multi-scale models, explanation mechanisms, and rigorous real-world evaluation.
ReliableEfficientExplainableReal-world
Methods and technical capabilities
Methods and technical capabilities
CNN and transfer learningVision attention mechanismsImage classification and segmentationFeature extraction and dimensionality reductionModel benchmarking and ablation studiesClassical machine learning and ensemble methodsBrowser-based computer vision demonstrationsResponsible-use and limitation framing
Selected publications
Selected publications
2026Conference Paper
Does Explicit Reasoning Improve Translation? Benchmarking Chain-of-Thought vs. In-Context Learning for Indonesian-English Language Pairs
IEEE conference proceedings
Google Scholar2025Conference Paper
CBAM-Enhanced Classification Head for VGG16-Based Pneumonia Detection
Conference proceedings
Google Scholar2025Journal Article
Integrating VGG Re-trained Feature Extraction with Machine Learning for Knee Osteoarthritis Severity Levels Detection Using X-ray Images
ITEGAM-JETIA
Google Scholar2025Proceedings Article
SE-VGG16: Enhanced classification of skin cancer lesions using VGG16 backbone with squeeze-and-excitation blocks
Procedia Computer Science
Google Scholar2024Journal Article
Color and Attention for U: Modified Multi Attention U-Net for a Better Image Colorization
JOIV: International Journal on Informatics Visualization
Google Scholar2024Journal Article
Overcoming Overfitting in CNN Models for Potato Disease Classification Using Data Augmentation
EMACS Journal
Google ScholarContact
Start with the intended outcome.
For research collaboration, laboratory discussion, invited speaking, educator development, academic review, or applied AI prototyping.