1. Project Overview

This project focuses on AI-powered intelligent systems with three main technical applications:

High-quality surface inspection for non-planar industrial products (e.g., optical lenses, ski goggles) – detecting defects down to 1×1 pixel with 98% accuracy using models like VGG-16, ResNet, and YOLO.

Medical retinal imaging – enhancing blood vessel segmentation to help diagnose diabetic retinopathy, glaucoma, and age-related macular degeneration using improved U-Net-based models.

AI-enhanced temperature monitoring – estimating wire temperatures in data centers to prevent thermal-related failures using AI-driven filtering and noise suppression.

2. Scholarships & Costs

Scholarship amount: 15,000 NTD per student per month (New Taiwan Dollars).

Other costs: Participants should budget separately for airfare, accommodation, meals, health insurance, and personal expenses.

Schedule / Duration: Minimum duration of 2 months

3. Educational background & major: Engineering, computer science, AI, biomedical engineering, or related technical fields.

4. Language proficiency: English proficiency sufficient for research and lab communication is expected.

5. Source universities (Vietnam – strong collaboration):

University of Science and Technology – Da Nang University

Hanoi University of Science and Technology

Vietnam National University, Hanoi

Electric Power University

6. Motivation: Students should be eager to gain international research experience, enhance technical skills, and participate in cross-cultural academic activities.

7. Other Important Considerations

Hands-on research: Participants will work directly with experienced researchers on laboratory projects involving deep learning (VGG-16, ResNet, YOLO, U-Net), image processing, and AI-driven filtering.

Data & tools: Use of public retinal datasets (ARIA, HRF, DRIVE, DR-HAGIS) and custom multi-frame image acquisition setups.

Academic bridge: The program encourages talented students to consider studying in Taiwan for the long term and strengthens institutional collaborations between Taiwan and Vietnam.

8. Preparation recommended: Familiarity with Python, C++, deep learning frameworks (TensorFlow/PyTorch), and basic image processing.