Kuanwei Chen

AI Researcher | Computer Science @ National Central University

I'm passionate about Natural Language Processing and Computer Vision, currently pursuing my Master's in CS with a specialization in Artificial Intelligence. My research focuses on sign language translation, multi-biometric authentication, and deep learning architectures.

Email GitHub
Kuanwei Chen

Education

National Central University (NCU)

Now

Master in Computer Science — Artificial Intelligence

Taoyuan, Taiwan · Sep 2025

National Changhua University of Education (NCUE)

Bachelor in Electrical Engineering

Changhua, Taiwan · Sep 2021 – Jun 2025 · GPA: 3.8 / 4.0

Graduation Certificate

Experience

Project Assistant

Artificial Intelligence Application Laboratory, NCUE

Supervisor: Prof. Yang Wen Ren

  • Assisted in TEEP and IIPP international collaboration projects, working with 8 international interns
  • Contributed to research on multi-biometric authentication systems using deep learning
  • Developed action recognition systems for invigilation using computer vision
  • Led a team of international interns to publish research on voice and facial recognition
Certificate

Publications

Two-step Authentication: Multi-biometric System Using Voice and Facial Recognition

First Author

IET (Institution of Engineering and Technology), 2024

Designed CNN architecture based on VGG-16 with ResNet-inspired modifications in TensorFlow. Achieved 95.1% accuracy for face recognition and 98.9% accuracy for voice recognition with 3.4% equal error rate.

Paper GitHub ArXiv Certificate

HET5: Hierarchy-Based Embedding with T5 for Efficient Sign Language Translation

Pre-print

Pre-print Article

Developed a transformer architecture for sign language translation. Implemented patch embedding and window multi-head self-attention mechanisms, significantly reducing computational complexity while maintaining translation quality.

Projects

How2sign YouTube-ASL Dataset Preprocess

Built a data preprocessing pipeline for American Sign Language videos. Extracted and processed 85-point skeletal landmarks to NumPy format using MediaPipe.

Python MediaPipe
View on GitHub

Invigilation Using Lightweight Semantic Analysis

Utilized Swin Transformer for image segmentation and ShuffleNet for classification. Developed a monitoring system using computer vision for identifying and tracking examinees.

Python PyTorch Swin Transformer ShuffleNet
View on GitHub

Taiwan Stock Predictor

Crawl data from the Taiwan stock market and analyze through machine learning models for prediction and trend analysis.

Python Web Scraping Machine Learning
View on GitHub

Recognition System Using Deepspeaker and Dlib

An implementation of a verification system with CAPTCHAs using face and voice recognition for multi-factor biometric authentication.

Python Deepspeaker Dlib CAPTCHA
View on GitHub

Awesome Italian Brainrot

A curated collection cataloging all the Italian brainrot memes. Community-driven with 224+ stars.

Collection Community
View on GitHub

Skills

Languages

English (IELTS 6.5) Mandarin (Native) Hokkien (Native)

Programming

Python C/C++ HTML5 MATLAB

Frameworks & Libraries

PyTorch TensorFlow HuggingFace Scikit-learn NumPy Pandas MediaPipe

Architecture

CNN RNN Transformer VGG ResNet Swin Transformer

Tools

Git / GitHub Arduino Linux

Awards & Certifications

Micro-Program in Green Energy Technology

Electrical Engineering, National Changhua University of Education

Completed courses including Introduction to Green Energy Technology, Circuit Theory Experiments, Circuit Theory, and Special Topic for Industry.

Teaching Assistant — Linear Algebra

National Changhua University of Education

Created and delivered supplementary materials for 40+ students. Assisted professor with exam invigilation and grading.

Certificate