Projects

A selected list of projects I’ve designed and built across personal, academic, and collaborative contexts.

From real-world projects and side experiments to course-driven work and research collaborations, these reflect my interest in building meaningful applications across full-stack development, applied AI, and product design.

Image-Text Co-Decomposition for Text-Supervised Semantic Segmentation (CVPR2024) ↗ ↖

Research Computer-Vision Deep-Learning Semantic-Segmentation Vision-Language-Models Pytorch Python MMCV

My undergraduate research project at VLLab, later accepted at CVPR2024. We developed an image-text co-decomposition method for segmenting objects from free-form text without any ground-truth masks.

SYNC - Crowdsourcing Platform for News Co-Editing

Crowdsourcing News-and-Media NLP Full-Stack Vue.js Node.js Javascript Firebase Python MongoDB

A collaborative news editing platform that leverages crowdsourcing and AI to reduce media bias, enabling readers to compare multiple perspectives and collectively create more balanced reporting.

Pocket Café: Your favorite cafés in your pocket!

Coffee Taiwan Node.js PostgreSQL LINE-API Vue.js Heroku AWS

Pocket Café: A LINE chatbot that helps Taiwan users find, save, and share great cafés nearby. Built with Node.js and PostgreSQL, it provides café ratings, amenities info, and location-based recommendations—perfect for coffee lovers seeking their next favorite spot.

Spotify Tracks Dataset Visualization

Data-Visualization Javascript D3.js

Spotify Tracks DataVis is an interactive web tool for exploring trends and insights from over 160,000 Spotify tracks. Built with JavaScript and D3.js, it visualizes attributes like genre, popularity, and audio features using bar charts, radar plots, histograms, and more.

Categorizing Chinese News Articles

Machine-Learning NLP News-and-Media Python Scikit-Learn

An experimental project exploring various machine learning approaches to classify Chinese news articles into six distinct topics.