Hi, I'm Kexin

PosePromptFX: LLM-powered 2D Pose-driven VFX Generation System

An LLM-powered system that generates 2D visual effects around human body based on user inputs.

Echo-FAS: Acoustic-based Face Anti-Spoofing 🏅

An acoustic face anti-spoofing system that uses a novel two-branch deep learning framework to detect spoofs by analyzing global and local frequency clues in audio signals.

M3FAS: An Accurate and Robust MultiModal Mobile Face Anti-Spoofing System

A deep-learning-based face anti-spoofing system that uses cameras, speakers, and microphones to detect attacks.

Hitchhiker's Guide to Property Purchase in Singapore

Visualizing and analyzing Singapore property data using data mining techniques.

Mood: A Multi-modal Song Emotion Recognition System

We compared different feature extraction approaches and deep learning architectures including CNN, LSTM and Transformer.

ThermEarhook: Investigating Spatial Thermal Haptic Feedback on the Auricular Skin Area

An ear-worn device that provides customized hot and cold thermal feedback patterns for communication.

CREATOR: Sketch-based 3D Animation System with Augmented Reality

A mobile AR application that converts 2D sketches into 3D models that can be animated in an augmented reality environment.

Sketching 3D animation for Real Indoor Environment

An interactive system for creating 3D avatar animation easily given a scanned 3D environment.