U-Time: A Fully Convolutional Network for Time Series Segmentation

Official code for "[The Fatigue Status Feature of Bicycle Movement Based on Deep Learning and Signal Processing Technology]". This work presents a non-invasive, vision-based method using Keypoint R…

This project contains the code used in my thesis, "IAmMuse: A signal-processing-based method to estimate arm positions with mmWave radar".

Open-source spatial intelligence for live theater. Fuses IR tracking + mmWave radar + AI to provide real-time performer positioning, haptic mark confirmation, crew coordination, and lighting sugges…

ESP32-C3 firmware for controlling ATX-powered embedded systems via a PC power button, with optional ACPI shutdown and deep sleep wakeup support. Ideal for NAS kits and other boards needing smart po…

FastAPI + Streamlit stack that chains OCR→Gene→Medical→Intervention→Dialogue services with shared pipelines, local logs, and docker-compose orchestration for health-report analysis.

PhyCL-Net: physics-inspired lightweight fall-detection on wearable inertial signals, emphasizing dynamic kernels, attention, and contrastive training without MSPA.

Privacy-first Smart Health Companion for elderly care - Edge AI system with mmWave radar fall detection, local LLM/RAG on RK3588, featuring ZeroMQ IPC and real-time monitoring console

专为 Rockchip NPU （RK3588） 优化的轻量级人脸检测库， 支持实时人脸框检测与5个关键点定位。

🖼️ Implement FPGA image processing modules in Verilog, featuring basics to advanced vision algorithms, all equipped with testbenches for easy validation.

AI-powered yoga mat with Bluetooth connectivity, pressure sensors, LED panels, and inbuilt speakers. Customizes workouts based on user data like weight, BMI, and fitness level. Provides real-time …

MuscleMap: An Open-Source, Community-Supported Consortium for Whole-Body Quantitative MRI of Muscle

Python implementations from my LifeStep biomechanics internship, including: 1) Processing and analyzing squat.mat and cutting.mat data, 2) Extracting joint variables (knee angle, hip angle, etc.) f…

SleepLiteCNN: Energy-Efficient Sleep Apnea Subtype Classification with 1-Second Resolution Using Single-Lead ECG

The dataset and code for the paper "Wearable intelligent throat enables natural speech in stroke patients with dysarthria"

Datasets and Codes for "A deep learning-enabled smart garment for versatile and accurate sleep conditions monitoring in daily life"

The supporting code and data for "AI-driven multimodal smart home platform for continuous monitoring and assistance of post-stroke patients"

Smart Toothbrush

A Python package for the analysis of biopsychological data.

Library for HRV data analysis based on measurements from Polar wearable devices

a portable biomedical device capable of acquiring, processing, and visualizing real-time biosignals(ECG, EEG, EMG, and EDA).

Tyee: A Unified, Modular, and Fully-Integrated Configurable Toolkit for Intelligent Physiological Health Care