NoiseCare is an AI-powered infant cry recognition and translation system combining deep learning algorithms with smart hardware. The project is currently developing a cry classification model, a baby-side monitoring and soothing device, a parent-side notification interface (including wearable alerts), and a mobile application. The system identifies common emotional needs behind infant cries—such as hunger, fatigue, or discomfort—and provides real-time responses to assist caregivers. The solution is also designed to support hearing-impaired parents, offering inclusive access to early parenting technology.
How it is being developed
Building a multi-label infant cry dataset from real-world audio samples
Applying CNN + TSI algorithms for feature extraction from spectrograms
Training cry classification models using deep learning frameworks (e.g., PyTorch)
Developing embedded hardware prototypes for both baby-side and parent-side devices
Creating a cross-platform mobile app (based on Flutter + Firebase) for interaction and visualization
Expected outcome
The expected outcome is an integrated AI-powered infant cry recognition and response system, consisting of:
A baby-side device mounted on cribs or strollers to detect crying, classify emotional needs (e.g., hunger, discomfort), and play soothing sounds automatically
A wearable parent-side wristband that delivers real-time alerts through vibration and lights
A mobile app that visualizes cry classification results, parenting suggestions, and daily behavioral trends
A visually-accessible interface for hearing-impaired parents, translating cry signals into readable text and alerts
The final product will offer high accuracy, real-time interaction, and strong usability, enabling both commercial application and inclusive social impact.
Brief Introduction
NoiseCare is an AI-powered infant cry recognition and translation system combining deep learning algorithms with smart hardware. The project is currently developing a cry classification model, a baby-side monitoring and soothing device, a parent-side notification interface (including wearable alerts), and a mobile application. The system identifies common emotional needs behind infant cries—such as hunger, fatigue, or discomfort—and provides real-time responses to assist caregivers. The solution is also designed to support hearing-impaired parents, offering inclusive access to early parenting technology.
How it is being developed
Building a multi-label infant cry dataset from real-world audio samples
Applying CNN + TSI algorithms for feature extraction from spectrograms
Training cry classification models using deep learning frameworks (e.g., PyTorch)
Developing embedded hardware prototypes for both baby-side and parent-side devices
Creating a cross-platform mobile app (based on Flutter + Firebase) for interaction and visualization
Expected outcome
The expected outcome is an integrated AI-powered infant cry recognition and response system, consisting of:
A baby-side device mounted on cribs or strollers to detect crying, classify emotional needs (e.g., hunger, discomfort), and play soothing sounds automatically
A wearable parent-side wristband that delivers real-time alerts through vibration and lights
A mobile app that visualizes cry classification results, parenting suggestions, and daily behavioral trends
A visually-accessible interface for hearing-impaired parents, translating cry signals into readable text and alerts
The final product will offer high accuracy, real-time interaction, and strong usability, enabling both commercial application and inclusive social impact.