I’m building a lean proof-of-concept that flags sleep-apnea events in children under ten by streaming SpO₂, ECG, and snoring audio from a wearable device. Everything will run in Python. Here’s the help I need: • A compact Region-based CNN that ingests the three synchronized signals and returns per-window apnea probabilities. • Lightweight, real-time preprocessing for SpO₂, ECG, and audio (basic filtering, resampling, and normalization only—no heavyweight feature engineering). • Simple publishing code that pushes a mobile-app notification the moment an apnea episode is detected (Firebase Cloud Messaging or any equally light SDK is fine). • The publicly available single-subject datasets you train on, plus direct links so I can review or replicate your work. • Well-commented source code with brief inline explanations and a short README describing setup, model architecture, and how to run an end-to-end demo. I want all the codes which u have used(each and very small code) Keep the solution minimal so it runs comfortably on modest hardware; I’m looking for a functional prototype, not a production-grade suite.