مشروع سنة دراسية

Decoding Baby Crying Using Audio Signal Processing and Deep Learning

Menna Ahmed Lead: Audio Data Pipeline & CRNN Model Development

My Core Responsibilities (Menna Ahmed)

•

Full ownership of the entire audio dataset (collection, cleaning, annotation)

•

Built the complete preprocessing and feature extraction pipeline

•

Designed and trained the top-performing CRNN model (highest accuracy in the entire project)

•

Extracted Mel-Spectrograms and MFCCs that became the foundation for all models

Project Summary An end-to-end AI system that classifies infant cries into 8 different needs in real time: Hunger · Belly Pain · Burping · Discomfort · Tired · Cold/Hot · Lonely · Scared