Sleep Audio Analysis: Snore & Respiration Detection
Budget: €750 – €1,500 EUR
Project Title
AI-Powered Audio Analysis Software for Sleep Respiratory Detection & Snoring Classification
Project Description
I am looking for an expert in Audio Signal Processing and Machine Learning to develop a specialized software tool for analyzing 8-hour sleep audio recordings. The system must process audio data to identify, categorize, and log specific respiratory events using provided training datasets.
1. Detection & Classification Requirements
The software must accurately detect and quantify the following:
• Normal Breathing: Total event count.
• Labored Breathing: Total event count.
• Snoring Intensity Levels:
o Low Snoring: 10 – 30 dB
o Moderate Snoring: 30 – 50 dB
o Heavy Snoring: 50 – 80 dB
o Explosive Snoring: > 80 dB
• Apneic Events:
o Apnea: Total cessation of breath for more than 10 seconds.
o Hypopnea: At least 50% reduction in breathing amplitude.
2. Features & Analytics
• Manual Validation Module: An interactive interface that allows the user to manually review, confirm, or reject detected events (Manual Override) for data verification.
• Statistical Reporting: Automated calculation of average duration and maximum duration for both Apnea and Hypopnea events.
• Automated Visualization: A dashboard that generates a color-coded Pie Chart showing the percentage (%) distribution of all detected categories.
3. Data & Training
I will provide an extensive, high-quality dataset containing labeled samples of all categories (Snoring levels, Apnea, Normal, and Labored breathing) for model training and testing.
AI-Powered Audio Analysis Software for Sleep Respiratory Detection & Snoring Classification
Project Description
I am looking for an expert in Audio Signal Processing and Machine Learning to develop a specialized software tool for analyzing 8-hour sleep audio recordings. The system must process audio data to identify, categorize, and log specific respiratory events using provided training datasets.
1. Detection & Classification Requirements
The software must accurately detect and quantify the following:
• Normal Breathing: Total event count.
• Labored Breathing: Total event count.
• Snoring Intensity Levels:
o Low Snoring: 10 – 30 dB
o Moderate Snoring: 30 – 50 dB
o Heavy Snoring: 50 – 80 dB
o Explosive Snoring: > 80 dB
• Apneic Events:
o Apnea: Total cessation of breath for more than 10 seconds.
o Hypopnea: At least 50% reduction in breathing amplitude.
2. Features & Analytics
• Manual Validation Module: An interactive interface that allows the user to manually review, confirm, or reject detected events (Manual Override) for data verification.
• Statistical Reporting: Automated calculation of average duration and maximum duration for both Apnea and Hypopnea events.
• Automated Visualization: A dashboard that generates a color-coded Pie Chart showing the percentage (%) distribution of all detected categories.
3. Data & Training
I will provide an extensive, high-quality dataset containing labeled samples of all categories (Snoring levels, Apnea, Normal, and Labored breathing) for model training and testing.