maraco-api 0.1.0

Marine Acoustic Classification API - CPU-optimized marine sound classification

MarACO - Marine Acoustic Classification API

A Python package for marine acoustic sound classification optimized for CPU usage with balanced speed and accuracy.

Features

• Fast Classification: <200ms per 5-second audio clip on CPU
• High Accuracy: >90% accuracy on marine acoustic sounds
• Noise Reduction: Built-in noise reduction and audio preprocessing
• Multiple Classes: Fin Whale, Humpback Whale, Right Whale, Sonar, Vessels, Explosions, Physical Noise
• Easy Integration: Simple API for quick integration into your projects
• CPU Optimized: Designed specifically for CPU-only environments

Installation

pip install maraco-api

Quick Start

from maraco import MarineAcousticClassifier

# Initialize classifier
classifier = MarineAcousticClassifier()

# Classify a single audio file
result = classifier.predict("audio_file.wav")
print(f"Class: {result['class']}, Confidence: {result['confidence']:.2f}")

# Batch processing
results = classifier.predict_batch(["file1.wav", "file2.wav"])
for result in results:
    print(f"File: {result['filename']}, Class: {result['class']}")

Supported Audio Formats

• WAV, AIFF, MP3, FLAC
• Sample rates: 600Hz - 48kHz (automatically resampled)
• Duration: 1-30 seconds (optimal: 2-5 seconds)

Performance

• Processing Speed: <200ms per clip
• Memory Usage: <500MB RAM
• Model Size: <50MB
• Accuracy: >90% on test data

License

MIT License