File size: 5,777 Bytes
1bb6efc | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 | """
Baby Cry AI - Dataset Explorer
Step 2: Explore and analyze the available datasets
"""
import os
import librosa
import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
from pathlib import Path
import json
class BabyCryDataExplorer:
def __init__(self, data_path="data"):
self.data_path = Path(data_path)
self.categories = []
self.file_info = {}
def explore_dataset(self):
"""Explore the structure and content of baby cry datasets"""
print("🔍 Exploring Baby Cry Dataset...")
print("=" * 50)
if not self.data_path.exists():
print(f"❌ Dataset path not found: {self.data_path}")
print("📥 Please download datasets first. See README.md for instructions.")
return False
# Get all categories
self.categories = [d for d in os.listdir(self.data_path)
if os.path.isdir(os.path.join(self.data_path, d))]
if not self.categories:
print("❌ No categories found in dataset directory")
return False
print(f"📁 Found {len(self.categories)} categories:")
for category in self.categories:
category_path = self.data_path / category
files = [f for f in os.listdir(category_path) if f.endswith(('.wav', '.mp3', '.m4a', '.flac'))]
self.file_info[category] = {
'count': len(files),
'files': files[:5] # Store first 5 files for analysis
}
print(f" • {category}: {len(files)} audio files")
return True
def analyze_audio_features(self):
"""Analyze audio features of sample files"""
print("\n🎵 Analyzing Audio Features...")
print("=" * 50)
features_summary = {}
for category in self.categories[:3]: # Analyze first 3 categories
if self.file_info[category]['files']:
sample_file = self.file_info[category]['files'][0]
file_path = self.data_path / category / sample_file
try:
# Load audio
y, sr = librosa.load(file_path)
# Extract features
duration = len(y) / sr
mfccs = librosa.feature.mfcc(y=y, sr=sr, n_mfcc=13)
spectral_centroids = librosa.feature.spectral_centroid(y=y, sr=sr)
zero_crossing_rate = librosa.feature.zero_crossing_rate(y)
rms = librosa.feature.rms(y=y)[0]
features_summary[category] = {
'duration': duration,
'sample_rate': sr,
'mfcc_mean': np.mean(mfccs, axis=1).tolist(),
'spectral_centroid_mean': float(np.mean(spectral_centroids)),
'zcr_mean': float(np.mean(zero_crossing_rate)),
'rms_mean': float(np.mean(rms))
}
print(f"\n📊 {category.upper()} Sample Analysis:")
print(f" • Duration: {duration:.2f} seconds")
print(f" • Sample Rate: {sr} Hz")
print(f" • MFCC Shape: {mfccs.shape}")
print(f" • Spectral Centroid Mean: {np.mean(spectral_centroids):.2f}")
print(f" • Zero Crossing Rate: {np.mean(zero_crossing_rate):.4f}")
print(f" • RMS Energy: {np.mean(rms):.4f}")
except Exception as e:
print(f"❌ Error analyzing {category}: {e}")
return features_summary
def create_data_summary(self):
"""Create a summary of the dataset"""
total_files = sum(info['count'] for info in self.file_info.values())
summary = {
'total_categories': len(self.categories),
'total_files': total_files,
'categories': self.categories,
'files_per_category': {cat: info['count'] for cat, info in self.file_info.items()},
'average_files_per_category': total_files / len(self.categories) if self.categories else 0
}
print(f"\n📈 Dataset Summary:")
print(f" • Total Categories: {summary['total_categories']}")
print(f" • Total Files: {summary['total_files']}")
print(f" • Average Files per Category: {summary['average_files_per_category']:.1f}")
return summary
def save_analysis_report(self, filename="data_analysis_report.json"):
"""Save analysis results to a JSON file"""
summary = self.create_data_summary()
features = self.analyze_audio_features()
report = {
'summary': summary,
'audio_features': features,
'timestamp': pd.Timestamp.now().isoformat()
}
with open(filename, 'w') as f:
json.dump(report, f, indent=2)
print(f"\n💾 Analysis report saved to: {filename}")
return report
if __name__ == "__main__":
# Initialize explorer
explorer = BabyCryDataExplorer("data")
# Explore dataset
if explorer.explore_dataset():
features = explorer.analyze_audio_features()
summary = explorer.create_data_summary()
report = explorer.save_analysis_report()
print("\n✅ Dataset exploration complete!")
print("Ready to proceed with model development.")
else:
print("\n❌ Dataset exploration failed. Please check your data directory.")
|