Gunshot Detection

python 2.7

Install requisites

pip install -r requirements.txt

Initial setup

python setup.py

Features

• Chromagram {c_stft} (Default - 12 coeff) (chroma_coeff)
• Constant Q-Chromagram {c_cqt} (Default - 12 coeff) (chroma_coeff)
• Chroma Energy Normalized {c_cens} (Default - 12 coeff) (chroma_coeff)
• Melspectrogram {mel_spec} (128 coeff)
• Mel Frequency Cepstral Coefficients {mfcc} (Default - 20 coeff) (mfcc_coeff)
• Root Mean Square Energy {rmse} (1 coeff)
• Spectral Bandwidth {bandwidth} (1 coeff)
• Spectral Centroid {centroid} (1 coeff)
• Spectral Contrast {contrast} (Default - 6 bands (bands+1 coeff)) (contrast_bands)
• Spectral Rolloff {rolloff} (1 coeff)
• Poly features {poly} (Default - 1 order (order+1 coeff)) (poly_order)
• Tonnetz {tonnetz} (6 coeff)
• Zero Crossing Rate {zcr} (1 coeff)

Vector Reduction

• Mean {mean}
• IQR {iqr} (Default - 50 percentile)
• Variance {var}

Feature Reduction (Default - 10 Features)

• Unsupervised

  • Principal Component Analysis {pca}
  • Incremental PCA {ipca}
  • Kernel PCA {kpca}
  • Fast Independent Component Analysis {fica}
  • Truncated SVD {tsvd}
  • Non-negative Matrix Factorization {nmf}
  • Sparse PCA {spca}
  • Canonical Correlation Analysis {cca}
  • Partial Least Square SVD {plssvd}
  • Gaussian Random Projection {grp}
  • Sparse Random Projection {srp}

• Supervised

  • Linear Discriminant Analysis {lda} (classes-1 Features)

Neural Network

• Multilayer Perceptron

Support Vector Machines

• Types

  • C SVM
  • Nu SVM

• Kernels

  • Linear {linear}
  • Polynomial {poly} (Default 3 degree) (poly_degree)
  • Radial Basis Function {rbf}
  • Sigmoid {sigmoid}