Skip to content

Latest commit

 

History

History
28 lines (16 loc) · 928 Bytes

File metadata and controls

28 lines (16 loc) · 928 Bytes

Feature Compression using PCA and SVD in Images from Scratch

  • Performed lossy compression using PCA and using SVD from scratch on images with varying number of principal components.
  • Generated Eigen Faces from Olivetti dataset using PCA with different number of eigen values.

Dataset

Output with Top-123 Eigen Values: 5% Error

Output with Top-5 Eigen Values

Error vs Number of PCA Components Considered

Contributions

Saurabh Chatterjee
MTech, Signal Processing and Machine Learning
Indian Institute of Technology (IIT) Kharagpur