Image retrieval using locality preserving projections

Abstract

In web surfing, image retrieval algorithms play a vital role. Many algorithms are existing but they have their drawbacks and also low retrieval rates. Locality preserving projections is a novel method for image retrieval proposed in this study. As the method includes principal component analysis, nearest neighbour and eigen features in extracting the hyper parameters for classification, it has better retrieval rate. This novel approach is tested using the Yale facial image dataset and the performance of the algorithm is tested by using different classifiers like Decision Tree etc.

Keywords

• decision trees
• image classification
• feature extraction
• image retrieval
• face recognition
• principal component analysis
• locality preserving projections
• web surfing
• image retrieval algorithms
• low retrieval rates
• principal component analysis
• retrieval rate
• yale facial image dataset