Comprehensive Analysis of Fused Descriptors for Image Retrieval

Content Based Image Retrieval (CBIR) systems are used to retrieve similar images to the query image from a large database. This paper represents a CBIR model which has been tested with multiple feature descriptors such as Scale-Invariant Feature Transform (SIFT), Speeded Up Robust Features (SURF), Binary Robust Independent Elementary Features (BRIEF), Oriented FAST and rotated BRIEF (ORB) and combinations of them. Multiple linguistic processing techniques such as Bag of Words and Topic modelling have been used for optimizing the image retrieval and making them meaningful based on human semantics. Using a combination of descriptors and Latent Dirichlet Allocation, our model has proven to yield high precision when tested against standard image retrieval data set.