Traditional Indian Painting Retrieval System Based On Curvelet Transform And Comparison The Result Using Gabor Filter
Priya Saxena Pooja Manjarekar · Zenodo (CERN European Organization for Nuclear Research) · 2016
Today the technology in which relevant images from a large databases are searched according to the user’s interest is famous by the name of Content based Image Retrieval or CBIR. Since last two decades it has become an active and fast advancing field amongst the researchers. Last decade is witness of the progress achieved in both theoretical as well as in system development. However this area of technology is still full of challenges that researchers from multiple disciplines are being continuously attracted to work with. As we are well known about the hierarchy of spectral methods of texture feature extraction starting from Fourier Transform (FT) to the latest Gabor filter transform, which became very popular for many useful applications still was found to lag the curved point singularities or can say curve lines along the edges of images. In this paper to overcome this problem we have chosen to work with Curvelet transform. The dataset of retrieval system we are presenting here is made up of three Indian traditional paintings named as Warli, Madhubani and Fadd. As mentioned earlier Curvelet transform will be applied to get the result. In the second part of this paper we shall also compare the result with Gabor transform.