Application of Bandelet Transform in Image and Video Compression

Abstract

The need for large-scale storage and transmission of data is growing exponentially With the widespread use of computers so that efficient ways of storing data have become important. With the advancement of technology, the world has found itself amid a vast amount of information. An efficient method has to be generated to deal with such amount of information. Data compression is a technique which minimizes the size of a file keeping the quality same as previous. So more amount of data can be stored in memory space with the help of data compression. There are various image compression standards such as JPEG, which uses discrete cosine transform technique and JPEG 2000 which uses discrete wavelet transform technique. The discrete cosine transform gives excellent compaction for highly correlated information. The computational complexity is very less as it has better information packing ability. However, it produces blocking artifacts, graininess, and blurring in the output which is overcome by the discrete wavelet transform. The image size is reduced by discarding values less than a prespecified quantity without losing much information. But it also has some limitations when the complexity of the image increases. Wavelets are optimal for point singularity however for line singularities and curve singularities these are not optimal. They do not consider the image geometry which is a vital source of redundancy. Here we analyze a new type of bases known as bandelets which can be constructed from the wavelet basis which takes an important source of regularity that is the geometrical redundancy.The image is decomposed along the direction of geometry. It is better as compared to other methods because the geometry is described by a flow vector rather than edges. it indicates the direction in which the intensity of image shows a smooth variation. It gives better compression measure compared to wavelet bases. A fast subband coding is used for the image decomposition in a bandelet basis. It has been extended for video compression. The bandelet transform based image and video compression method compared with the corresponding wavelet scheme. Different performance measure parameters such as peak signal to noise ratio, compression ratio (PSNR), bits per pixel (bpp) and entropy are evaluated for both Image and video compression.