Data-Dependent Subband Coder for Image Compression
Subband coding is a popular technique extensively used in the areas of image and video communications. This technique is widely employed to achieve data compression and efficient data transmission. The general principle of subband coding involves dividing
- PDF / 184,110 Bytes
- 8 Pages / 439.37 x 666.142 pts Page_size
- 86 Downloads / 222 Views
Abstract Subband coding is a popular technique extensively used in the areas of image and video communications. This technique is widely employed to achieve data compression and efficient data transmission. The general principle of subband coding involves dividing the input image based on its frequency, where the correlation between the subbands still exists. In this paper, we propose a new technique where the subband coder divides the image such that the subbands will have only decorrelated data, thus reducing the redundancy and achieving image compression. To exploit for decorrelation between the subbands, we adopt a polynomial eigenvalue decomposition technique which is popularly known as Sequential Best Rotation (SBR2C) algorithm. It is an iterative technique with a set of delay and rotation operations. Using this approach, we design the analysis filter bank whose operation is data dependent and thus splits the input image into subbands that are strongly decorrelated. At the receiver end, we design the synthesis bank by imposing perfect reconstruction condition in the absence of quantization errors. We compare our proposed algorithm with the popular subband coding technique, discrete wavelet transform (DWT). Simulation results show that proposed method can perform better compression because of data-dependent operation of filter banks.
Keywords Subband coding Polynomial eigen value decomposition Strong decorrelation Perfect reconstruction
1 Introduction Image compression is an important task in the field of image and video processing which allows efficient storage and transmission of signals in multimedia applications [1, 2]. The goal of image compression is to represent the image with less number of bits while the reconstructed image is still similar to original image. J.D.K. Abel D. Samiappan (&) N. Ponnusamy Department of ECE, SRM University, Kattankulathur 603203, India e-mail: [email protected] © Springer Nature Singapore Pte Ltd. 2018 H.S. Saini et al. (eds.), Innovations in Electronics and Communication Engineering, Lecture Notes in Networks and Systems 7, https://doi.org/10.1007/978-981-10-3812-9_19
185
186
J.D.K. Abel et al.
Thus, the digital images can be stored using less memory and transmitted using smaller bandwidth. Image compression can be achieved in several ways. One of the techniques is transform coding where image prior to transmission is subjected to invertible transforms with an aim to remove redundant data. Discrete Cosine Transform (DCT) [3–5] and Vector Quantization (VQ) [6–8] coding are some of popular transform coding approaches. The other popular technique employed to achieve image compression is subband coding. Subband coding is proven to give pleasing reconstructions compared to transform coding. The principle involved in subband coding is to decompose the input image into subbands based on their frequency. The human eyes are less sensitive to high-frequency components and thus compression can be achieved by reducing such components. Several subband coding techn
Data Loading...
