[attachment=13143]
Adaptive And Energy Efficient Wavelet Image Compression For Mobile Multimedia Data Services
ABSTRACT
To enable wireless Internet and other data services using mobile appliances, there is a critical need to support contentrich cellular data communication, including voice, text, image and video. However, mobile communication of multimedia content has several bottlenecks, including limited bandwidth of cellular networks, channel noise, and battery constraints of the appliances. In this paper, we address the energy and bandwidth bottlenecks of image data communication. We present an energy efficient, adaptive data codec for still images that can significantly minimize the energy required for wireless image communication, while meeting bandwidth constraints of the wireless network, the image quality, and latency constraints of the wireless service.
Based on wavelet image compression, we propose an energy efficient wavelet image transform algorithm (EEWITA) for lossy compression of still images, enabling significant reductions in computation as well as communication energy needed, with minimal degradation in image quality. We also present a dynamic configuration methodology that selects the optimal set of parameters to minimize energy under network, service, and appliance constraints
1. INTRODUCTION
To enable new wireless data services such as mobile multimedia email, mobile Internet access, mobile commerce, mobile data sensing in sensor networks, home and medical monitoring services, and mobile conferencing, there will be a growing demand for content-rich cellular data communication, including voice, text, image and video. One of the major challenges in enabling mobile multimedia data services will be the need to process and wirelessly transmit very large volumes of data. This will impose severe demands on the battery resources of multimedia mobile appliances as well as the bandwidth of the wireless network. While significant improvements in achievable bandwidth are expected with future wireless access technologies, improvements in battery technology will lag the rapidly growing energy requirements of future wireless data services. One approach to mitigate this problem is to reduce the volume of multimedia data transmitted over the wireless channel via data compression techniques.
Since images will constitute a large part of future wireless data, we focus in this paper on developing energy efficient and adaptive image compression and communication techniques. Based on a popular image compression algorithm, namely, wavelet image compression, we present an energy efficient wavelet image transform algorithm (EEWITA), consisting of techniques to eliminate computation of certain high-pass coefficients of an image. As shown by our experiments, the use of EEWITA can significantly reduce both (i) computation energy, by minimizing the computation needed to compress an image, and (ii) communication energy, consumed by the RF component of the mobile appliance, which is proportional to the number of bits transmitted. The reduction in energy is obtained with minimally perceptible loss in image quality.
We identify several parameters of EEWITA that can be varied, and analyze their effects on computation and communication energy, and image quality during wireless image communication. Based on EEWITA and its parameters, we have developed an adaptive image codec, which minimizes energy consumption and air time (service cost) needed for an image-based data service, while meeting bandwidth constraints of the wireless network, and the image quality and latency constraints of the wireless service. We demonstrate the effectiveness of the energy efficient, adaptive codec by applying it to image communication over multiple wireless access technologies, with significant energy and air time (service cost) savings compared to the use of a statically configured wavelet transform based codec.
2. WAVELET IMAGE COMPRESSION
In this section, we first present an overview of image compression. We then describe a typical wavelet transform algorithm, and analyze its energy consumption.
A) Background
Fig. 2.1 illustrates the main block diagram of the image compression (source coding) process. The image sample goes first through a transform, which generates a set of frequency coefficients. The transformed coefficients are then quantized (or divided by a certain fixed value) to reduce the volume of encoded data. The output of this step is a stream of integers, each of which corresponds to an index of a particular quantized binary. Encoding is the final step, where the stream of quantized data is converted to a sequence of binary symbols in which shorter binary symbols are used to encode integers that occur with relatively high probability. This helps reduce the number of bits transmitted. A number of different encoding schemes are available, such as Huffman coding and run length coding (RLC) .
Image compression can be implemented using a variety of algorithms, such as transform-based schemes, vector quantization and subband coding. The selection of an image compression algorithm for multimedia mobile communication depends not only on the traditional criteria of achievable compression ratio and the quality of reconstructed images, but also on associated energy consumption and robustness to higher bit error rates. By optimizing algorithmic features of the transform step, performance and energy requirements of the entire image compression process can be significantly improved. For this reason, we target the wavelet transform step to minimize the energy consumption.
We next describe a typical wavelet transform algorithm and then go on to analyze its energy consumption.
B) Wavelet Transform Overview
The forward wavelet-based transform uses a 1-D subband decomposition process where a 1-D set of samples is converted into the low-pass subband (Li) and high-pass subband (Hi). The low-pass subband represents a downsampled low-resolution version of the original image. The high-pass subband represents residual information of the original image, needed for the perfect reconstruction of the original image from the low-pass subband. The 2-D subband decomposition is just an extension of 1-subband decomposition. The entire process is carried out executing a 1-D subband decomposition twice, first in one direction (horizontal), then in the orthogonal (vertical) direction. For example, the low-pass subband (Li) resulting from the horizontal direction is further decomposed in the vertical direction, leading to LLi and LHi subbands. Similarly, the highpass subband (Hi) is further decomposed into HLi and HHi. After one level of transform, the image can be further decomposed applying the 2-D subband decomposition to the existing Lli subband. This iterative process results in multiple “transform levels”. For example, in Fig. 2.2(a), the first level of transform results in LH1, HL1, and HH1, in addition to LL1, which is further decomposed into LH2, HL2, HH2, LL2 at the second level, and the information of LL2 is used for the third level transform. We refer to the subband LLi as a low-resolution subband and high-pass subbands LHi, HLi, HHi as horizontal, vertical, and diagonal subband respectively since they represent the horizontal, vertical, and diagonal residual information of the original image. An example of three-level decomposition into subbands of the image CASTLE is illustrated in Fig. 2.2(b).
(b) Demonstration using image CASTLE
Having described the operation of the wavelet transform algorithm, we now address its efficiency from an energy standpoint.
C) Analysis of Energy Consumption
We choose the Daubechies 5-tap/3-tap filter for embedding in the forward wavelet transform. The main property of the wavelet filter is that it includes neighborhood information in the final result, thus avoiding the block effect of DCT transform . It also has good localization and symmetric properties, which allow for simple edge treatment, high-speed computation, and high quality compressed image. In addition, this filter is amenable to energy efficient hardware implementation because it consists of binary shifter and integer adder units rather than multiplier/divider units. The following equation represents the Daubechies 5-tap/3-tap filter.
To determine the energy efficiency of each algorithm, we use a metric that is independent of the detailed implementation of the algorithm. We analyze energy efficiency by determining the number of times certain basic operations are performed for a given input, which in turn determines the amount of switching activity, and hence the energy consumption. We model the energy consumption of the low/high-pass decomposition by counting the number of operations and denote this as the computational load. Thus 8S + 8A units of computational load are required in a unit pixel of the low-pass decomposition and 2S + 4A units for the high-passes.
For a given input image size of M N and wavelet decomposition applied through L transform levels, we can estimate the total computational load as follows. Suppose we first apply the decomposition in the horizontal direction. Since all even-positioned image pixels are decomposed into the low-pass coefficients and odd-positioned image pixels are decomposed into the high-pass coefficients, the total computational load involved in horizontal decomposition is 1/2MN(10S+12A). The amount of computational load in the vertical decomposition is identical. Using the fact that the image size decreases by a factor of 4 in each transform level, the total computational load can be represented as follows:
Computational load :
We estimate the data-access load by counting the total number of memory accesses during the wavelet transform. At a transform level, each pixel is read twice and written twice. Hence, with the same condition as the above estimation method, the total data-access load is given by the number of read and write operations:
Data-access load :
The overall computation energy is computed as a weighted sum of the computational load and data-access load. The add operation requires two times more energy consumption than the shift operation, and the energy cost of the data-access load is 2.7 times more than the computational load. We also estimate the
The overall computation energy is computed as a weighted sum of the computational load and data-access load. From our implementation experiments, we found that the add operation requires two times more energy consumption than the shift operation, and the energy cost of the data-access load is 2.7 times more than the computational load. We also estimate the communication energy by C*R, where C is the size of the compressed image (in bits) and R is the per bit transmission energy consumed by the RF transmitter.
Having analyzed the sources and magnitude of energy consumption in the wavelet transform, we next present techniques to minimize the computation energy as well as communication energy needed in wavelet-based image compression and wireless transmission