An Analysis of Codebook Optimization for Image Compression: Modified Genetic Algorithm and Particle Swarm Optimization Algorithm

Billions of images are uploaded daily, and it requires a large storage space. Utilization of better storage capacity and to improve uploading/downloading time, researchers have designed an image compression model. Many researchers have implemented various approaches to improve the image compression ratio of an image. This paper presents an analysis of various optimization algorithms based on vector quantization (VQ). The first algorithm is a modified genetic algorithm. It is based on Darwin’s principle which is natural characteristics. Those who are fit can survive and use it to optimize the codebook. A second algorithm for optimization of the codebook is particle swarm optimization (PSO). PSO algorithm is superior to finding the codeword vectors of codebook from the training image samples for image compression. In the PSO algorithm, the selection approach plays an important role to select the particle based on the fitness of the population. Training images from the standard image database are used for the design of the codebook. The input image set is 4 × 4 or 8 × 8 blocks and is represented as vectors. They are referred to as codewords in the codebook, and it is a component of a code. The codebook size is measured by codewords. The block size is decided by the length of the codeword. These codewords generate the codebook by entering the vector value. Compression is done with the help of sending indices to the decoder. Likewise, analysis of quality measures is presented to the modified GA and PSO algorithms based on mean square error, peak signal-to-noise ratio, structural similarity index, and average difference. In this work, we have calculated bits per pixel (BPP), the compression ratio (CR), and the % compression ratio. The experimental results are validated.