A Comparative Analysis of Image Compression Techniques: K Means Clustering and Singular Value Decomposition

The global drive to digitize almost all the existing processes has mandated the conversion of all concerned analog data into their respective digital formats. One such crucial data that is being digitized on a priority in today’s world is image. An image is a type of data which is composed of picture elements called pixels. It can be represented as a matrix for the manipulating process. The storage of a vast database of image files occupies a huge memory space in the disk. To overcome this hassle, image files can be compressed and saved. This image compression process is aimed at reducing the data size in terms of bytes and enable the efficient storage and transmission of image files. Image compression can be achieved through several algorithms. In this paper, we discuss two such algorithms, namely k means clustering and singular value decomposition. K Means Clustering technique helps in minimizing the colour components of the image. Singular Value Decomposition technique can be carried out by low rank approximation of the image matrix. This research work is performed using the Python platform and subsequently the efficiency of both the methods is compared. The comparative analysis of the simulation results are further compared with the existing methods to show the competence of different methodologies. Thus, this work strives to be of learned assistance to the concerned aspirants in choosing the best algorithm for their applications.