Generally the methods have been used in face recognition and processing a group of images which belongs to similar objects, such as face image databases. In this paper a new method is introduced to compress one coloured image using PCA technique. | ISSN:2249-5789 Mohammad Mofarreh Bonab et al, International Journal of Computer Science & Communication Networks,Vol 2(1), 111-116 A New Technique for Image Compression Using PCA Mohammad Mofarreh-Bonab1, Mostafa Mofarreh-Bonab2 1. University of Bonab, East Azerbaijan, Iran 2. Shahid Beheshti University, Tehran, Iran 1. 2. Abstract PCA is a statistic approach which widely used in many fields of study. Recently, this technique is used in image processing as a powerful tool especially in face recognition and face database compression. By working on this method in the recently researches, this method has been improved and kernel PCA, 2D-PCA and some other methods has been introduced. Generally the methods have been used in face recognition and processing a group of images which belongs to similar objects, such as face image databases. In this paper a new method is introduced to compress one coloured image using PCA technique. 1. Introduction PCA method for images is called HOTELLING or KL transform. Common KL transform focuses on similarities between different images in the database. In this paper, we will show that by applying some changes in KL transform, we can use this technique for compressing just one coloured image. Keywords: Image compression, HOTELLING, KL transform, database compression, PCA. matrix that its elements are the intensity values of images. F1 X= ⋮ , Fi = xi1 , xi2 , , xiQ 1×Q Eq. 1 FM M×Q The term Fi indicates .
