Hybrid of DCT And PCA For Image Size Compression

One of a person’s most crucial biometric traits is their face. Face recognition is a crucial first step for applications in many fields, such as biometrics, user interfaces for computers, and surveillance. Most of these techniques fail due to high memory usage. Therefore, image compression is very important today as it is used in most applications that are concerned with the amount of memory usage and speed of transferring information. In this research, the speed of data transmission was increased and the size of the images was reduced while maintaining accuracy by using discrete cosine transform (DCT) hybrid with principal component analysis (PCA). Face recognition techniques compare the desired face image with a set of face images stored in a database. An equalizer followed by the median filter has been used for suitable intensity and removing the noise from the faces images as a first step before the face recognition process. Different databases are implemented to evaluate the performance of the system’s algorithms such as ‘face94’. Concerned with memory amount, experimental results showed that the use of compression and dimensionality reduction with face recognition process produced better results compared with the applying the compression alone. DCT compressed the image memory size by 60% (from 6Kbyte to 2.4Kbyte), while based on Eigen vectors of 90, DCT+PCA compressed the image size by 83.5%. This last statement is the contribution of this research.