Color image compression using partial pixels elimination and hybrid pattern recognition

Abstract EN

The increased usage of media in our daily life has forced the need for efficient image storage and transmission.

It has been shown previously that eliminating some image columns and rows of pixels can be recovered by the use of neural network.

In this article, the method is further investigated for color images.

Many of existing methods consume relatively large execution time for image compression.

The computation time using this method is negligible compared to other method.

Additionally, it takes less than half of the execution time than SPIHT to perform image compression at a compression quality greater than 0.75.