GPU Accelerated Image and Video Processing

Abstract

Central Processing Unit (CPU) is called the brain of the computers and the mobile devices and designed in such a way that it performs best in executing tasks sequentially. So, is the CPU alone fast enough when it comes to perform an operation on high quality images and videos to provide the best user experience? Graphics Processing Unit (GPU) is a dedicated hardware which is mainly used for rendering graphics; however, its general-purpose nature allows it to be used in other generic tasks where high level of parallelism is required.

The objective of this thesis is to find out how the GPU plays essential role when it comes to process the images and the videos and what makes the GPU so efficient to process them compared to the CPU. A comparative study is made in this thesis to analyse why the GPU has an edge over the CPU when images and videos need to be processed fasters. This is verified with the help of some tests which were performed on a laptop computer equipped with Intel CPU and Nvidia GPU. Tests results clearly show that the GPU outperforms the CPU with 4 to 6 times faster processing rate. But the same time processing power of the GPU can not be accessed directly, the CPU’s involvement is something which is a must in order to access GPU’s functionality.