Design and development of de-blur/de-noising filter using multiple image captures on a massively parallel architecture

Now a days mobile devices are capable of performing more operations in less amount of time. More number of GPU and CPU cores are added to the mobile devices. Due to increase in processing power various image processing algorithms are targeted at mobile devices. The users of these devices demand quality that is similar to that of a professional camera. There are definite limitations on the sensors that can be used in a smart phone. However, with the increase in computing power of smart phones/tablets, better software can be developed to compensate for these shortcomings. The image capture in low light condition can be enhanced by using guided filter. In this paper, we present the implementation of the Guided filter on Nexus 4. It has been observed that the speedup increases with respect to increase in size of image and size of window.