PROJECT TITLE :

Efficient Scalable Median Filtering Using Histogram-Based Operations - 2018

ABSTRACT:

Median filtering could be a smoothing technique for noise removal in pictures. Whereas there are various implementations of median filtering for one-core CPU, there are few implementations for accelerators and multi-core systems. Several parallel implementations of median filtering use a sorting algorithm for rearranging the values within a filtering window and taking the median of the sorted worth. While using sorting algorithms permits for easy parallel implementations, the price of the sorting becomes prohibitive as the filtering windows grow. This makes such algorithms, sequential and parallel alike, inefficient. In this work, we introduce the first software parallel median filtering that's non-sorting-based mostly. The new algorithm uses efficient histogram-based mostly operations. These cut back the computational needs of the new algorithm whereas also accessing the image fewer times. We tend to show an implementation of our algorithm for both the CPU and NVIDIA's CUDA supported graphics processing unit (GPU). The new algorithm is compared with many alternative leading CPU and GPU implementations. The CPU implementation has near good linear scaling with a three.seven× speedup on a quad-core system. The GPU implementation is several orders of magnitude faster than the opposite GPU implementations for mid-size median filters. For tiny kernels, 3 × three and 5 × 5, comparison-based mostly approaches are preferable as fewer operations are needed. Lastly, the new algorithm is open-supply and will be found within the OpenCV library.