Image Sequences Filtering using a New Nonlinear Method

M. Saeidi, S.A. Motamedi, B. Saeidi, A. Behrad, Roghayeh Saeidi, and Reza Saeidi (Iran, USA)


Image sequences filtering, non-linear functions, low density noise, Median Filter.


In this paper, we will propose a novel spatiotemporal non linear method in image sequences filtering. The Median filter, using consecutive frames is more efficient than using only one frame in removing medium and high density Salt and Pepper noise. However, simple median filter using only one frame removes low-density Salt and Pepper noise more effectively than using consecutive frames. This results from the fact that, the median filter isn’t able to efficiently utilize the correlation of pixels among the consecutive frames under low-density Salt and Pepper noise. We will propose a solution to improve the performance of median filters using image sequences under low-density Salt and Pepper noise. The proposed method utilizes non-linear functions to improve the efficiency of median filters using consecutive frames. Our proposed filter applies consecutive median to the selected pixels in each spatiotemporal window and then applies consecutive non-linear functions.

Important Links:

Go Back