Improvement of quantized adaptive switching median filter for impulse noise reduction in gray-scale digital images

Abstract

Digital images may suffer from fixed value impulse noise due to several causes. The noise significantly degrades the quality of the image, which may affect the subsequence image processing. Therefore, a noise reduction technique is required to restore the image. In this paper, a new method, which is called improvement of quantized adaptive switching median filter (IQASMF), has been proposed to reduce the fixed value impulse noise from gray-scale digital images. The implementation of IQASMF has five processing blocks. The first processing block is the noise detection block, where the noise pixel candidates are detected based on the intensity value. Then estimation of the local noise density is done by the second processing block. Next, the third processing block filters the corrupted pixel candidates with filters of predefined size, depending on the local noise density. After that, the noise mask is updated in the fourth processing block. Finally, the fifth processing block processes the noise residuals from the third processing block by using a size adaptive filter.Experimental results from twenty standard gray-scale images of various sizes have shown that IQASMF has the ability to restore images for up to 99% of the impulse noise corruption. As compared with the other five median filter-based methods, from the measures of mean squared error (MSE) and structural similarity index (SSIM), it is shown that the performance of IQASMF is equivalent to the performance of other methods at low and medium levels of corruption.However, at high corruption levels, IQASMF has demonstrated the best performance in terms of MSE and SSIM. The outputs from IQASMF also have the best visual appearance.

Median Filtering Using First-Order and Second-Order Neighborhood Pixels to Reduce Fixed Value Impulse Noise from Grayscale Digital Images

“…Impulse noise can be divided into two categories, which are random value impulse noise and fixed value impulse noise [13][14][15]. Random value impulse noise is usually caused by errors in timing or synchronization of analog to digital converter [15,16], malfunctioning pixel elements in the camera sensors or faulty transmission [3,[15][16][17]. The intensity values of the damaged pixels in the image by random value impulse noise ranged from the minimum value to the maximum value in the dynamic range [13,15,18,19].…”