A Two-Stage Algorithm for the Detection and Removal of Random-Valued Impulse Noise Based on Local Similarity

Abstract

A two-stage denoising algorithm based on local similarity is proposed to process lowly and moderate corrupted images with random-valued impulse noise in this paper. In the noise detection stage, the pixel to be detected is centered and the local similarity between the pixel and each pixel in its neighborhood is calculated, which can be used as the probability that the pixel is noise. By obtaining the local similarity of each pixel in the image and setting an appropriate threshold, the noise pixels and clean pixels in the damaged image can be detected. In the image restoration stage, an improved bilateral filter based on local similarity and geometric distance is designed. The pixel detected as noise in the first stage is filtered and the new intensity value is the weighted average of all pixel intensities in its neighborhood. A large number of experiments have been conducted on different test images and the results show that compared with the mainstream denoising algorithms, the proposed method can detect and filter out the random-value impulse noise in the image more effectively and faster, while better retaining the edges and other details of the image.

Learning-to-augment incorporated noise-robust deep CNN for detection of COVID-19 in noisy X-ray images

“…The run-time of the methods that improve the quality of noisy X-ray images is of great important [40] , especially for the point-of-care machines in the clinical environment. Typically, X-ray images corrupted by the impulse noise are enhanced in two phases [24] , [41] , [42] . In the first phase, noise-free or noisy pixels are identified.…”

Defect Recognition of Roll-to-Roll Printed Conductors Using Dark Lock-in Thermography and Localized Segmentation

“…However, the process of finding the optimal parameters for the Canny edge detector are challenging, because the efficacy of the Canny edge detector is highly dependent on the low and high thresholding as well as the Gaussian filter used. In addition to the thresholding and edge-based detection methods, there are other methods proposed based on various algorithms [23][24][25][26], such as region-based methods [27], clustering-based methods [28] and deep-learning methods [29][30][31]. However, they all have their own drawbacks.…”

“…The schematic diagram for this process is depicted in Figure 14. In addition to the thresholding and edge-based detection methods, there are other methods proposed based on various algorithms [23][24][25][26], such as region-based methods [27], clustering-based methods [28] and deep-learning methods [29][30][31]. However, they all have their own drawbacks.…”

The alternative irregularity reduction algorithm built on 2-stage identification with AMF on FMIO

“…By ordinary, the impulse irregularity is destroyed in digital depictions by deterioration machinery as an outcome of depiction keeping proceeding, distribution proceeding, computerized flare or etc. From logical opinion, the impulse irregularity component [22]- [26] may be identified into 2 classes: random magnitude impulse outlier (RMIO) and fix magnitude impulse outlier (FMIO). The RMIO may be any magnitude over the dynamic magnitude, which is varied from "0" to "255" thereupon the RMIO may be more difficultly identified than FMIO where the FMIO may completely be "0" or "255" while the FMIO may be promptly identified than RMIO.…”

“…For natural depictions, the generous allotment of depiction pixels is steady territory nonetheless the diminutive allotment is discontinuous territory. Accordingly, in the early years, most irregularity reduction algorithms [22]- [26] were offered for reducing the outlier depiction from impulsive irregularity.…”