TY - GEN
T1 - Wavelet domain statistical order filter using the tri-state median filter algorithm
AU - Jesús, Martínez Valdes
AU - Javier, Gallegos Funes Francisco
AU - Antonio, Acevedo Mosqueda Marco
PY - 2005
Y1 - 2005
N2 - This paper proposes a new statistical order filter in the wavelet domain that suppress the impulsive noise obtaining better results than the methods developed by other researchers [1,2]. This filter takes the basis of three algorithms [3-5] and joins them to get a powerful method for reducing the impulsive noise present in color images while preserving edge and important details information. At first, this filter employs a method known as redundancy of approaches [3] to reduce the noise present in the image low frequencies, after the previous task, it applies a second method [4] as much in low as in high frequencies to continue suppressing impulsive noise contained in the image, this method [4] is able to detect a corrupted sample (first stage) and to differentiate if the sample corresponds to noise or edge information (second stage), in this case and for getting better results, the last part of this method is replaced by a Tri-State Median Filter Algorithm [5] that improves the performance of the original algorithm [4]. All mentioned above and the results obtained prove that this wavelet domain statistical order filter is an important tool that must be considered by anyone who wants to solve the impulsive noise suppression task.
AB - This paper proposes a new statistical order filter in the wavelet domain that suppress the impulsive noise obtaining better results than the methods developed by other researchers [1,2]. This filter takes the basis of three algorithms [3-5] and joins them to get a powerful method for reducing the impulsive noise present in color images while preserving edge and important details information. At first, this filter employs a method known as redundancy of approaches [3] to reduce the noise present in the image low frequencies, after the previous task, it applies a second method [4] as much in low as in high frequencies to continue suppressing impulsive noise contained in the image, this method [4] is able to detect a corrupted sample (first stage) and to differentiate if the sample corresponds to noise or edge information (second stage), in this case and for getting better results, the last part of this method is replaced by a Tri-State Median Filter Algorithm [5] that improves the performance of the original algorithm [4]. All mentioned above and the results obtained prove that this wavelet domain statistical order filter is an important tool that must be considered by anyone who wants to solve the impulsive noise suppression task.
KW - Discrete Wavelet Transform
KW - Wavelet Coefficients
UR - http://www.scopus.com/inward/record.url?scp=33748892159&partnerID=8YFLogxK
U2 - 10.1109/ICEEE.2005.1529566
DO - 10.1109/ICEEE.2005.1529566
M3 - Contribución a la conferencia
AN - SCOPUS:33748892159
SN - 0780392302
SN - 9780780392304
T3 - 2nd International Conference on Electrical and Electronics Engineering, ICEEE and XI Conference on Electrical Engineering, CIE 2005
SP - 32
EP - 35
BT - 2nd International Conference on Electrical and Electronics Engineering, ICEEE and XI Conference on Electrical Engineering, CIE 2005
T2 - 2nd International Conference on Electrical and Electronics Engineering, ICEEE and XI Conference on Electrical Engineering, CIE 2005
Y2 - 7 September 2005 through 9 September 2005
ER -