A near-optimal wavelet-based estimation technique for video sequences

wavelet;video;multichannel;denoise

Bonham, Melody I. ; Kamalabadi ; Farzad

关键词: wavelet;    video;    multichannel;    denoise;   

Others  :  https://www.ideals.illinois.edu/bitstream/handle/2142/18501/Bonham_Melody.pdf?sequence=1&isAllowed=y

美国|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

This thesis presents a method for estimation of a video signal given a data set with Poissonnoise. The cameras used in creating video sequences are often charge-coupled devices,which produce data by way of a counting process, leading to noise with a Poissondistribution. Because many applications using video require data with less noise, a methodof reducing the noise and estimating the original signal is desired. The method presentedin this thesis attempts to accomplish this goal without using a Wienerlter, which cande-noise signals and is optimal in the mean-square error sense, but is hard to implementbecause second-order statistics may be unknown and because of the inversion of a possiblylarge matrix. Instead, an approximation of the Wienerlter is accomplished byrstperforming a one-dimensional discrete Fourier transform in order to decorrelate the videosequence between each two-dimensional frame or across each channel, and then performinga two-dimensional discrete wavelet transform on each of the resulting frames. Thresholdingis then implemented, and the inverse transform is applied in order to recover an estimate ofthe original signal. It is shown that this scheme is e ective in improving signal-to-noiseratio in synthetic video sequences and video captured by a camera.