The Adaptive 3-D Noise Reduction System using spatial and temporal filter by motion and noise detection

Generally spatial noise reduction(NR), though if is simple to implement, has a image blurring artifact. Meanwhile temporal NR removes a noise effectively but that method is not effective for a impulse noise or a plenty of noise condition. To solve these problems, a new NR algorithm for removing noise of image is presented. In the proposed algorithm, the spatial NR and temporal NR are organically combined by detection of noise and motion in the temporal NR stage. The spatial NR filter decomposes the noisy signal to data and noise(edge) by detection of noise level in signal. This detection of noise is measured in the temporal NR. Thus, the soft threshold value of the spatial NR determines the amount of noise by detection of the noise in temporal NR. The proposed algorithm improves the performance of noise reduction without image blurring artifact. As a result, the proposed algorithm has achieved more increasing the 5dB gains than other algorithms for noisy image and can be performed robustly and adaptively in many cases of image.