Adaptive spatio-temporal video noise filtering for high quality applications

High-quality digital video filtering applications pose interesting challenges. Noise reduction cannot come at the expense of losing picture detail or introducing post-filtering artifacts. We address these issues by proposing an adaptive spatio-temporal method for reducing noise in digital video. The novelty of this method is that it aims to achieve an "even" amount of noise reduction at each pixel while preserving picture detail. The emphasis on even noise reduction helps reduce post-filtering artifacts arising from uneven filtering. We further propose an algorithm for automatically estimating the key filtering parameters, given two intuitive high-level inputs, filtering strength and a tolerance value. The high-level inputs provide user-friendly "knobs" that could be adjusted to obtain the desired filtering performance.