Noise reduction pre-processor for digital video using previously generated motion vectors and adaptive spatial filtering

作者: Siu-Wai Wu , Jing Yang Chen

DOI:

关键词:

摘要: Noise is reduced in a video system by applying motion compensated temporal filtering using previously generated vectors and adaptive spatial at scene change frames. Various types of noise can be introduced into prior to compression transmission. Artifacts arise from recording signal manipulation, terrestrial or orbital communications, during decoding. image interferes with performance subsequently impairs performance. While generally reduces image, it also reduce edge definition leading loss focus. Filtering tax throughput, since increased computational complexity often results schemes. Furthermore, the movement objects within frames, as defined groups pixels, complicates reduction process adding additional complexity. In addition improvements made FIR filtering, present invention improves on previous techniques Infinite Impulse Response (IIR) while maintaining definition. It uses calculated part first-pass encoding alternatively transcoding for P-frame B-frame preprocessing. Single stage double are presented.