Abstract:
Described herein is a method for block coding of a flow (I) of data, in particular video data, via a compression operation (100; 200; 300) that comprises: applying to input-data blocks (B, CB; MB) in said data flow (I) a discrete-cosine-transform (DCT) operation (120; 220) and a quantization operation (130; 230) to produce compressed-data blocks (DB); subjecting said compressed-data blocks (DB) to a coding operation (235, 240) to obtain compressed output flows (O); and moreover applying an inverse-quantization operation (190; 290) and an inverse-discrete-cosine-transform (IDCT) operation (195; 295) on said compressed-data blocks (DB) to obtain reconstructed blocks (RB). The method envisages controlling (1000; 1100) generation of mismatch errors from said input-data blocks (B, CB; MB) via the operations of: - detecting from said input-data blocks (B, CB; MB) and compressed-data blocks (DB) data blocks (DB; MB) that are liable to cause mismatch errors; and - modifying said block (DB; MB) that are liable to cause mismatch errors prior to said coding operation (235, 240).
Abstract:
A digital image processing system and method for removing motion effects from images of a video sequence, and generating corresponding motion compensated images.