摘要 |
The method involves spatially sub-sampling a video sequence (101), and compressing the sub-sampled video sequence to form a compressed sequence (109), which is decompressed and oversampled. A difference between the original video sequence and the decompressed video sequence is determined to produce an intermediate result. A mask is applied to an area of interest (102) of the result to produce a fused video sequence (103) that is compressed to form another compressed sequence (104). The compressed sequences are multiplexed to form a single compressed sequence (105) of low bandwidth. |