Motion vector de-interleaving method

Abstract

The method includes following steps: decoding a image content into multiple macro data pieces; from the multiple macro data pieces, calculating a motion vector, which is as a judgment value; from the said judgment value, generating multiple deinterleaved labels; implementing time stretching operation for generated deinterleaved labels. The time stretching operation includes following actions further: setting up multiple buffers of deinterleaved labels; determining whether current deinterleaved label is a field combination format (Weave); determining whether other buffers of deinterleaved labels are single field interpolation format (BOB); and setting up the deinterleaved labels as derived single field interpolation format (BOB 2); carrying out destructive operation to destroy multiple isolated deinterleaved labels, and making them possess correlativity in spatial motion. The invention prevents phenomena of mouse teeth in more degree.

Description

technical field

[0001] The invention relates to a method for de-interlacing motion vectors, in particular to a method for decoding MPEG-2 images. Background technique

[0002] At present, the source of multimedia is quite extensive for us, such as TV programs, movies, video discs (VCD) and digital versatile discs (DVD) and so on. Among so many sources, we have found that these sources can be roughly divided into two categories: progressive (non-interlaced) images and interlaced (interlaced) images. Most of these images follow the coding principles of MPEG-2 to encode these images.

[0003] In the process of MPEG-2 video encoding / decoding, it is currently the general processing method to convert the interlaced video content into a de-interlacing format (De-interlacing), and the more common de-interlacing formats are the scene combination format (Weave) and the single format. The field interpolation format (BOB), the field combination format (Weave) simply combines the data ...

Images