Browse Prior Art Database

Method of Logo Insertion in MPEG Vido Transcoding

IP.com Disclosure Number: IPCOM000015612D
Original Publication Date: 2002-Jan-07
Included in the Prior Art Database: 2003-Jun-20
Document File: 5 page(s) / 89K

Publishing Venue

IBM

Abstract

Method of Logo Insertion In MPEG Video Transcoding What is disclosed is a low cost and efficient method for logo insertion in MPEG video transcoding. We consider both cases of inserting the transparent and non-transparent logos. The non-transparent logo totally replaces the part of the image data covered by the logo. The transparent logo only overlays on the original image data that still can be seen through the logo. In efficient and low cost video transcoding, the information obtained from the input bit stream is reused. This reduces the computational complexity tremendously. In order for our logo insertion scheme to be applicable to as many different transcoders as possible, our scheme also reserve and reuse the information decoded from the input stream as much as possible. Our logo insertion scheme does not require estimating motion vectors again for the logo affected areas. Instead, it only makes simple but effective modifications on motion vectors and macroblock modes for the affected macroblocks, and the quantization of these areas. 1. Modification of Motion Information We assume, without losing the generality, that the logo part occupies an integer number of macroblocks, especially for covered logo. There is very little difference in modifying the motion information for these two types of logos. We will describe them in one piece. As metioned, some of the information decoded from the compressed bit stream is preserved in most of the transcoders, such as the picture type and the motion information. Since the logo usually only affects a small portion of a limited number of pictures in the video sequence, we reuse the old motion vectors for the part unaffected by the logo. So we only need to modify the motion vectors of the part affected by the logo, which includes the part covered by the logo and the part outside of the logo but in motion compensation used the area inside the logo (only for P, B frames). We name them "logo part" and "logo-affected part" respectively in the following discussion. We address on how to deal with these two situations separately.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 25% of the total text.

Page 1 of 5

Method of Logo Insertion in MPEG Vido Transcoding

Method of Logo Insertion In MPEG Video Transcoding

What is disclosed is a low cost and efficient method for logo insertion in MPEG video transcoding. We consider both cases of inserting the transparent and non-transparent logos. The non-transparent logo totally replaces the part of the image data covered by the logo. The transparent logo only overlays on the original image data that still can be seen through the logo. In efficient and low cost video transcoding, the information obtained from the input bit stream is reused. This reduces the computational complexity tremendously. In order for our logo insertion scheme to be applicable to as many different transcoders as possible, our scheme also reserve and reuse the information decoded from the input stream as much as possible. Our logo insertion scheme does not require estimating motion vectors again for the logo affected areas. Instead, it only makes simple but effective modifications on motion vectors and macroblock modes for the affected macroblocks, and the quantization of these areas.

1. Modification of Motion Information

We assume, without losing the generality, that the logo part occupies an integer number of macroblocks, especially for covered logo. There is very little difference in modifying the motion information for these two types of logos. We will describe them in one piece. As metioned, some of the information decoded from the compressed bit stream is preserved in most of the transcoders, such as the picture type and the motion information. Since the logo usually only affects a small portion of a limited number of pictures in the video sequence, we reuse the old motion vectors for the part unaffected by the logo. So we only need to modify the motion vectors of the part affected by the logo, which includes the part covered by the logo and the part outside of the logo but in motion compensation used the area inside the logo (only for P, B frames). We name them "logo part" and "logo-affected part" respectively in the following discussion. We address on how to deal with these two situations separately.

    To work on the part of the picture affected the logo, first we need to find out the range of video images that is possibly affected by the logo. Without loss of generality, we considered open GOP structure in this paper, with a typical such sequence: BBIBBPBBP....... Our invention can also be used for closed GOP structure with little changes. Assume to insert a logo in the range [l, h] of the video sequence which starts from the sequence number zero, here, l and h are the lowest and highest frame numbers of the video sequence in display order. Thus, the possible range that affected by the logo [L, H] is decided by the following equations: L = Previous reference frame + 1;

H = Next reference frame if next reference frame is not I frame;

Next reference frame -1 otherwise.

Here, L is the B frame with the smallest frame number...