Television – Format conversion – Changing number of fields for standard conversion
Reexamination Certificate
2000-12-22
2003-08-26
Hsia, Sherrie (Department: 2614)
Television
Format conversion
Changing number of fields for standard conversion
C348S452000, C348S455000, C348S700000, C348S699000
Reexamination Certificate
active
06611294
ABSTRACT:
TECHNICAL FIELD
In general, the present invention relates to a frame count transformation method and a frame count transformation apparatus of a picture signal. More particularly, the present invention relates to a frame count transformation method of a picture signal and a frame count transformation apparatus adopting the method which are properly applicable to transformation of the number of frames of the picture signal by special signal processing such as movement correction.
BACKGROUND ART
With the development of multimedia technologies going on in recent years, it becomes necessary to provide the television receiver with a function for displaying a variety of signals with different picture formats such as television signals and PC picture signals of various systems. As a function to keep up with an increasing demand for high quality pictures and plain surface displays, a picture display function based on sequential scanning is required. In order to implement these functions, special signal processing such as frame count transformation to convert a variety of input signals into a format of a picture display unit and the sequential scanning is required.
With regard to the frame count transformation, there are provided simple signal processing methods such as frame repetition and frame dropping. However, these methods cause picture quality deterioration such as the so-called motion judder disturbance whereby the smooth motion of a moving picture is lost.
A frame count transformation method of a movement correction type has been introduced to eliminate such disturbances. With this method, the positions of pictures of the preceding and succeeding frames are moved by a movement vector in order to generate a signal of an interpolation frame. With regard to the method and the configuration of such signal processing, a number of proposals have been made. For example, in Japanese Patent Laid-open No. Hei-7-170496, there is disclosed a technology to efficiently search for a movement vector. In addition, in Japanese Patent Laid-open No. Hei-7-336650, there is disclosed a technology for preventing degradation such as resolution deterioration at edges of a moving picture which is inherent in movement correction.
In the conventional technology disclosed in Japanese Patent Laid-open No. Hei-7-170496, however, the precision of the movement detection is an issue while, in the technology disclosed in Japanese Patent Laid-open No. Hei-7-336650, the complexity of the signal processing is an issue. These issues become a big problem encountered in the implementation of a high-quality and low-cost frame count transformation apparatus.
It is thus an object of the present invention to provide a movement correction frame count transformation method and a movement correction frame count transformation apparatus of a picture signal that are capable of implementing signal processing to convert the number of frames of the picture signal with a high degree of quality and at a low cost.
DISCLOSURE OF INVENTION
The frame count transformation method provided by the present invention comprises the steps of:
detecting movement of a picture signal and searching for a block unit movement vector;
correcting the block unit movement vector in dependence on the magnitude of a movement correction error component of the block unit movement vector;
generating a movement vector selected on the basis of an error component between a movement correction signal of a current frame and a movement correction signal of an immediately preceding frame calculated by using the corrected block unit movement vector as a pixel unit movement vector;
generating a movement correction interpolation frame signal in dependence on the magnitude of an error component between a movement correction signal of a current frame and a movement correction signal of an immediately preceding frame calculated by using the pixel unit movement vector; and using the movement correction interpolation frame signal for carrying out conversion of the number of frames of a picture signal.
In addition, the frame count transformation apparatus provided by the present invention comprises:
a block unit movement vector searching unit for detecting movement of a picture signal and searching for a block unit movement vector;
a movement vector correcting unit for correcting the block unit movement vector in dependence on the magnitude of a movement correction error component of the block unit movement vector;
a pixel unit movement vector generating unit for generating a movement vector selected on the basis of an error component between a movement correction signal of a current frame and a movement correction signal of an immediately preceding frame calculated by using the corrected block unit movement vector as a pixel unit movement vector; and
a movement correction interpolation frame signal generating unit for generating a movement correction interpolation frame signal in dependence on the magnitude of an error component between a movement correction signal of a current frame and a movement correction signal of an immediately preceding frame calculated by using the pixel unit movement vector.
To put it concretely, in the present invention, the following technological means are used. To be more specific, in order to substantially reduce the amount of processing, in a search of movement vectors required in the transformation of the number of frames of a picture signal, signal processing is carried out at 3 stages, namely, a search for a block unit movement vector, correction of the block unit movement vector and generation of a pixel unit movement vector.
First of all, in the search for a block unit movement vector, preprocessing is carried out to distinguish a static picture block from a moving picture block. A static picture block is a block in which movement is not detected by a frame differential signal. On the other hand, a dynamic block is a block in which movement is detected by a frame differential signal. Only a moving picture block is subjected to one of the following 3 kinds of search processing.
[1] Reference vector search processing: Movement vectors of blocks adjacent to the current block are taken as reference vectors. A specific reference vector, for which a predicted error component is smallest among the reference vectors and is smaller than a threshold value, is identified and an area in close proximity to this specific reference vector is subjected to block matching processing in order to detect a movement vector of a block unit.
[2] Movement vector distribution adaptive search processing: In the case of a reference vector with a predicted error component equal to or greater than the threshold value, on the other hand, a movement vector of a block unit is detected by carrying out block matching processing in accordance with a search mode which is determined in dependence on a frequency of generation of movement vectors in an immediately preceding frame. To put in detail, a search area and a layout of representative movement vectors in the block matching area are changed depending on the search mode.
[3] Converted vector search processing: Movement vectors per frame generated in vector conversion processing from movement vector information used in a picture encoding process are taken as reference movement vectors and a reference movement vector with a smallest predicted error component is detected as a movement vector of a block unit.
Subsequently, in the correction of the block unit movement vector, the detected block unit movement vector with a predicted error component equal to or greater than a threshold value is divided into miniblocks in the horizontal and vertical directions. Then, correction is made by carrying out miniblock division search wherein one of movement vectors of a current block and blocks adjacent to the current block with a smallest predicted error component computed for the associated miniblocks is selected as a movement vector of the miniblocks.
Then, in the generation of a pixel unit movement
Hirano Yasuhiro
Ishikura Kazuo
Kanehachi Takashi
Matono Takaaki
Nakajima Mitsuo
Antonelli Terry Stout & Kraus LLP
Hitachi , Ltd.
Hsia Sherrie
LandOfFree
Method and device for converting number of frames of image... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and device for converting number of frames of image..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and device for converting number of frames of image... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3119082