JPH0646406A - Method and device for correcting error of digital image signal - Google Patents

Abstract

PURPOSE:To provide an error correcting device for digital images which detects and corrects the miscorrection of errors of the image coding data due to the noises, etc., and reduces the influences of these errors. CONSTITUTION:An error correcting device 3 is provided with an error miscorrection detector 1 which detects the miscorrection of the error of the coding data predicted from a movement compensated reference image and coded, and a decoder 2 which decodes the data with a predictive error signal set at zero when the error miscorrection is detected out of a predictive error signal train and with a movement vector and the estimated error signal set at zero when the error miscorrection is detected out of the movement vector respectively. Therefore the omission of the block data when detected can be replaced with a movement compensated block. Then the errors of the digital image signals can be corrected with reduced deterioration due to the errors of the natural movements.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a digital signal which corrects the digital signal for an error caused by noise or the like when transmitting and recording coded data obtained by coding a digital image signal to reduce the influence of the error. The present invention relates to an image signal error correction method and an error correction device.

[0002]

2. Description of the Related Art Before transmitting or recording encoded data obtained by encoding a digital image signal, it is common to add an error correction code to detect or correct an error in the encoded data caused by noise or the like. is there. When decoding encoded data to which an error correction code is added, an error correction circuit first detects or corrects an error.

When the error can be detected but the error cannot be corrected, it is possible to take a countermeasure against the error by reading the encoded data again if there is a time margin. is there. However, if there is not enough time, only information that an error has occurred can be obtained. In this case, a countermeasure is taken by interpolating from the adjacent digital image signals.

Further, in encoding a digital image signal composed of a plurality of temporally continuous images, one or a plurality of images in the digital image signal is divided into blocks composed of a plurality of pixels, and each block is divided into blocks. A method of performing motion compensation to form a reference image, predicting an image to be encoded from the reference image, and encoding a motion vector and a prediction error signal of each block used for motion compensation is generally performed. There is. In this case, a decoding error may propagate into a block and cause block data to be lost, but measures are taken against block block data loss by interpolating from a temporally previous or subsequent image. .

[0005]

However, in the case of a moving image signal sequence in which the image region is moving with the passage of time, when data loss in block units occurs due to an error during decoding, it is possible to simply If interpolation is performed in the subsequent image area, the motion becomes unnatural, and there is a problem that the image quality deteriorates significantly.

In view of the above point, the present invention replaces with block data motion-compensated by a motion vector when data loss in block units occurs due to noise or the like, and performs error correction with little deterioration in the loss of block data. It is an object to provide a method and an error correction device.

[0007]

According to the error correcting method of the present invention, when one or a plurality of first images in a digital image signal sequence consisting of a plurality of images are to be encoded, the digital image signal One or a plurality of second images in a row is divided into blocks composed of a plurality of pixels, and each block is motion-compensated from the first image to form a reference image, and the first image is aforesaid After detecting and correcting an error in the coded data due to noise when decoding a coded data sequence that is predicted from a reference image and coded the motion vector data of the block used for the motion compensation and the prediction error signal A method for correcting the digital image signal sequence erroneously decoded due to a correction error, wherein when the error correction error is detected in the prediction error signal, the prediction error signal is decoded as zero. And if the correct leakage of error is detected in the motion vector is a method of decoding the motion vector and the prediction error signal as a zero.

Further, the error correction device of the present invention, when one or a plurality of first images in a digital image signal sequence consisting of a plurality of images are to be encoded, one of the digital image signal sequences One or a plurality of second images is divided into blocks made up of a plurality of pixels, motion compensation is performed from the first image for each block to form a reference image, and the first image is predicted from the reference image. However, when decoding a coded data string obtained by coding the motion vector data of the block used for the motion compensation and the prediction error signal, an error in the coded data due to noise or the like is detected and corrected. A device that corrects the digital image signal sequence that is erroneously decoded, a correction error detector that detects an error correction error of the encoded data, and the prediction error by the correction error detector. If the error correction omission is detected in the signal, the prediction error signal is decoded as zero, and if the error correction omission is detected in the motion vector, the motion vector and the prediction error signal are set to zero. It is composed of a decoder for decoding.

[0009]

As a result, even if data is lost in units of blocks due to noise or the like, replacement with block data that has been motion-compensated by a motion vector allows correction with little deterioration to the loss of block data. be able to.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of an error correction device for implementing the error correction method of the present invention will be described in detail below with reference to FIG. Here, a method of encoding a digital image signal sequence consisting of a plurality of temporally continuous images is divided into one or a plurality of images in the digital image signal into blocks composed of a plurality of pixels, and each block is divided into blocks. In this method, motion compensation is performed to form a reference image, an image to be encoded is predicted from the reference image, and a motion vector and a prediction error signal of each block used for motion compensation are encoded.

FIG. 1 shows an error correction leak detector 1 and a decoder 2.
It shows an error correction device 3 composed of and an error correction circuit 4, an image storage device 5, a motion compensation device 6, and a decoding device composed of an adder 7. The error correction circuit 4 divides the encoded data into segments having a predetermined number of bits, and detects and corrects an error in the encoded data due to noise or the like using the error correction code added to each segment. The error-correction-leakage detector 1 detects an error-correction-leakage when an error can be detected but cannot be corrected.

When it is detected that error correction omission has occurred in the motion vector data, the decoder 2 decodes the prediction error signal as zero even if it can be correctly decoded, and the motion vector data is decoded. Also decrypts as zero.
That is, in this case, the motion compensation is not performed, and the block is replaced with the block at the same position in the reference image. When the motion vector is correctly decoded, but the prediction error signal sequence has error correction omission, the prediction error signal sequence is decoded as zero. That is, in this case, the block of the reference image for which motion compensation has been performed is replaced. As described above, when the motion vector is correctly decoded, the image quality deterioration due to a decoding error is significantly improved by substituting the motion-compensated image block.

Next, using the motion vector decoded by the decoder 2, the motion compensation device 6 motion-compensates the image which is stored in the image storage device 5 and is temporally preceding or posterior in time to form a reference image. To do. Next, the prediction error signal decoded by the decoder 2 and the reference image are added by the adder 7 to obtain a reproduced image.

It should be noted that the motion vector and the prediction error signal must be distinguished from each other to detect the error correction omission, but the coded data of the prediction error signal is arranged after all the motion vector data of one image. Then, the motion vector data and the prediction error signal can be distinguished. That is, motion vector data is collected at the beginning of the encoded data of each image. Although error detection and correction of encoded data are performed in units of a segment of a plurality of bits, one or several segments from the beginning of encoded data of an image is occupied by motion vector data or the like. Therefore, it is possible to detect whether or not there is an error correction omission in the motion vector based on whether or not the error correction omission occurs in the data of one or several segments from the head of the encoded data of the image.

As described above, according to the present embodiment, when the block data is missing, the block is replaced by the motion-compensated block, so that the error can be corrected with a natural motion and less deterioration.

In order to distinguish the motion vector from the prediction error signal, the motion vector is collected at the beginning of the coded data of each image, but the present invention is not limited to this, and the data is identified by the added identification code. Anything is acceptable as long as it can distinguish the motion vector and the prediction error signal.

[0017]

As described above, according to the error correction method of the present invention, when one or a plurality of first images in a digital image signal sequence consisting of a plurality of images are to be encoded, the digital image One or a plurality of second images in the signal sequence is divided into blocks composed of a plurality of pixels, and a motion compensation is performed from the first image to form a reference image for each block, and the first image is Predicting from the reference image, and detecting and correcting an error in the coded data due to noise or the like when decoding a coded data sequence in which the motion vector data of the block used for the motion compensation and the prediction error signal are coded A method of correcting the digital image signal sequence that is erroneously decoded due to a correction error, and when the error correction error is detected in the prediction error signal, the prediction error signal is decoded as zero. In addition, when an error correction error is detected in the motion vector, the motion vector and the prediction error signal are decoded as zero, so that the motion compensation is performed by the motion vector even if the block data is lost. By replacing the block data with the above-described block data, it is possible to perform error correction with a natural movement and less deterioration.

Further, the error correction apparatus of the present invention, when one or a plurality of first images in a digital image signal sequence consisting of a plurality of images are to be encoded, one of the digital image signal sequences is One or a plurality of second images is divided into blocks made up of a plurality of pixels, motion compensation is performed from the first image for each block to form a reference image, and the first image is predicted from the reference image. However, when decoding a coded data string obtained by coding the motion vector data of the block used for the motion compensation and the prediction error signal, an error in the coded data due to noise or the like is detected and corrected, and A device that corrects the digital image signal sequence that is erroneously decoded, a correction error detector that detects an error correction error of the encoded data, and the prediction error by the correction error detector. If the error correction error is detected in the signal, the prediction error signal is decoded as zero, and if the error correction error is detected in the motion vector, the motion vector and the prediction error signal are changed to zero. By configuring with a decoder for decoding, even if a block data loss occurs, it is possible to perform error correction with less natural deterioration of motion by replacing with block data that has been motion-compensated by a motion vector. it can.

[Brief description of drawings]

FIG. 1 is a block diagram of a decoding device using an error correction device according to a first embodiment of the present invention.

[Explanation of symbols]

 1 Error Correction Leakage Detector 2 Decoder 3 Error Correction Device 4 Error Correction Circuit 5 Image Storage Device 6 Motion Compensation Device 7 Adder

Claims (2)

[Claims] 1. When one or a plurality of first images in a digital image signal sequence consisting of a plurality of images are to be encoded, one or a plurality of second images in the digital image signal sequence. Is divided into blocks made up of a plurality of pixels, motion compensation is performed from the first image for each block to form a reference image, the first image is predicted from the reference image, and the reference image is used for the motion compensation. After detecting and correcting an error in the coded data due to noise or the like when decoding the coded data sequence in which the motion vector data of the block and the prediction error signal are coded, the digital image erroneously decoded by the correction error A method of correcting a signal sequence, wherein when a prediction error error is detected in the prediction error signal, the prediction error signal is decoded as zero, and an error correction error is detected in the motion vector. Error correction method when it is detected the digital image signal, characterized by decoding the motion vector and the prediction error signal as a zero. 2. When one or a plurality of first images in a digital image signal sequence consisting of a plurality of images are to be encoded, one or a plurality of second images in the digital image signal sequence. Is divided into blocks made up of a plurality of pixels, motion compensation is performed from the first image for each block to form a reference image, the first image is predicted from the reference image, and the reference image is used for the motion compensation. Detecting and correcting an error in the coded data due to noise when decoding a coded data sequence in which motion vector data of a block and a prediction error signal are coded, and the digital image signal erroneously decoded due to correction error A device for correcting a column, wherein a correction leak detector for detecting error correction leak of the encoded data, and error correction leak for the prediction error signal is detected by the correction leak detector. In the case where the prediction error signal is decoded as zero, and when the error correction error is detected in the motion vector, the motion vector and the prediction error signal are decoded as zero. Characteristic error correction device.