FR2701623A1 - Method of recording reduced-rate images and device for implementing it. - Google Patents

Abstract

The method according to the invention consists in recording a prediction error signal (e), corresponding to the difference between a real image (X) to be recorded and an image predicted from the previous image, using a main circuit (1) coding-decoding comprising means for recording and reading a digital recorder 6, to (CF DRAWING IN BOPI) single inter-picture prediction loop (2), in this loop (2), to be predicted the following image from the previous image taken, and reinjecting the image resulting from the prediction (P) both at the input and at the output of the main circuit (1) respectively by subtracting the image resulting from the prediction (P) to the real image (X) to be recorded and adding the same image to the error signal read and decoded (ed) in the main circuit (1) to reconstitute (CF DRAWING IN BOPI) recording. The applications go to digital recorders, and in particular to high definition digital video recorders.

Description

Method of recording reduced-rate images and device
for its implementation
The present invention relates to a reduced-rate image recording method and a device for its implementation. Applications include digital recorders for high definition television.

 Known methods for the transmission and recording of video signals and in particular television images use image coding and decoding devices called CODEC, abbreviation for coder decoder.

 These devices use known image coding techniques exploiting the principle of flow compression.

Much of the information in an image is redundant. Redundancy can be defined as the part of an unnecessary message in the sense that if it were missing, the message could not be reconstructed. Most scanning systems for example type
MIC abbreviation for Coded Pulse Modulation, do not take into account the intrinsic properties of the image transmitter, or source, nor those of the visual receiver, for example the eye.

 Off, the transmitted signal is not random. There exists between each image sample, representing one or more pixels, properties of correlations. Moreover, the eye does not perceive all information in the same way.

 Among the coding techniques using rate compression, the psychovisual codings take into account properties related to the visual receiver, that is to say the eye of the observer. They consist in deceiving the eye as best as possible, by restoring a signal different from the original signal, but containing the relevant information for the observer. This type of coding is irreversible.

 The results obtained correspond to reconstructed images of good quality exploitable for example by television studios for image processing.

 In psychovisual encoding techniques, the adaptive differential coding is not to encode the video signal itself, but the difference between that signal and a prediction signal that is known to the receiver.

 A known type of architecture of an adaptive differential coding device for the transmission of images uses two inter-image prediction loops, also called inter loops, at the input and at the output of the main coding and decoding circuit.

 The purpose of each loop is to reconstruct an estimated image from the current image taken by each loop. The reconstructed image is then compared to the next image when it is transmitted to the input and output of the transmission device. The input inter loop generally comprises from the input to the output, a decoder, called a local decoder, an addition operator and an image prediction device.

 The image taken by the input loop is decoded by the local decoder and then added to the estimated prediction signal from a previous image to reconstruct an image that is compared to the current image at the input of the device. transmission. The image prediction device comprises an image memory and optionally an image estimator with motion compensation.

 The signal transmitted to the input of the transmission channel therefore corresponds to a coded error signal resulting from the difference between the current image present at the input of the device and the prediction signal. At the output of the transmission channel, the error signal is decoded by a second decoder, called remote decoder, then injected on the second loop which adds to the decoded error signal a prediction signal identical to that delivered by the first loop. . The prediction signal of the second loop is estimated from the previous transmitted image, reconstituted at the output of the device, taken by the second loop.

 To record or store one or more images, it is known to insert in the transmission channel, a digital image recorder for example a digital video recorder or an image memory for the storage of images in the medical field, etc. ...

 The image transmission devices generally comprise, at the input and at the output of the transmission channel, a device for adapting the flow rate between the codageldecoding main circuit and the transmission channel, which also integrates a function for correcting errors due to transmission through the transmission channel. Likewise, digital recorders are generally provided with input and output of error correction devices.

 However, the errors due to the recording technology used by the recorder, for example resulting in a loss of digital information, or bytes, are not taken into account by the type of error correction devices mentioned above. This type of architecture for the recording of images is therefore not very robust with respect to the errors due to the recording. On the other hand, the hardware used for the recording quickly becomes very large and the error correction devices, mainly based on the introduction of a redundancy of digital information are very consumption of flow.

 The invention aims to overcome the aforementioned drawbacks.

 For this purpose, the subject of the invention is a method for recording images with reduced bit rate, using coding and differential decoding with inter-image adaptive prediction, characterized in that it consists, after having introduced means for recording and reading digital information in a main coding and decoding circuit of images, recording a prediction error signal resulting from the comparison between an image to be recorded and a predicted image from the read image on at least one recording medium, and to read after recording, the prediction error recorded on the recording medium, and in that it consists in reconstructing an image from the error signal of prediction, read and decoded at the output of the main circuit and an image derived from a single prediction delivered by a single prediction loop taking the reconstructed preceding image, and comparing the image resulting from the has the same prediction with the next image to be recorded at the input of the main circuit taking into account the recording errors to avoid their propagation from one image to another.

Other features and advantages of the invention will appear more clearly in the following description given by way of non-limiting example and with reference to the appended figures which represent:
- Figure 1, a block diagram of an embodiment of a
device for implementing the method according to the invention, and
FIG. 2, an example of transmission means in a mode
embodiment of the device for implementing the method
according to the invention.

 The device according to the invention has the advantages that it makes it possible on the one hand to reduce the overall size of the image recording device by using a single decoder, a single prediction loop and, by including the data channel. record playback of a digital recorder in the main coding and decoding circuit. It also makes it possible to correct registration errors so that they do not propagate from one image to another.

 An embodiment of a device for implementing the method according to the invention is illustrated in FIG. 1.

 It uses the techniques of psychovisual coding and in particular differential coding with inter-image prediction. This embodiment comprises a main circuit 1 coding and decoding delimited by a discontinuous closed line and a single loop 2 also delimited by a discontinuous closed line which reinjects on the input and the output of the main circuit 1 a portion of the signal of output of the device according to the invention.

 A video signal X is injected at the input of the main circuit 1 on a first operand input of a subtraction operator 3. The video signal X is for example a sequence of television images. The output signal e of the subtraction operator 3 is the result of the subtraction of a current image X injected at the input of the main circuit 1 with a prediction P coming from the prediction loop 2. The corresponding signal e a prediction error signal is injected at the input of a coding means 4, or coder, digitally coding the signal e received at the input of the encoder 4.

The output of the encoder 4 is coupled to first means 5 variable length coding and error correction, hereinafter referred to as VLC abbreviation for Variable Length Coding, to adapt the flow of digital information from the encoder 4 to that of recording and reading means 6 at the input of which the output of the encoder 4 is coupled.

 The output of the recording and reading means 6 is coupled to the input of a second VLC 7, performing the inverse function of the first VLC 5, which is also associated with an error correction function. The output of the second VLC 7 is coupled to the input of a decoding means, or decoder 8, the output of which is outputted the error signal ed. The output of the decoder 8 is coupled to a first operand input of an addition operator 9.

 The image X reconstituted at the output of the addition operator 9 is picked up by the loop 2 and injected at the input of prediction means 10 comprising an image storage device, not shown, possibly associated with a device of FIG. estimation, also not shown, performing motion compensation between two consecutive images. The output of the prediction means 10 is coupled to both the input and the output of the main circuit 1 respectively by a second operand input of the subtraction operator 3 and by a second operand input of the addition operator 9. The image k thus reconstituted takes into account the recording errors between two consecutive images.

 An embodiment of the recording and reading means 6 of a device for implementing the method according to the invention is illustrated in FIG. 2.

 This embodiment comprises an input E common to a first channel 11, called the direct channel, and to a second channel 12, called the recording channel. The output of the recording channel 12 and the output of the direct channel 11 are respectively coupled to a first input RAN and a second input M of a switching means 13 making it possible to select one of the two channels 11 and 12 for recording followed by the reading or for the direct reading of a signal between the input E and the output S of the recording and reading means 6.

The two channels 11 and 12 can each be assimilated to a small transmission channel belonging to a recording device using only one decoder, both local and remote. On the other hand, the recording device according to the invention comprising this embodiment makes it possible to integrate the coding and decoding functions in the same box.

Thus the notions of "local" and "remote" decoders known in the transmission devices lose their meaning.

 In a recorder type VCR, a device commonly known in the English language "scanner" generally comprises separate writing heads and reading. This makes it possible to check, in operating mode "Read after Write" English name for reading after writing, the validity of the recording in progress at any time.

In this mode of operation, the read and write heads continuously scan the recording medium, for example a magnetic tape, thus allowing the instantaneous reading of what has just been recorded on the recording medium. In operating mode "monitoring" English name meaning "control", the recording function of the video recorder is disconnected.

In this operating mode only the direct playback function via the VCR's direct playback channel is operational for controlling the information to be recorded.

 The modes "Read after Write" RNV and "monitoring" M are selected by the switching device 13. In the operating mode "Read after Write" R! VV of the recorder, the recording channel read 12 allows to write digital information on the recording medium, such as a magnetic tape, and then read this information almost instantly after recording on the tape.

 The information read on the magnetic tape is then transmitted through the magnetic recording channel 12 of the video recorder, used as a transmission channel, at the output S of the recording and reading means 6. The information transmitted suffers a simple delay due to the processing which in most cases, for television images for example, is of the order of the duration of a video frame is about 20 ms. This delay is compensated in loop 2.

 Since the recording channel 12 is part of the main coding / decoding circuit 1, the errors related to the recording are taken into account by the device at the time of reading, or when the information is transmitted to the output S of the recording means. recording and playback 6.

 In the "monitoring" operating mode M, the recording recording channel 12 is disconnected. The direct reading channel 11 of the recorder selected by the switching device 13 is then used as a transmission channel to allow the device according to the invention to continue to operate in a closed loop.

Claims (6)

 A method for recording reduced-rate images, using coding and differential decoding with inter-image adaptive prediction, characterized in that it consists, after introducing means for recording and reading digital information. in a main picture coding and decoding circuit (1), recording a prediction error signal (e) resulting from the comparison between an image (X) to be recorded and a predicted image from the image previously read on at least one recording medium, and to read after recording, the prediction error (ed) recorded on the recording medium, and in that it consists in reconstructing an image from the signal of prediction error (ed), read and decoded at the output of the main circuit (1) and of an image derived from a single prediction (P) delivered by a single prediction loop (2) taking the reconstituted preceding image , and compare the image from the m same prediction (P) with the next image to be recorded at the input of the main circuit (1) taking into account the recording errors to avoid their propagation from one image to another.  2. Device for implementing the method according to claim 1, characterized in that it comprises a main circuit (1) comprising in the order, an image coding means (4) coupled to first means ( 5) adapted to rate and error correction coupled to the input of the digital information recording and reading means (6), coupled to an output at second rate matching means (7) and error correction and decoding means (8), and in that it comprises a single inter-image prediction loop (2) taking a reconstituted image (X) at the output of the device and comprising prediction means ( 10) for predicting the next image from the reconstructed image (X) taken by the loop (2), means (3) for subtracting the image from the prediction (P) to the actual image to be recorded , and means (9) for adding to the prediction error (ed), after decoding, the image resulting from the same prediction tion (P).  3. Device according to claim 2, characterized in that the prediction means (10) comprise an image storage device.  4. Recorder according to claim 3, characterized in that the prediction means (10) further comprises a motion estimation device for compensating the movements between two consecutive images to be recorded.  5. Device according to any one of claims 2 to 4, characterized in that the recording and reading means (6) comprises an input common to at least a first channel (11), called direct reading channel of a digital recorder, and at least one second channel (12), called the recorder channel of a digital recorder, the output of each channel (11 and 12) is coupled respectively to a first input and a second input of switching means (13) for selecting one of the two channels (11 and 12).  6. Device according to claim 5, characterized in that the first channel (11) is a direct reading channel of a digital video recorder, in that the second channel (12) is a magnetic recording channel for playback of the digital video recorder using a magnetic tape as recording medium, and in that the switching means (13) is a switch for selecting a first mode of operation of the digital video recorder, called "monitoring" mode (M) using the direct channel (11) , or a second mode of operation of the digital video recorder, called "Read after Write "(RNV) using the magnetic recording channel (12).