Prediction of an image by compensation during forward movement

A method comprising:

• encoding a sequence of images, which implements, for a current image of said sequence, a method of predicting an image by forward motion compensation from at least one reference image, delivering a predicted image, wherein, for at least one block of a reference image, called a reference block, said method of predicting comprises:
• determining at least one piece of shift information, representing a shift between the reference block and a corresponding block in the image to be predicted;
• determining at least one enlarged block, comprising said reference block and at least one point of a block adjacent to said reference block in said reference image;
• and
• determining a value of prediction for at least one point of the image to be predicted, from at least one enlarged block and from the piece or pieces of shift information determined for said reference block.

A device for encoding a sequence of images, implementing a prediction by forward motion compensation from at least one reference image, wherein, for at least one block of a reference image, called a reference block, said device comprises:

• means for determining at least one piece of shift information, representing a shift between the reference block and a corresponding block in the image to be predicted;
• means for determining at least one enlarged block, comprising said reference block and at least one point of a block adjacent to said reference block in said reference image;
• and
• means for determining a value of prediction for at least one point of the current image from said at least one enlarged block and said piece or pieces of shift information determined for said reference block, delivering a predicted image.

A method comprising:

• generating a signal representing a sequence of images, comprising by predicting an image by forward motion compensation from at least one reference image, wherein, for at least one block of a reference image, called a reference block, said predicting comprises the following steps:
• determining at least one piece of shift information, representing a shift between the reference block and a corresponding block in the image to be predicted;
• determining at least one enlarged block, comprising said reference block and at least one point of a block adjacent to said reference block in said reference image;
• and
• determining a value of prediction for at least one point of the image to be predicted, from at least one enlarged block and from the piece or pieces of shift information determined for said reference block,
• wherein the signal comprises at least one piece of information representing the use of at least one enlarged block and/or of the enlargement assigned to the enlarged block for a given reference block:
• and transmitting said signal.

A method comprising:

• decoding a signal representing a sequence of images, which implements a method of predicting an image by forward motion compensation from at least one reference image, delivering a predicted image, wherein, for at least one block of a reference image, called a reference block, said method of predicting comprises:
• determining at least one piece of shift information, representing a shift between the reference block and a corresponding block in the image to be predicted;
• determining at least one enlarged block, comprising said reference block and at least one point of a block adjacent to said reference block in said reference image;
• and
• determining a value of prediction for at least one point of the image to be predicted, from at least one enlarged block and from the piece or pieces of shift information determined for said reference block.

A device for decoding a signal representing a sequence of images, implementing a prediction by forward motion compensation from at least one reference image, wherein, for at least one block of a reference image, called a reference block, said device comprises:

• means for determining at least one piece of shift information, representing a shift between said reference block and a corresponding block in an image to be predicted;
• means for determining at least one enlarged block, comprising said reference block and at least one point of a block adjacent to said reference block in said reference image;
• and
• means for determining a value of prediction for at least one point of the image to be predicted, from said at least one enlarged block and said piece or pieces of shift information determined for said reference block, delivering a predicted image.

A non-transitory computer-readable medium comprising instructions stored thereon for implementing a method of predicting an image by forward motion compensation from at least one reference image when this program is executed by a processor, wherein, for at least one block of a reference image, called a reference block, said method comprises:

• determining at least one piece of shift information, representing a shift between the reference block and a corresponding block in the image to be predicted;
• determining at least one enlarged block, comprising said reference block and at least one point of a block adjacent to said reference block in said reference image;
• and
• determining a value of prediction for at least one point of the image to be predicted, from at least one enlarged block and from the piece or pieces of shift information determined for said reference block.