Moving picture encoding method and apparatus, moving picture decoding method and apparatus and recording medium

Claim 2

A picture encoding apparatus in which a picture constituting a moving picture is divided into an object picture in which the main concern is a picture making up the moving picture and the other picture, said object picture is split into a texture picture representing brightness and color hue of the picture and a shape picture representing the shape of an object, said texture picture and the shape picture being encoded independently, comprising: first motion vector detection means for detecting the motion vector of the shape picture and to output the motion vector of the shape picture; first encoding means for encoding said shape picture based on the motion vector of the shape picture to output encoded data of the shape picture, said first encoding means also encoding the motion vector of said shape picture to output encoded data of the motion vector of the shape picture; second motion vector detection means for detecting the motion vector of the texture picture and for outputting the motion vector of the texture picture; and second encoding means for encoding said texture picture based on the motion vector of said texture picture to output encoded data of said texture picture, said second encoding means also encoding the motion vector of the texture picture using the motion vector of the shape picture to output encoded data of the motion vector of the texture picture.

Claim 3

The picture encoding apparatus as claimed in claim 1 wherein said first and second motion vector detection means detects each motion vector from block to block; said first encoding means executing motion-compensated predictive coding using a locally decoded shape picture obtained on locally decoding encoded shape picture data to encode the difference shape motion vector obtained on calculating the difference between the motion vector of the shape picture of a block being encoded and the motion vector of the shape picture of a neighboring block; said second encoding means executing motion-compensated predictive coding using a locally decoded texture picture obtained on locally decoding encoded texture picture data to encode the difference texture motion vector obtained on calculating the difference between the motion vector of the texture picture of a block being encoded and the motion vector of the shape picture of the same block.

Claim 4

The picture encoding apparatus as claimed in claim 2 wherein said second motion vector detection means searches a range in the vicinity of a position specified by a motion vector of said shape picture to detect the motion vector of said texture picture.

Claim 5

The picture encoding apparatus as claimed in claim 2 wherein said first motion vector detection means has calculating means for calculating the evaluation value to detect the motion vector of said shape picture; said calculating means including first calculating means for calculating the evaluation value to detect the motion vector of the shape picture using plural pixels in a block of said shape picture being encoded and plural pixels in a block of said locally decoded shape picture; second calculating means for adding to said evaluation value a value taking into account the length of the motion vector of said shape picture to calculate a new evaluation value; and third calculating means for subtracting a pre-set value from said new evaluation value if the motion vector of said shape picture is of the same value as the motion vector of the shape picture of the neighboring block.

Claim 6

The picture encoding apparatus as claimed in claim 4 wherein said calculating means further includes fourth calculating means for calculating an evaluation value for detecting the motion vector of a texture picture using plural pixels in a block of the texture picture to be encoded and plural pixels in a block of said locally decoded texture pixel; said second calculating means summing to said evaluation value a value which takes into account the length of the motion vector of the shape picture and said evaluation value for detecting the motion vector of said texture picture to calculate a new evaluation value.

Claim 7

A picture encoding method in which a picture constituting a moving picture is divided into an object picture in which the main concern is a picture making up the moving picture and the other picture, said object picture is split into a texture picture representing brightness and color hue of the picture and a shape picture representing the shape of an object, said texture picture and the shape picture being encoded independently, comprising: a first motion vector detection step for detecting the motion vector of the shape picture and to output the motion vector of the shape picture; a first encoding step for encoding said shape picture based on the motion vector of the shape picture to output encoded data of the shape picture, said first encoding step also encoding the motion vector of said shape picture to output encoded data of the motion vector of the shape picture; a second motion vector detection step for detecting the motion vector of the texture picture and for outputting the motion vector of the texture picture; and a second encoding step for encoding said texture picture based on the motion vector of said texture picture to output encoded data of said texture picture, said second encoding means also encoding the motion vector of the texture picture using the motion vector of the shape picture to output encoded data of the motion vector of the texture picture.

Claim 8

A picture encoding method in which a picture constituting a moving picture is divided into an object picture in which the main concern is a picture making up the moving picture and the other picture, said object picture is split into a texture picture representing brightness and color hue of the picture and a shape picture representing the shape of an object, said texture picture and the shape picture being encoded independently, comprising: a first motion vector detection step for detecting the motion vector of the shape picture and to output the motion vector of the shape picture; a first encoding step for encoding said shape picture based on the motion vector of the shape picture to output encoded data of the shape picture, said first encoding step also encoding the motion vector of said shape picture to output encoded data of the motion vector of the shape picture; a second motion vector detection step for detecting the motion vector of the texture picture and for outputting the motion vector of the texture picture; and a second encoding step for encoding said texture picture based on the motion vector of said texture picture to output encoded data of said texture picture, said second encoding means also encoding the motion vector of the texture picture using the motion vector of the shape picture to output encoded data of the motion vector of the texture picture.

Claim 9

The picture encoding method as claimed in claim 6 wherein said first and second motion vector detection steps detecting each motion vector from block to block; said first encoding step executing motion-compensated predictive coding using a locally decoded shape picture obtained on locally decoding encoded shape picture data for encoding the difference shape motion vector obtained on calculating the difference between the motion vector of the shape picture of a block being encoded and the motion vector of the shape picture of a neighboring block; said second encoding step executing motion-compensated predictive coding using a locally decoded texture picture obtained on locally decoding encoded texture data for encoding the difference texture motion vector obtained on calculating the difference between the motion vector of the texture picture of a block being encoded and the motion vector of the shape picture of the same block.

Claim 10

The picture encoding method as claimed in claim 7 wherein said second motion vector detection step searches a range in the vicinity of a position specified by a motion vector of said shape picture to detect the motion vector of said texture picture.

Claim 11

The picture encoding method as claimed in claim 7 wherein said first motion vector detection step has a calculating step for calculating the evaluation value for detecting the motion vector of said shape picture; said calculating step including a first calculating sub-step for calculating the evaluation value for detecting the motion vector of the shape picture using plural pixels in a block of said shape picture being encoded and plural pixels in a block of said locally decoded shape picture; a second calculating sub-step for adding to said evaluation value a value taking into account the length of the motion vector of said shape picture to calculate a new evaluation value; and a third calculating sub-step for subtracting a pre-set value from said new evaluation value if the motion vector of said shape picture is of the same value as the motion vector of the shape picture of the neighboring block.

Claim 12

The picture encoding method as claimed in claim 9 wherein said calculating step further includes a fourth calculating sub-step for calculating an evaluation value for detecting the motion vector of a texture picture using plural pixels in a block of the texture picture to be encoded and plural pixels in a block of said locally decoded texture pixel; said second calculating sub-step summing to said evaluation value a value which takes into account the length of the motion vector of the shape picture and said evaluation value for detecting the motion vector of said texture picture to calculate a new evaluation value.

Claim 13

A picture decoding apparatus for decoding an encoded signal representing encoded moving picture signals, in which the encoded signal is made up of encoded data of a shape picture, encoded data of a motion vector of the shape picture, encoded data of a texture picture, and encoded data of a motion vector of the texture picture, and in which each of said encoded data is generated by dividing a picture constituting a moving picture into an object picture in which the main concern is a picture making up the moving picture and the other picture, splitting said object picture into a texture picture representing brightness and the color hue of the picture and a shape picture representing the shape of an object, detecting the motion vector of the shape picture to output the motion vector of the shape picture, encoding said shape picture based on the motion vector of the shape picture to output encoded data of the shape picture, encoding the motion vector of said shape picture based on the motion vector of the shape picture to output encoded data of the motion vector of the shape picture, encoding the motion vector of the shape picture to output encoded data of the motion vector of the shape picture, detecting the motion vector of the texture picture to output the motion vector of the texture picture, encoding said texture picture based on the motion vector of the texture picture to output encoded data of the texture picture and by encoding the motion vector of the texture picture using the motion vector of said shape picture to output encoded data of the motion vector of said texture picture, said apparatus comprising: first decoding means for decoding encoded data of the motion vector of the shape picture to output a decoded motion vector of the shape picture, said first decoding means also decoding encoded data of said shape picture based on the decoded motion vector of the shape picture to output a decoded shape picture; and second decoding means for decoding encoded data of the motion vector of the texture picture using the decoded motion vector of the shape picture to output a decoded motion vector of the texture picture, said second decoding means also decoding encoded data of said texture picture based on the decoded motion vector of the texture picture to output a decoded texture picture.

Claim 14

A picture decoding apparatus for decoding an encoded signal representing encoded moving picture signals, in which the encoded signal is made up of encoded data of a shape picture, encoded data of a motion vector of the shape picture, encoded data of a texture picture, and encoded data of a motion vector of the texture picture, and in which each of said encoded data is generated by dividing a picture constituting a moving picture into an object picture in which the main concern is a picture making up the moving picture and the other picture, splitting said object picture into a texture picture representing brightness and the color hue of the picture and a shape picture representing the shape of an object, detecting the motion vector of the shape picture to output the motion vector of the shape picture, encoding said shape picture based on the motion vector of the shape picture to output encoded data of the shape picture, encoding the motion vector of said shape picture based on the motion vector of the shape picture to output encoded data of the motion vector of the shape picture, encoding the motion vector of the shape picture to output encoded data of the motion vector of the shape picture, detecting the motion vector of the texture picture to output the motion vector of the texture picture, encoding said texture picture based on the motion vector of the texture picture to output encoded data of the texture picture and by encoding the motion vector of the texture picture using the motion vector of said shape picture to output encoded data of the motion vector of said texture picture, said apparatus comprising: first decoding means for decoding encoded data of the motion vector of the shape picture to output a decoded motion vector of the shape picture, said first decoding means also decoding encoded data of said shape picture based on the decoded motion vector of the shape picture to output a decoded shape picture; and second decoding means for decoding encoded data of the motion vector of the texture picture using the decoded motion vector of the shape picture to output a decoded motion vector of the texture picture, said second decoding means also decoding encoded data of said texture picture based on the decoded motion vector of the texture picture to output a decoded texture picture.

Claim 15

The picture decoding apparatus as claimed in claim 11 wherein the motion vector of said shape picture and the motion vector of said texture picture are each detected from block to block; the encoded data of the motion vector of the shape picture being obtained by calculating a difference between the motion vector of the shape picture of a block to be encoded and the motion vector of the shape picture of the neighboring block and encoding the resulting difference shape motion vector; the encoded data of the motion vector of the texture picture being obtained by calculating a difference between the motion vector of the texture picture of a block to be encoded and the motion vector of the shape picture of the same block and encoding the resulting difference shape motion vector; said first decoding means executing motion compensated predictive decoding using said shape decoded picture and summing the motion vector of the difference texture of the block to be encoded to the motion vector of the shape picture of the neighboring block to restore the motion vector of the shape picture; and said second encoding means executing motion compensated predictive decoding using said texture decoded picture and summing the motion vector of the difference texture of the block to be encoded to the motion vector of the shape picture of the neighboring block to restore the motion vector of the texture picture.

Claim 16

A picture decoding method for decoding an encoded signal representing encoded moving picture signals, in which the encoded signal is made up of encoded data of a shape picture, encoded data of a motion vector of the shape picture, encoded data of a texture picture, and encoded data of a motion vector of the texture picture, and in which each of said encoded data is generated by dividing a picture constituting a moving picture into an object picture in which the main concern is a picture making up the moving picture and the other picture, splitting said object picture into a texture picture representing brightness and color hue of the picture and a shape picture representing the shape of an object, detecting the motion vector of the shape picture to output the motion vector of the shape picture, encoding said shape picture based on the motion vector of the shape picture to output encoded data of the shape picture, encoding the motion vector of the shape picture to output encoded data of the motion vector of the shape picture, detecting the motion vector of the texture picture using the motion vector of the shape picture to output the motion vector of the texture picture, encoding said texture picture based on the motion vector of the texture picture to output encoded data of the texture picture and by encoding the motion vector of the texture picture using the motion vector of said shape picture to output encoded data of the motion vector of said texture picture, said apparatus comprising: a first decoding step for decoding encoded data of the motion vector of the shape picture to output a decoded motion vector of the shape picture, said first decoding step also decoding encoded data of said shape picture based on the decoded motion vector of the shape picture to output a decoded shape picture; and a second decoding step for decoding encoded data of the motion vector of the texture picture using the decoded motion vector of the shape picture to output a decoded motion vector of the texture picture, said second decoding step also decoding encoded data of said texture picture based on the decoded motion vector of the texture picture to output a decoded texture picture.

Claim 17

A picture decoding method for decoding an encoded signal representing encoded moving picture signals, in which the encoded signal is made up of encoded data of a shape picture, encoded data of a motion vector of the shape picture, encoded data of a texture picture, and encoded data of a motion vector of the texture picture, and in which each of said encoded data is generated by dividing a picture constituting a moving picture into an object picture in which the main concern is a picture making up the moving picture and the other picture, splitting said object picture into a texture picture representing brightness and color hue of the picture and a shape picture representing the shape of an object, detecting the motion vector of the shape picture to output the motion vector of the shape picture, encoding said shape picture based on the motion vector of the shape picture to output encoded data of the shape picture, encoding the motion vector of the shape picture to output encoded data of the motion vector of the shape picture, detecting the motion vector of the texture picture using the motion vector of the shape picture to output the motion vector of the texture picture, encoding said texture picture based on the motion vector of the texture picture to output encoded data of the texture picture and by encoding the motion vector of the texture picture using the motion vector of said shape picture to output encoded data of the motion vector of said texture picture, said apparatus comprising: a first decoding step for decoding encoded data of the motion vector of the shape picture to output a decoded motion vector of the shape picture, said first decoding step also decoding encoded data of said shape picture based on the decoded motion vector of the shape picture to output a decoded shape picture; and a second decoding step for decoding encoded data of the motion vector of the texture picture using the decoded motion vector of the shape picture to output a decoded motion vector of the texture picture, said second decoding step also decoding encoded data of said texture picture based on the decoded motion vector of the texture picture to output a decoded texture picture.

Claim 18

The picture decoding method as claimed in claim 13 wherein the motion vector of said shape picture and the motion vector of said texture picture are each detected from block to block; the encoded data of the motion vector of the shape picture being obtained by calculating a difference between the motion vector of the shape picture of a block to be encoded and the motion vector of the shape picture of the neighboring block and encoding the resulting difference shape motion vector; the encoded data of the motion vector of the texture picture being obtained by calculating a difference between the motion vector of the texture picture of a block to be encoded and the motion vector of the shape picture of the same block and encoding the resulting difference shape motion vector; said first decoding step executing motion compensated predictive decoding using said shape decoded picture and summing the motion vector of the difference texture of the block to be encoded to the motion vector of the shape picture of the neighboring block to restore the motion vector of the shape picture; and said second encoding step executing motion compensated predictive decoding using said decoded texture picture and summing the motion vector of the difference texture of the block to be encoded to the motion vector of the shape picture of the same block to restore the motion vector of the texture picture.

Claim 19

A recording medium decodable by a decoding device, said recording medium having recorded thereon encoded signals representing encoded moving picture signals, the encoded signal is made up of encoded data of a shape picture, encoded data of a motion vector of the shape picture, encoded data of a texture picture, and encoded data of a motion vector of the texture picture, and in which each of said encoded data is generated by dividing a picture constituting a moving picture into an object picture in which the main concern is a picture making up the moving picture and the other picture, splitting said object picture into a texture picture representing brightness and color hue of the picture and a shape picture representing the shape of an object, detecting the motion vector of the shape picture to output the motion vector of the shape picture, encoding said shape picture based on the motion vector of the shape picture to output encoded data of the shape picture, encoding the motion vector of said shape picture to output encoded data of the motion vector of the shape picture, detecting the motion vector of the texture picture using the motion vector of the shape picture to output the motion vector of the texture picture, encoding said texture picture based on the motion vector of the texture picture to output encoded data of the texture picture and by encoding the motion vector of the texture picture using the motion vector of said shape picture to output encoded data of the motion vector of said texture picture.

Claim 20

A recording medium decodable by a decoding device, said recording medium having recorded thereon encoded signals representing encoded moving picture signals, the encoded signal is made up of encoded data of a shape picture, encoded data of a motion vector of the shape picture, encoded data of a texture picture, and encoded data of a motion vector of the texture picture, and in which each of said encoded data is generated by dividing a picture constituting a moving picture into an object picture in which the main concern is a picture making up the moving picture and the other picture, splitting said object picture into a texture picture representing brightness and color hue of the picture and a shape picture representing the shape of an object, detecting the motion vector of the shape picture to output the motion vector of the shape picture, encoding said shape picture based on the motion vector of the shape picture to output encoded data of the shape picture, encoding the motion vector of said shape picture to output encoded data of the motion vector of the shape picture, detecting the motion vector of the texture picture using the motion vector of the shape picture to output the motion vector of the texture picture, encoding said texture picture based on the motion vector of the texture picture to output encoded data of the texture picture and by encoding the motion vector of the texture picture using the motion vector of said shape picture to output encoded data of the motion vector of said texture picture.

Claim 21

The recording medium as claimed in claim 15 wherein the motion vector of said shape picture and the motion vector of said texture picture are each detected from block to block; the encoded data of the motion vector of the shape picture being obtained by calculating a difference between the motion vector of the shape picture of a block to be encoded and the motion vector of the shape picture of the neighboring block and encoding the resulting difference shape motion vector; the encoded data of the motion vector of the texture picture being obtained by calculating a difference between the motion vector of the texture picture of a block to be encoded and the motion vector of the shape picture of the same block and encoding the resulting difference shape motion vector.