Picture signal encoding system, picture signal decoding system and picture recognition system

• US 5,959,672 A
• Filed: 09/27/1996
• Issued: 09/28/1999
• Est. Priority Date: 09/29/1995
• Status: Expired due to Term

• Current Assignee: Nippondenso Co., Ltd. (DENSO Corporation)
• Original Assignee: Nippondenso Co., Ltd. (DENSO Corporation)
• Inventors: Sasaki, Mikio
• Primary Examiner(s): BRITTON, HOWARD W
• Application Number:
  US08/722,559
• Time in Patent Office: 1,096 Days
• Field of Search: 348/390, 348/414, 348/417, 348/418, 348/422, 382/243
• US Class Current: 375/240.23
• CPC Class Codes: H04N 19/107 between spatial and temporal predictive coding, e.g. picture refresh
• CPC Class Codes: H04N 19/115 Selection of the code volume for a coding unit prior to coding
• CPC Class Codes: H04N 19/124 Quantisation
• CPC Class Codes: H04N 19/13 Adaptive entropy coding, e.g. adaptive variable length coding [AVLC] or context adaptive binary arithmetic coding [CABAC]
• CPC Class Codes: H04N 19/137 Motion inside a coding unit, e.g. average field, frame or block difference
• CPC Class Codes: H04N 19/146 Data rate or code amount at the encoder output
• CPC Class Codes: H04N 19/152 by measuring the fullness of the transmission buffer
• CPC Class Codes: H04N 19/17 the unit being an image region, e.g. an object
• CPC Class Codes: H04N 19/176 the region being a block, e.g. a macroblock
• CPC Class Codes: H04N 19/20 using video object coding
• CPC Class Codes: H04N 19/44 Decoders specially adapted therefor, e.g. video decoders which are asymmetric with respect to the encoder
• CPC Class Codes: H04N 19/503 involving temporal prediction adaptive coding with adaptive selection between spatial and temporal predictive coding H04N19/107; adaptive coding with adaptive selection among a plurality of temporal predictive coding modes H04N19/109
• CPC Class Codes: H04N 19/537 Motion estimation other than block-based
• CPC Class Codes: H04N 19/61 in combination with predictive coding
• CPC Class Codes: H04N 19/91 Entropy coding, e.g. variable length coding [VLC] or arithmetic coding

Abstract

A picture signal encoding system capable of transmitting a motion picture at an extremely low rate while sufficiently controlling the occurrence of a retransmission mode and the occurrence of a picture freeze. The encoding system comprises an encoding control section, an attribute memory, an attribute prediction section and an area extraction and recognition processing section. For realizing the extremely low rate transmission, the encoder system integrally performs the alteration of the syntax, the substitute of the code word, the adaptive control for the prediction of the current frame encoding attribute based on the past encoding attribute and the attribute decision, the object area extraction based on the motion and a model, the area-separated quantization control, and the control of the necessary number of transformation coefficients according to the use mode, the transmission rate and the motion occurrence quantity. A protocol transformer is provided to match the current picture compression standard (H.261), so that the encoding system can be made with a simple arrangement.

Claims

1. A picture signal encoding system which compresses a picture signal with a variable-length code to produce and transmit encoded bit stream data, said system comprising:

mode control means retaining a mode condition for setting an incidental situation with a predetermined code, said mode control means performing an encoding processing of said picture signal under said mode condition; and

mode setting means for adding a code corresponding to said mode condition to said encoded bit stream data when said mode control means executes said encoding processing under said mode condition.

2. A system as defined in claim 1, further comprising:

person information storage means for storing and reading out data corresponding to a person in a scene; and

background information storage means for storing and reading out background information located at a stationary area other than an object area in said scene,

wherein said mode control means sets a use environment mode which makes as said incidental situation a limited condition corresponding to a use environment in the form of a code, and performs an encoding processing on the basis of said limited condition in said use environment mode and further on the basis of necessary data of one said person information storage means and said background information storage means.

3. A system as defined in claim 2, wherein said mode control means is constructed to set as said use environment mode a fixed mode and a movable mode for each of an interior of a motor vehicle, an indoor spot and an outdoor spot,

said fixed mode for said motor vehicle interior and said indoor place performing said encoding processing by setting a person mode for when a person is in a main area,

wherein said movable mode for said motor vehicle interior and said indoor place performs said encoding processing by setting a landscape mode for a landscape in addition to said person mode,

said fixed mode for said outdoor place performs said encoding process to enable wire and wireless transmissions, and

said movable mode for said outdoor place performs said encoding processing to enable a wireless transmission.

4. A system as defined in claim 1, further comprising:

person information storage means for storing and reading out data corresponding to a person in a scene; and

background information storage means for storing and reading out background information located at a stationary area other than an object area in said scene,

wherein said mode control means sets a subject mode which makes as said incidental situation a limited condition corresponding to a subject in the form of a code, and performs an encoding processing on the basis of said limited condition in said subject mode and further on the basis of necessary data of one said person information storage means and said background information storage means.

5. A system as defined in claim 4, wherein said mode control means sets as said subject mode a person mode, a landscape mode and a stationary object mode,

wherein each of said person mode and said landscape mode makes a limited condition presupposing the existence of a person and a landscape in a main area of a scene in the form of a code to perform said encoding process under said condition, and

said stationary object mode sets one of a motion picture mode and a high-resolution still picture mode to perform said encoding processing.

6. A system as defined in claim 5, wherein in said person mode said mode control means performs the switching among a head portion mode, an upper-half portion mode, a full-length mode and a plural-person mode in accordance with a distance up to a person existing in a main area and adds the corresponding mode information for said encoding processing.

7. A system as defined in claim 1, wherein said mode control means sets an encoding control mode in which said encoding processing is executed under a specific condition for limiting information quantity during encoding of said picture signal.

8. A system as defined in claim 7, wherein in said encoding control mode said mode control means sets a motion picture mode, a quasi-motion picture mode, a still picture mode and a background mode, and the mode setting is selectively carried into effect in accordance with a transmission condition and a picture quality condition to perform said encoding processing.

9. A system as defined in claim 7, wherein said mode control means sets as said encoding control mode a scene center preference mode to perform said encoding processing in which a larger quantity of bits are assigned to a center area of a scene.

10. A system as defined in claim 7, wherein said mode control means sets as said encoding control mode an object area preference mode on the basis of area information set in such a way that a specific area of a scene given in a different mode setting is treated as an object area, and said encoding processing is executed in a state that a larger quantity of bits are assigned to said object area.

11. A system as defined in claim 10, wherein said mode control means sets as said encoding control mode a motion preference mode for maintaining a frame rate in such a manner that said scene center preference mode and said object area preference mode are selectively switched to suppress the occurrence of a code quantity in accordance with an occurrence quantity of a motion area in said scene.

12. A system as defined in claim 7, wherein said mode control means sets three-dimensional model data previously indicative of a shape of an object through a three-dimensional rough model, and sets a model-based mode to produce as encoding data positional information and color information relative to said three-dimensional data.

13. A system as defined in claim 12, wherein in said model-based mode said mode control means calculates distance information on the basis of a two-dimensional template corresponding to a scene set in advance in producing a model-based prediction picture of said object.

14. A system as defined in claim 13, wherein in said model-based mode said mode control means previously sets a feature area of said two-dimensional template so that a model-based prediction is made by roughly estimating a positional attitude of said three-dimensional model from moved position data of said feature area on the basis of two-dimensional motion information.

15. A system as defined in claim 12, wherein said mode control means transmits forced INTRA frame data after the start of transmission of a picture signal and before the start of conversation and subsequently obtains necessary data in said model-based mode to make a model-based prediction of an object frame immediately before the start of the conversation.

16. A system as defined in claim 1, wherein said mode control means determines a transition from a set mode state to a different mode state in accordance with a given transition decision condition.

17. A system as defined in any one of claims 1 to 16, wherein said mode control means is made to allow manual and forced setting of a transition from a set mode state to a different mode state.

18. A system as defined in claim 16, wherein said mode control means determines said transition from said set mode state to said different mode state in accordance with a decision made according to an analysis of said picture signal.

19. A system as defined in claim 16, wherein, when information necessary for said decision according to said picture signal analysis lacks, said mode control means determines said transition from said set mode state to said different mode state on the basis of a transition probability value set in advance according to its level.

20. A system as defined in claim 19, wherein, when an evaluation calculation is made of the mode transition decision based on said transition probability value and shows that an error value goes large, said mode control means updates said transition probability value so that said error value becomes smaller.

21. A system as defined in claim 20, wherein said mode control means performs said evaluation calculation to said decision result through the employment of one of a two-dimensional template, a color analysis result and a velocity and acceleration evaluation.

22. A system as defined in claim 16, further comprising a human interface for information exchange with a user, said mode control means gives set information through said human interface to said user so that an incidental situation corresponding to a given mode setting state takes place.

95. A system as defined in claim 6, wherein said person information storage means is made such that a person picture of a specific person is registered as individual information when needed and is read out at a proper timing.

96. A system as defined in claim 95, wherein said mode control means uses person picture data being said individual information stored in said person information storage means, in said model-based mode.

23. A picture signal encoding system which compresses a picture signal with a variable-length code to produce and transmit encoded bit stream data, said system comprising:

object area extraction means for extracting an area of an object existing in a scene; and

encoding means for adding attribute data of the object area extracted by said object area extraction means to said picture signal to reduce information quantity in encoding processing.

24. A system as defined in claim 23, wherein said object area extraction means extracts said object area in accordance with the presence or absence of the occurrence of a motion vector.

25. A system as defined in claim 24, wherein said object area extraction means extracts said object area by detecting an occurrence area of said motion vector over three or more frames.

26. A system as defined in claim 24, wherein, for extracting said object area, said object area extraction means detects a block in which a maximum motion vector occurs and estimates as a main area an area including the maximum motion vector occurrence block.

27. A system as defined in claim 26, wherein said object area extraction means is equipped with panning vector calculation means for calculating a panning vector indicative of a motion quantity of the whole scene on the basis of an average value of motion vectors at a peripheral portion of said scene, and the detection of said maximum motion vector is made by subtracting said panning vector given through said panning vector calculation means.

28. A system as defined in claim 24, wherein for extracting said object area said object area extraction means decides a block pattern comprising motion blocks to calculate its feature quantity and scales a corresponding template when template information is present.

29. A system as defined in claim 28, wherein said object area extraction means performs a model-based transmission of data representative of attribute information in block unit as said template information in correspondence with a model.

30. A system as defined in claim 29, wherein for extracting said object area said object area extraction means analyzes a pattern taken when said object in said scene moves two-dimensionally to simply detect the motion.

31. A system as defined in claim 23, wherein for extracting said object area said object area extraction means produces a two-dimensional template on the basis of a three-dimensional model.

32. A system as defined in claim 23, wherein said object area extraction means decides a category of said object area on the basis of a control state based on mode information.

33. A system as defined in claim 32, wherein said object area extraction means decides said category of said object area on the basis of a state transition condition in mode control.

34. A system as defined in claim 32, wherein said object area extraction means decides said category of said object area on the basis of attribute information set in units of a block.

35. A system as defined in claim 23, wherein said object area extraction means always performs the area extraction processing at the time of transmission of a forced INTRA frame and designates and transmit information on a background area estimated from said object area at that time.

36. A system as defined in claim 35, wherein said object area extraction means designates said background area information to selectively perform transmission of said background area in accordance with an information transmission quantity.

37. A system as defined in claim 36, wherein, when said information transmission quantity undergoes a restriction, said object area extraction means transmits data whereby information designated as said background area information is combined with person information being under transmission.

38. A picture signal encoding system comprising,

picture compression and recognition processing means for compressing an input picture signal and for performing recognition processing;

a variable-length coding means for coding an output picture signal from said picture compression and recognition processing means;

a fixed-length coding means for coding said output picture signal from said picture compression and recognition processing means;

combining means responsive to output signals from said variable-length coding means and said fixed-length coding means for combining said output signals on the basis of syntacs;

a buffer responsive to an output signal from said combining means;

coding control means responsive to the capacity of said buffer and to processing attribute obtained by said picture compression and recognition processing means for controlling the operation of said variable-length coding means, said fixed-length coding means and said combining means.

39. A system as defined in claim 38, wherein said encoding control means encodes said picture signal by reducing the number of codes indicative of a header and a number for a block group comprising one frame to half of the number of block groups.

40. A system as defined in claim 38, wherein said encoding control means sets a length of a variable-length code indicative of an encoding attribute of a macroblock designated in units of a block in the order from a higher occurrence probability in relation to a person picture to control the occurrence code quantity.

41. A system as defined in claim 38, wherein said encoding control means is equipped with adaptive control means for executing adaptive control of a decision characteristic taken for a decision of an attribute set in units of a block.

42. A system as defined in claim 41, wherein said adaptive control means executes said adaptive control on the basis of area information in a frame.

43. A system as defined in claim 41, wherein said adaptive control means executes said adaptive control in accordance with a buffer quantity to make a decision on whether to choose an INTER frame or an INTRA frame in encoding said picture signal.

44. A system as defined in claim 41, wherein said adaptive control means executes said adaptive control in accordance with a buffer quantity to make a decision on whether or not to make a motion compensation in encoding said picture signal.

45. A system as defined in claim 40, wherein said encoding control means predicts an attribute of a frame to be encoded, on the basis of encoding attribute data obtained in terms of the past frame to control the code occurrence quantity.

46. A system as defined in claim 45, wherein said encoding control means performs said attribute prediction in units of a block on the basis of each of the past encoding attribute data in a macroblock layer.

47. A system as defined in claim 45, wherein said encoding control means performs said attribute prediction in units of a block on the basis of area information.

48. A system as defined in claim 40, wherein said encoding control means is provided with quantization control means for controlling a quantization step in encoding said picture signal.

49. A system as defined in claim 48, wherein said quantization control means assigns said quantization step at every area being one of an object area and a scene center.

50. A system as defined in claim 48, wherein said quantization control means assigns said quantization step in accordance with a buffer quantity.

51. A system as defined in claim 48, wherein said quantization control means assigns said quantization step in accordance with an encoding rate of said picture signal.

52. A system as defined in claim 48, wherein said quantization control means assigns said quantization step in accordance with the information occurrence quantity in encoding said picture signal.

53. A system as defined in claim 52, wherein said quantization control means decides said information occurrence quantity on the basis of the result of a calculation of a degree of an intensity of motion of the current frame with respect to that of the past frame.

54. A system as defined in claim 52, wherein said quantization control means decides said information occurrence quantity on the basis of the result of calculation of the degree of intensity of a color variation of the current frame with respect to that of the past frame.

55. A system as defined in claim 51, wherein said quantization control means makes a decision of the information occurrence quantity in accordance with the result of a calculation of an increment of a virtual buffer quantity made on the basis of one of the degree of intensity of motion and the degree of intensity of a color variation which are respectively said information occurrence quantity.

56. A system as defined in claim 40, wherein said encoding control means is provided with limiting means for controlling the number of significant coefficients of transformation coefficients after quantization in units of a block so that it is below an upper limit value determined in accordance with a given condition.

57. A system as defined in claim 40, wherein said encoding control means is provided with switching means for switching a frame rate in accordance with the information occurrence quantity.

58. A system as defined in claim 38, wherein said encoding control means executes said encoding control based on a picture encoding under one of the H.261 standard and the MPEG standard.

59. A system as defined in claim 58, wherein said encoding control means is equipped with data transform means made according to the H.261 standard.

60. A picture signal recognition system comprising,

object region extracting means responsive to an input picture signal for generating region information;

region evaluating means responsive to said region information for obtaining an evaluation result;

mode control means for providing mode setting information;

3-dimensional shape database including 3-dimensional models;

first memory means for storing person region picture data;

second memory means for storing scenery region picture data;

model base predicting means responsive to said evaluation result, said person region picture data, said scenery region picture data, said mode setting information, and said 3-dimensional models for generating a predicted picture; and

inter-frame coding means responsive to said input picture signal and said predicted picture for effecting inter-frame coding to produce coding data, said inter-frame coding means also generating and sending coded reproduction pictures to said model base prediction means,

thereby said model base predicting means outputs recognition results by obtaining a position posture by way of incidental situation which is obtained from said mode control means and said 3dimensional models, on the basis of color information accompanying a model or color information obtained by said coded reproduction picture, and said model base predicting means generates said prediction picture of a succeeding frame, said model base prediction means supplying said object region extraction means with a presumption region so that said object region extracting means outputs said region information on the basis of said presumption region and said input picture signal, said 3-dimensional shape data base supplies said model base prediction means with a suitable 3-dimensional model, person region pictures and scenery region pictures obtained in said model base prediction means during the operation thereof are able to be newly stored in said first and second memories respectively.

61. A system as defined in claim 60, wherein in said model-based mode said mode control means calculates distance information on the basis of a two-dimensional template corresponding to a scene set in advance for producing a model-based prediction picture of an object.

62. A system as defined in claim 61, wherein in said model-based mode said mode control means, for performing a model-based prediction, previously sets a feature area of said two-dimensional template and roughly estimates a positional attitude of said three-dimensional model from position data on motion of said feature area on the basis of two-dimensional motion information.

63. A system as defined in claim 60, wherein, when information necessary for said decision according to said picture signal analysis lacks, said mode control means determines said transition from said set mode state to said different mode state on the basis of a transition probability value set in advance according to its level.

64. A system as defined in claim 63, wherein, when an evaluation calculation is made with respect to the mode transition decision based on said transition probability value and shows that an error value goes large, said mode control means updates said transition probability value so that said error value becomes smaller.

65. A system as defined in claim 64, wherein said mode control means performs said evaluation calculation to said decision result through the employment of one of a two-dimensional template, a color analysis result and a velocity and acceleration evaluation.

66. A system as defined in claim 60, further comprising a human interface for information exchange with a user, said mode control means gives set information through said human interface to said user so that an incidental situation corresponding to a given mode setting state takes place.

67. A picture signal recognition system which analyzes a picture signal in the process of encoding said picture signal to recognize an object existing in an area of a frame, said system comprising:

object area extraction means for extracting an area of an object existing in a scene; and

encoding means for setting attribute data of the object area extracted by said object area extraction means to reduce information quantity in encoding processing.

68. A system as defined in claim 67, wherein said object area extraction means extracts said object area in accordance with the presence or absence of the occurrence of a motion vector.

69. A system as defined in claim 68, wherein said object area extraction means extracts said object area by detecting an occurrence area of said motion vector over three or more frames.

70. A system as defined in claim 68, wherein for extracting said object area said object area extraction means detects a block in which a maximum motion vector occurs and estimates as a main area an area including the maximum motion vector occurrence block.

71. A system as defined in claim 70, wherein said object area extraction means is equipped with panning vector calculation means for calculating a panning vector indicative of a motion quantity of the whole scene on the basis of an average value of motion vectors at a peripheral portion of said scene, and the detection of said maximum motion vector is made by subtracting said panning vector given through said panning vector calculation means.

72. A system as defined in claim 67, wherein for extracting said object area said object area extraction means decides a block pattern comprising motion blocks to calculate its feature quantity and scales a corresponding template when template information is present.

73. A system as defined in claim 72, wherein said object area extraction means performs a model-based transmission of data representative of attribute information in block unit as said template information.

74. A system as defined in claim 73, wherein for extracting said object area said object area extraction means analyzes a pattern taken when said object in said scene moves two-dimensionally for simply detecting a motion of the object.

75. A system as defined in claim 74, wherein for extracting said object area said object area extraction means produces a two-dimensional template on the basis of a three-dimensional model.

76. A system as defined in claim 67, wherein said object area extraction means decides a category of said object area on the basis of a control state according to mode information.

77. A system as defined in claim 76, wherein said object area extraction means decides said category of said object area on the basis of a state transition condition in mode control.

78. A system as defined in claim 76, wherein said object area extraction means decides said category of said object area on the basis of attribute information set in units of a block.

79. A system as defined in claim 67, wherein said object area extraction means always performs the area extraction processing at the time of transmission of a forced INTRA frame and designates and transmit information on a background area estimated from said object area at that time.

80. A system as defined in claim 79, wherein said object area extraction means designates said background area information to selectively perform transmission of said background area in accordance with an information transmission quantity.

81. A system as defined in claim 80, wherein, when said information transmission quantity undergoes a restriction, said object area extraction means transmits data whereby information designated as said background area information is combined with person information being under transmission.

82. A picture signal decoding system which receives a picture signal transmitted in the form of an encoded bit stream compressed with a variable-length code to decode it to reproduce said picture signal, said system comprising:

NFX transform means for detecting macroblock address (MBA) data indicative of a position of a block included in said picture signal formed in a state that pixel data corresponding to one frame is divided into a plurality of blocks and for transform between fixed (FXD) data and non-fixed (NFX) data on the basis of the detected MBA data;

prediction means for predicting an NFX pattern of a frame being currently under decoding, on the basis of a transform result of the past frame by said NFX transform means;

calculation means for calculating the degree of similarity between the NFX data pattern predicted by said prediction means and a decoded result and for calculating the degree of reliability of the calculated similarity;

evaluation means for evaluating said decoded result on the basis of the calculation result by said calculation means; and

correction means for correcting an error of the decoded result in accordance with the evaluation result by said evaluation means.

83. A picture signal decoding system which receives a picture signal transmitted as an encoded bit stream compressed with a variable-length code to decode it to reproduce said picture signal, said system comprising:

MTP detection means for detecting macroblock type (MTP) data indicative of an attribute of a block, included in said picture signal made in a state that pixel data corresponding to one frame is divided into a plurality of blocks, in the form of a code, and for producing an MTP pattern on the basis of a value of the detected MTP data;

comparative area setting means for setting a comparative area by making a motion compensation of a portion corresponding to said comparative area on the basis of the MTP pattern of the past frame detected by said MTP detection means;

calculation means for calculating the degree of similarity between said MTP data of the past frame in said comparative area set by said comparative area setting means and the MTP data being currently under decoding as a score representative of the degree of coincidence among a plurality of attribute data indicated by code values, and further for calculating the degree of reliability of the calculated similarity;

evaluation means for evaluating a decoded result on the basis of the calculation result by said calculation means; and

error correction means for correction an error of said decoded result in accordance with the evaluation result by said evaluation means.

84. A picture signal decoding system which receives a picture signal transmitted as an encoded bit stream compressed with a variable-length code to decode it to reproduce said picture signal, said system comprising:

MV transform means for detecting difference motion vector data (MVD) indicative of a difference motion vector of a block included in said picture signal made in a state that pixel data corresponding to one frame is divided into a plurality of blocks to transform the detected MVD data into motion vector (MV) data representative of a motion vector;

MV data estimation means for obtaining estimated MV data by estimating a motion vector, being currently under decoding, from one of the MV data of a past frame transformed by said MV transform means and MV data of a peripheral block; and

error correction means for comparing the estimated MV data obtained by said MV data estimation means with said MV data being currently under decoding to perform an evaluation and error correction of a decoded result on the basis of a difference obtained as the comparison result therebetween.

85. A picture signal decoding system which receives a picture signal transmitted as an encoded bit stream compressed with a variable-length code to decode it to reproduce said picture signal, said system comprising:

YUV vector transform means for detecting an encoded block pattern (CBP) data indicative of color information of a block, included in said picture signal made in a state that pixel data corresponding to one frame is divided into a plurality of blocks, in the form of a code and further for transforming the detected CBP data into a YUV vector in Y, U and V directions which allows a comparison in value;

comparative area setting means for setting a comparative area by making a motion compensation of a portion corresponding to said comparative area on the basis of a motion vector with respect to said YUV vector of the past frame transformed by said YUV vector transform means;

YUV vector estimation means for estimating a YUV vector of a block being under decoding on the basis of one of the YUV vector of a past frame transformed by said YUV vector transform means and a YUV vector value of a peripheral block within the comparative area set by said comparative area setting means; and

error correction means for comparing the estimated YUV vector obtained by said YUV vector estimation means with a YUV vector being currently under decoding to calculate an error evaluation value, and further for performing an error correction of CBP data of a block decoded on the basis of the error evaluation value.

86. A picture signal decoding system which receives a picture signal transmitted as an encoded bit stream compressed with a variable-length code to decode it to reproduce said picture signal, said system comprising:

mode control means storing a mode condition setting an incidental situation in the form of a predetermined code, when mode information is not included in the received picture signal, said mode control means selectively setting said mode condition corresponding to said mode information in performing a decoding processing; and

correction means for detecting and correcting an error of decoded data obtained through said decoding processing of said picture signal under the mode condition selectively set by said mode control means.

87. A system as defined in claim 86, wherein said mode control means includes person information storage means for storing and reading out data corresponding to a person in a scene and background information storage means for storing and reading out background information being in a fixed area other than an object area in said scene, said mode control means executing mode control corresponding to a use environment mode setting condition through the use of said person information storage means and said background information storage means used according to a use environment.

88. A system as defined in claim 87, wherein said mode control means is constructed to set a fixed mode and a movable mode for each of an interior of a motor vehicle and an outdoor place as said use environment, in said fixed mode for the interior of said motor vehicle said mode control means making said correction means exert the error correction function under the condition using said person information storage means and said background information storage means, in said movable mode for the interior of said motor vehicle said mode control means carrying out an updating process for said background information storage means in addition to said function in the motor vehicle interior fixed mode, in said outdoor place fixed mode said mode control means making said error correction means exert the error correction function under the condition mainly using said background information storage means, and in said outdoor place movable mode said mode control means making said error correction means exert the error correction function under the condition that said person mode is established.

89. A system as defined in claim 86, wherein said mode control means is equipped with person information storage means which stores and reads out data corresponding to a person in a scene and background information storage means which stores and reads out background information being in a stationary area other than an object area in said scene, said mode control means executing mode control corresponding to a subject mode setting condition through the use of said person information storage means and said background information storage means needed according to a subject.

90. A system as defined in claim 89, wherein said mode control means is made to set a person mode and a landscape mode as said subject mode, in said person mode said mode control means executing a decoding processing on the basis of data stored in said person information storage means under a feature condition peculiar to a person set in advance, while in said landscape mode said mode control means executing the decoding processing on the basis of background information stored in said background information storage means.

91. A system as defined in claim 86, wherein said mode control means detects a code of an encoding mode, set in the process of an encoding of said picture signal, in the process of a decoding to carry out the decoding processing under a decoding condition corresponding to said encoding mode and further to make said error correction means implement the error correction.

92. A system as defined in claim 91, wherein said mode control means implements the decoding processing in accordance with one of a motion picture mode, a quasi-motion picture mode and a still picture mode set as said encoding mode, and error correction means exerts said error correction function in said motion picture mode and said quasi-motion picture mode while making a retransmission request of said picture signal when an error occurs in a setting state of said still picture mode.

93. A system as defined in claim 91, wherein, when as said encoding mode is set one of a template use mode in which information equivalent to person area information corresponds a shape prepared in advance and a model-based mode in which motion information is added to a three-dimensional model, said mode control means executes the decoding processing by setting one of said template use mode and said model-based mode in the process of decoding in said person mode.

94. A system as defined in claim 91, wherein said error correction means selectively transmits one of the following request signals (1) to (4) to the bit stream transmission side when an error occurs in the decoding process so that the decoding stops:

(1) a retransmission request signal of a motion picture beginning with a forced INTRA picture;

(2) a newly resuming request signal of a transmission of a motion picture starting with a forced INTRA picture;

(3) a transmission request signal of a quasi-motion picture relying on only a forced INTRA picture; and

(4) a transmission request signal of a motion parameter in a model-based encoding.

97. A picture signal decoding system which receives a picture signal transmitted in the form of an encoded bit stream compressed with a variable-length code to decode it to reproduce said picture signal, said system comprising:

color information estimation means for estimating color information of each of blocks of pixel data decoded in block units on the basis of one of past color information data of an object block and color information of a peripheral block;

error evaluation means for evaluating the color information of said object block on the basis of an error value between the color information data obtained by the decoding process of said object block and the color information data estimated by said color information estimation means; and

error correction means for replacing the color information data of said object block with the color information data estimated by said color information estimation means when an evaluation result by said error evaluation means shows that the color information involves an error.

98. A system as defined in claim 97, further comprising:

transformation coefficient decoding means for detecting and decoding orthogonal transformation coefficient data of each of blocks of said encoded bit stream which is described with a fixed-length code using one of a variable-length code and an ESC code of a specific bit pattern;

ESC position storage means for calculating the degree of similarity between said variable-length code and said ESC code in the process of the detection of said orthogonal transformation coefficient data by said transformation coefficient decoding means and for storing a position of said ESC code of said bit stream when a bit coincidence therebetween exceeds a predetermined value; and

reevaluation means for, when said error correction means detects the error of the color information, recognizing as fixed-length code data the data described at the position stored by said ESC position storage means to again perform the decoding processing prior to replacing the color information with the color information estimated by said color information estimating means.

99. A system as defined in claim 97, further comprising:

EOB detection means for retrieving an end-of-block (EOB) code described at an end of each of blocks of said encoded bit stream to estimate an error of one bit and again detect said EOB code when not detecting said EOB code within a given condition range in the object block and to specify it as EOB code when detecting it within said given condition range; and

returning means for performing the decoding processing for the next block on the basis of a position of the EOB code detected by said EOB detection means when said error correction means detects the error of the color information data so that the decoding processing stops.

100. A picture signal decoding system which receives a picture signal transmitted in the form of an encoded bit stream compressed with a variable-length code to decode it to reproduce said picture signal, said system comprising:

error correction means for, when color information of pixel data obtained in block units through a decoding processing involves an error, estimating color information from one of past color information and color information of a peripheral block for an error correction; and

error propagation prevention means for requesting from a transmission side periodical transmission of forced INTRA block data in terms of said picture signal in block units in a portion becoming an important area of a decoded picture in accordance with the frequency of the error correction by said error correction means.

101. A picture signal decoding system which receives a picture signal transmitted in the form of an encoded bit stream compressed with a variable-length code to decode it to reproduce said picture signal, said system comprising:

area specification means for specifying an object area on the basis of area information data described in a user data area of said encoded bit stream data; and

error correction means for correcting an error included in said picture signal in the process of a decoding processing of the object area specified by said area specification means, on the basis of said area information data.

102. A system as defined in claim 101, wherein said area specification means is equipped with a plurality of basic templates for specifying a shape of said object area, and selectively sets one of said plurality of basic templates on the basis of information on said basic template designated through said area information described in said encoded bit stream data from the transmission side and specifies a portion of said object area by deforming the set basic template so that it conforms to said object area.

103. A system as defined in claim 102, wherein said plurality of basic templates are hierarchically provided covering a range from a template wholly indicative of an object to a template representative of a detailed portion of said object and express the specified portion in the form of a two-dimensional coordinate in the basic templates.

104. A system as defined in claim 103, further comprising motion compensation means for, when said area specification means specifies an area including a block being currently under decoding, performing a motion compensation for said block on the basis of a motion vector of a block of an area already decoded.

105. A system as defined in claim 103, further comprising color information compensation means for, when said area specification means specifies an area including a block being currently under decoding, correcting color information of said block on the basis of color information including a luminance, a color and a texture in a block of an area already decoded.

106. A system as defined in claim 101, further comprising background information storage means for storing as background information picture information of a background area other than an object area specified by said area specification means, so that said error correction means performs the correction on the basis of said background information of said background information storage means when an error occurs during decoding of picture data in said background area.

107. A picture signal decoding system comprising,

first means responsive to a picture signal transmitted in the form of an encoded bit stream compressed with a variable-length code to decode it to reproduce said picture signal, said system comprising psuedo area specification means for specifying a psuedo object area by performing an evaluation calculation centering on the center of gravity of a picture on the basis of pattern information data of a past frame when an object is not specified from a decoded result of said picture signal;

second means for storing pattern information data;

third means for specifying pseudo area by performing an evaluation calculation with the center of gravity of a picture being used as a center on the basis of pattern information data of past frames read out from said second means when an object is not specified from z decoded result of said picture signal, said third means generating pattern information data to be stored in said second means, said pseudo area being provided to said first means so that said first means is able to efficiently perform decoding operation.

108. A picture signal decoding system comprising,

first means responsive to a picture signal transmitted in the form of an encoded bit stream compressed with a variable-length code to decode it to reproduce said picture signal, said first means having a buffer capable of storing bit stream data of at least one frame;

and

second means for retrieving a start code of a specific pattern, which start code is attached to be placed at a head position of said encoded bit stream data of one frame, thereby checking said bit stream data stored in said buffer globally.

109. A system as defined in claim 108, wherein for retrieving a group-of-blocks (GOB) start code of a specific bit pattern positioned at a head portion of a first block group included in said picture signal formed in a state that pixel data corresponding to one frame is divided into a plurality of block groups, said code check means globally checks the bit stream data accumulated in said buffer.

110. A system as defined in claim 108, wherein, when the bit error rate in the specific bit pattern is below a given value, said code check means specifies said bit pattern as said start code in globally checking the bit stream data.

111. A system as defined in claim 108, wherein for performing positional search said code check means detects data indicative of a necessary number of bits which is described at a position subsequent to one of said start code of said bit stream data and said GOB start code, and compares the detected data with the actual number of bits between the specified start code and the next start code to determine the start code.

112. A system as defined in claim 109, wherein said code check obtains 8-bit data representative of a necessary number of bits of a block group described at a position subsequent to said GOB start code as data in which a higher rank side and a lower rank side are interchanged to form an inversion pattern.