The present invention relates to a transmission device, a reception device and a
communication system.
[Background Art]
[0002]
Recently, as disclosed, for example, in Patent Literature 1 below, a method is known
in which an image for the left eye and an image for the right eye having parallax
therebetween are alternately supplied to a display at a predetermined interval, and the
images are viewed using glasses with liquid crystal shutters that are driven in
synchronization with the predetermined interval.
[0003]
Further, it is disclosed in Patent Literature 2 below that modes such as error control
by FEC (Forward Error Correction) or error control on the basis of an automatic
repeat request (ARQ) are changed in accordance with packet loss and error
generating conditions in a network, and a packet transfer is performed.
[0004]
hi addition, as stereoscopic image encoding methods, a method is known, for
example, in which the image for the left eye and the image for the right eye are each
encoded while being considered to be individual video, and a method is known in
which encoding is performed as multiview images by MVC (Multiview Video
Codmg). In Patent Literature 1 below, the method to perform encoding by MVC
with respect to a plurality of video image data, such as stereoscopic video images
etc., is disclosed. Further, in Patent Literature 2 below, a method is disclosed in
which FEC strength (redundancy) is changed in accordance with a degree of priority.
[Citation List]
[Patent Literature]
[0005]
[PTL 1]
Japanese Patent AppUcation Publication No. JP-A-9-138384
[PTL 2]
Japanese Patent No. 3757857
[Non Patent Literature]
[0006]
[NPLl]
2

Kimata Hideaki, "Trends in International Standardization of Multiview Video Coding
MVC," Institute of Image Information and Television Engineers Journal, VOL. 61,
No. 4 (2007), p. 426 - 430.
[NPL2]
Alexander E. Mohr, Student Member, IEEE, Eve A. Riskin, Senior Member, IEEE,
and Richard E. Lander, Member, IEEE, "Unequal Loss Protection: Graceful
Degradation of Image Quality over Packet Erasure Channels Through Forward Error
Correction", "IEEE JOURNAL ON SELECTED AREAS IN
COMMUNICATIONS", VOL. 18, NO. 6, JUNE 2000, p. 819 - 828.
[Summary of Invention]
[Technical Problem]
[0007]
However, in a case in which a plurality of camera input video images are encoded
and broadcast by live relay or transmitted by network, a problem arises in that an
effect of the video images is lost due to delay, for example, a transmission delay from
one camera relative to another, and packet loss. For example, in a case in which
stereoscopic video images are transmitted as the plurality of camera input video
images, a stereoscopic effect is strong, and when, due to network delay jitter or
packet loss, part or some of the data of the video images are lost on one side only of
the left and right video images having a large parallax therebetween and the video
images are displayed in that state, a problem arises in that the stereoscopic effect is
lost.
[0008]
Further, when FEC processing is performed for packet loss recovery in order to avoid
this type of problem, in a case of live relay etc. by a broadcast station, a delay of less
than one frame of the video images is required. As a result, it is necessary to perform
FEC processing while reducing the FEC encoding matrix to perform the FEC
processing. However, when the matrix size is reduced, a delay size and a recovery
rate have a trade off relationship, with a reduction in the packet loss recovery rate
and so on. A system is therefore required in which a plurality of video images can be
transmitted with high packet loss resilience and also low delay.
[0009]
The present embodiment has been made in light of the foregoing circumstances, and
3

the present embodiment aims to provide a novel and improved transmission device,
reception device and communication system that are capable of transmitting a
plurality of video image data with high packet loss resilience and also low delay.
[Solution to Problem]
[0010]
Summary of the Invention
[0011]
In accordance with one embodiment, a video image data encoder comprises an input
configured to receive stereoscopic image data. The stereoscopic image data includes
first and second image data having chronological correspondence. An error
correction encoding unit is configured to combine portions of the first image data and
corresponding portions of the second image data from the input unit into a common
encoding matrix. An error correcting code derived from the combined portions is
added to the encoding matrix.
[Advantageous Effects of Invention]
[0012]
According to the present embodiment, it is possible to provide a transmission
device, a reception device and a communication system that are capable of
transmitting a plurality of video image data with high packet loss resilience and also
low delay.
[Brief Description of Drawings]
[0013]
[Fig. 1]
Fig. 1 is a schematic diagram showing a basic configuration of a communication
system according to an embodiment of the present embodiment.
[Fig. 2]
Fig. 2 is a schematic diagram illustrating a low delay FEC processing method
according to the present embodiment.
[Fig. 3]
Fig. 3 is a schematic diagram showing a method to generate an FEC encoding
matrix.
[Fig. 4]
4

Fig. 4 is a schematic diagram showing a method to generate a priority-based FEC control encoding matrix. [Fig. 5]
Fig. 5 is a block diagram showing an example of a configuration to perform priority-based FEC control and CODEC-linked FEC control. [Fig. 6]
Fig. 6 is a flow chart showing a processing procedure of a transmission device. [Fig. 7]
Fig. 7 is a flow chart showing a processing procedure of a reception device. [Fig. 8]
Fig. 8 is a flow chart showing processing corresponding to the priority-based FEC control and the CODEC-linked FEC control, and processing corresponding to dynamic QoS control. [Fig. 9]
Fig. 9 is a flow chart showing another processing procedure of the reception device. [Fig. 10]
Fig. 10 is a schematic diagram showing information of an FEC header provided in a parity packet.
[Description of Embodiments]
[0014]
Hereinafter, a preferred embodiment will be described in detail with reference to the
appended drawings throughout which like parts are referred to by like references.
Note that, in this specification and the appended drawings, structural elements that
have substantially the same function and structure are denoted with the same
reference numerals, and repeated explanation of these structural elements is omitted.
[0015]
Note that the description will be given in the order shown below.
(1) Overview of present embodiment
(2) Example of system configuration
(3) Priority-based FEC control
(4) CODEC-linked FEC control
(5) Processing procedure of transmission device
5

(6) Processing procedure of reception device
[0016]
(1) Overview of present embodiment
In the present embodiment, in transmission of stereoscopic video images, when a
plurality of camera input video images are encoded and transmitted, in order to
transmit with low delay and also with high resilience to packet loss, packets of the
plurality of camera video images that are picked up at the same time are processed
in units of a same FEC encoding matrix. In addition, when packet loss recovery is
not possible with the encoding matrix, by performing error concealment on a packet
that has correlation in that matrix unit, it becomes possible to maintain the
stereoscopic video effect. Note that, in the following explanation, an example is
described in which stereoscopic video images are formed from images for the right
eye and images for the left eye as input video images from a plurality of cameras, but
the present embodiment is not limited to this example. The present embodiment can
be widely applied to systems that perform transmission of a plurality of video images
having a chronological correspondence, such as a plurality of video images that are
simultaneously acquired from a several cameras, and in this case, it is possible to
transmit a plurality of video image data with a high degree of packet loss resilience
and also with low delay.
[0017]
In addition, the encoding of the present embodiment is line-based encoding in which
a plurality of scan lines of video image data are bundled and processed in line block
units, and processing is possible with a low delay of less than a frame or a field.
Further, by processing in block units, such as an FEC encoding block or matrix, an
impact of packet loss that cannot be recovered can also be minimized to deterioration
of a single part of an image fi-ame.
[0018]
Furthermore, a priority is attached in accordance with the stereoscopic video effect,
namely in accordance with a size of the parallax, and concealment processing is
performed, such as increasing FEC parity with respect to packets with a high impact
on image quality (e.g. the left and right images represented by the packets are quite
different), and reducing parity when there is a small impact (e.g. there is relatively
little difference between the left and right images). For that reason, video images can
6

be provided without losing the stereoscopic video efifect.
[0019]
It is assumed that the system according to the present embodiment has a
configuration that is explained below. The video image encoding is low delay
encoding processing in which processing is performed on a line basis, namely, in line
block units of a plurality of lines (one line or more). Thus, the encoding processing
is not performed after data are buffered in video frame units, and, as transmission and
display are possible with a delay of less than a frame time period, it is particularly
suitable for adoption in systems that perform real time video image fransmission,
such as live video distribution, and relay broadcasting. In the present system,
stereoscopic video image (3D) compliant camera output video images are encoded
and packetized for IP transmission, then FEC encoded in order to perform packet loss
recovery in the network and transmitted. On the reception side, FEC decoding and
packet loss recovery are performed and then video image decoding and display are
performed.
[0020]
(2) Example of system configuration
Fig. 1 is a schematic diagram showing a basic configuration of a cormnunication
system 100 according to the present embodiment. As shown in Fig. 1, the
communication system 100 is formed of a transmission device 200 that has a camera
video image input and a reception device 300 that has a display output. The
transmission device 200 and the reception device 300 are connected via a network
400, such as the Internet.
[0021]
As the stereoscopic video images, video images are used that are captured using a
dedicated camera 500. In Fig. 1, an example is shown in which the stereoscopic
video images are obtained from a camera (R) 502 that captures an image for the right
eye R and a camera (L) 504 that captures an image for the left eye L. In this way,
video images that are captured by at least two cameras can be used for the
stereoscopic video images, but the system can be applied with even one camera, as
long as the camera is compliant with stereoscopic video image capture. For example,
even with a single reflex type 3D camera, if it is a system that obtains a 3D effect by
forming left and right video images, it can be applied to the present system.
7

[0022]
As shown in Fig. 1, the transmission device 200 is formed of a 3D processing portion
202, a video image coding portion 204, a packet processing portion 206, an FEC
encoding portion 208 and a transmission portion 210. The 3D processing portion
202 brings about the stereoscopic video effect by performing, on the video images
output from the camera 500, information processing such as calibration of video
image data of the left and right cameras, and processing for 3D video images.
[0023]
Compression processing is performed in the video image coding portion 204 on the
video images on which 3D processing has been performed. After compression
processing, packetization processing for conmiunication (RTP (Real-time Transport
Protocol) packetization) is performed by the packet processing portion 206. Packets
output from the packet processing portion 206 are arranged in an FEC encoding
matrix (to be explained later) by the FEC encoding portion 208 and FEC processing
is performed. In this way, a parity packet that is formed of parity only, is generated.
Header processing is performed on this parity packet, and is transmitted as a UDP/IP
packet from the transmission portion 210 to the network 400.
[0024]
Further, the reception device 300 is formed of a reception portion 302, an FEC
decoding portion 304, a packet processing portion 306 and a video image decoding
portion 308. The reception portion 302 receives IP packets transmitted from the
transmission device 200 via the network 400. If the reception portion 302
recognizes, by the packet header, the data on which FEC processing has been
performed, an FEC matrix is formed by the FEC decoding portion 304, and, if there
is packet loss, the parity in the parity packet is used and FEC decoding is performed.
Lost packets can be recovered in this way. A code stream obtained from the packet
data is decoded by the video image decoding portion 308 and left and right video
image data are output to a 3D display device 500. In the packet loss recovery
processing, after decoding, a substitute section is selected from a preceding frame
and display is performed.
[0025]
Fig. 2 is a schematic diagram illustrating a low delay FEC processing method
according to the present embodiment. In the illusfration here, an example is used in
8

which the two cameras (the camera 502 and the camera 504) are aligned and 3D image capture is performed of left and right video images (an R image and an L image). As shown in Fig. 2, packets Rl, R2 to R7 and so on are generated from the R image, and packets LI, L2 to L7 and so on are generated from the L image. In the present embodiment, when FEC processing is performed in packet units, as one example, a matrix is formed by aligning four of the packets in a column-wise direction, as shown in Fig. 2. [0026]
Fig. 3 is a schematic diagram showing a method to generate a prerequisite FEC encoding matrix. Processing on the transmission side is performed in the FEC encoding portion 208 and processing on the reception side is performed in the FEC decoding portion 304. In the present embodiment, as shown in Fig. 3, a single FEC encoding matrix is formed by four packets. On the side of the transmission device 200, as the FEC (as the forward error correction), encoding is performed using Reed-Solomon codes for example, the encoding being performed in byte units in the column-wise direction. Parity data are generated in accordance with a parity number. The parity data are aligned in the transverse direction and parity packets (FEC parity packets shown in Fig. 3) are generated. FEC headers are attached to the parity packets and the packets are transmitted. Further, on the side of the reception device 300, if it is detected by the FEC decoding portion 304 that an error (LOSS) has occurred of a given packet, the parity packet is used to cause recovery of the data in which the error occurred. [0027]
In the present embodiment, in a low delay CODEC, left and right image data are encoded in units of line blocks that are formed of a plurality of lines, then packetized and transmitted. As described above, the single FEC encoding matrix is formed of the four packets. For that reason, when the stereoscopic video images are the usiml two sets of data of the image for the right eye (the R image) and the image for the left eye (the L image), if it is assumed that the FEC encoding matrix is generated using only the R image shown in Fig. 2, a time to generate four packets (tl +12) is required to generate the matrix. In the present embodiment, by using packets in which the L image and the R image in the same time period tl are each encoded, data of the four
packets (the packets Rl, R2, LI and L2) can be obtained in only the time tl shown in
9

Fig. 2, and it is possible to generate the FEC encoding matrix with a delay of the time tl. As a result, according to the method of the present embodiment, it is possible to reduce delay caused by encoding to a minimum. [0028]
As an encoding method, MVC (Multiview "V^deo Coding) is used, for example. In MVC, a structure incorporates not only temporal direction prediction, but also incorporates prediction between a pliirality of cameras. In other words, not only frame prediction in a temporal direction, but also prediction using a difference between the left and right images L and R are used. Then, using video image encoded data of reference image data that serves as a base and parallax information with respect to the reference data, encoding data to generate a predicted image are respectively packetized and the packets are processed using the same FEC encoding matrix. Note that the system disclosed in Patent Literature 1 described above can be used as the MVC encoding method. [0029]
(3) Priority-based FEC control (ULP: Unequal Loss Protection) Next, priority-based FEC control according to the present embodiment will be explained. Fig. 4 is a schematic diagram showing a method to generate a priority-based FEC control encoding matrix. Further, Fig. 5 is a block diagram showing an example of a configuration to perform priority-based FEC control and CODEC-linked FEC control, which will be explained later. [0030]
In Fig. 5, the 3D processing portion 202 acquires, from among 3D video image data transmitted from the camera 500, parallax information, color-difference information etc. relating to the 3D video image, and sorts the video image data into degrees of priority based on the obtained information. The priority information is then input into the FEC encoding portion 208. In the FEC encoding portion 208, the FEC strength, namely the parity number (redxmdancy), is updated based on the priority information. [0031]
For example, in order from a highest degree of priority, it is assumed that a priority A, a priority B, and a priority C are set, and that it is established in advance that two
parity packets are attached to priority A video image data, one parity packet is
10

attached to priority B video image data and no parity packet is attached to priority C
video image data. In this case, with respect to the FEC encoding matrix, as shown in
Fig. 4, two parity packets are attached to the priority A time period tl, no parity
packet is attached to the priority C time period t2, and one parity packet is attached to
the priority B time period t3.
[0032]
In a case in which a visual perception of the amount of pop out and the degree of
depth of the stereoscopic video images is large, namely, in a case in which the
parallax between the L image and the R image is great, an effect of the stereoscopic
video images is high, and thus the priority is set high. As a result, with respect to the
video image data that have a high eflFect as stereoscopic video images, a higher
number of parity packets are attached, and, if data loss occurs, the data can be
reliably recovered. It is thus possible to avoid loss of the 3D effect. On the other
hand, in a case in which the visual perception of the amount of pop out and the
degree of depth of the stereoscopic video images is small, a degree of importance as
3D video images is relatively low, a low priority is set, and the parity packets are
fewer. In a similar manner, the larger a color difference or the larger a difference in
luminance between the L image and the R image, the higher the priority is set. In the
way described above, with the priority-based FEC control according to the present
embodiment, the parity packets can be attached in accordance with priority
information based on the video image data, and when loss occurs in the data with a
high priority, recovery can be reliably performed on the reception side.
[0033]
(4) CODEC-linked FEC control
Next, CODEC-linked FEC control according to the present embodiment will be
explained. In the CODEC-linked FEC control, a total amount of data to which the
parity packet is attached is calculated by the FEC encoding portion 208, based on the
priority information. Normally, a constant bit rate or a maximum bit rate etc. that is
set when transmitting video image data is determined based on band restrictions and
a reserved bandwidth and the like. For that reason, as shown in Fig. 5, the FEC
encoding portion 208 feeds back information indicating a total amount of the data to
a CODEC (the video image coding portion 204) and adjusts, in conjunction with the
video image coding portion 204, the bit rate after encoding. In this way, the total
11

amount of data after the FEC processing can be controlled to be equal to or less than
a constant range.
[0034]
When explaining using the example shown in Fig. 4, two parity packets are attached
in the priority A time period tl, but if the bit rate exceeds a fixed value, an encoding
rate is adjusted by the video image coding portion 204 to be within a range in which
the bit rate does not exceed an upper limit. In the video image coding portion 204,
the encoding rate is adjusted by performing processing such as data thinning-out,
interpolation etc. In a similar maimer, with respect to the priority C time period t2
and the priority B time period t3, the encoding rate is adjusted within the range in
which the bit rate does not exceed the upper limit. In time periods in which the
priority is low, if the bit rate is in a range that does not exceed the upper limit, a
greater number of parity packets may be attached than the niraiber that otherwise
would be attached for that low priority. In the above-described manner, as well as
attaching the parity packets in accordance with the priority set by the priority-based
FEC control, it is possible to perform transmission with a data amount that is equal to
or less than a fixed bit rate.
[0035]
(5) Processing procedure of transmission device
Next, a processing procedure according to the transmission device 200 of the present
embodiment will be explained. Fig. 6 is a flow chart showing the processing
procedure of the transmission device 200. First, at step SIO, it is determined whether
or not the video image data has been input from the camera 500. If the video image
data has been input, the processing advances to step SI 2, and if the video image data
has not been input, the processing is ended. At step SI2, 3D processing is performed
by the 3D processing portion 202, and the parallax information and color-difference
information etc. between the left and right images, which are included in the video
image data, are acquired. In addition, in the 3D processing at step SI2, processing is
performed to acquire audio data and to acquire subtitle data and the like. Further,
when the 3D data is to be converted to side-by-side data, top-and-bottom data etc.,
conversion processing is performed.
[0036]
Next, at step SI4, processing is performed by the video image coding portion 204 to
12

encode the video image data. Next, at step SI6, processing is performed by the
packet processing portion 206 to packetize the video image data. Further, at step S12
performed by the 3D processing portion 202, information (data Dl shown in Fig. 6)
is acquired that indicates whether or not the left and right video image data are of a
same time, and this information is used when performing FEC encoding processing
at step SI8. This information is used to determine whether these data are associated
with the video image data at the time of 3D processing and whether these data are
input into a same matrix as correlating packets, at the time of the FEC encoding
processing (step SI8).
[0037]
Then, at step SI8, the FEC encoding processing is performed by the FEC encoding
portion 208, and next, at step S20, transmission processing is performed by the
transmission portion 210.
[0038]
Further, when the CODEC-linked FEC control is being performed, the encoding
processing at step S14 is performed such that a bit rate of the video image data that is
acquired in the FEC encoding processing at step S18 does not exceed a
predetermined threshold.
[0039]
(6) Processing procedure of reception device
Next, a processing procedure according to the reception device 300 of the present
embodiment will be explained. Fig. 7 is a flow chart showing the processing
procedure of the reception device 300. First, at step S30, it is determined whether or
not the reception portion 302 has received a packet. If a packet has been received at
step S30, the processing advances to step S32. On the other hand, if a packet has not
been received at step S 30, the processing is ended.
[0040]
At step S32, FEC decoding processing is performed by the FEC decoding portion
304. Next, at step S34, it is determined whether or not recovery of data lost in the
FEC decoding processing is complete. When the data recovery is complete, the
processing advances to step S36 and packet processing is performed by the packet
processing portion 306.
[0041]
13

On the other hand, when the data recovery at step S34 is not complete, the processing
advances to processing from step S38 onwards. In other words, when a number of
packet losses is in excess of a number that can be recovered by the FEC processing,
complete data packet recovery cannot be achieved by FEC, and the processing
advances to a flow from step S38 onwards to perform interpolation processing.
[0042]
At step S38, it is determined whether or not, with respect to the left and right video
image data, correlating packets, which have a high degree of correlation, exist. For
example, packets of left and right video image data that are picked up at the same
time exhibit relatively high correlation. When correlating packets do not exist, the
processing advances to step S40, and interpolation processing is performed on data
that was not recovered. In this way, when there are no correlating packets, for
example, when there is no 3D effect in the left and right video images, and there is
no difference between information of the left and right video images, at step S40,
interpolation processing is performed only on lost packets. In this case, at step S40,
error concealment processing is performed in which, for example, data of a
preceding fiame is extracted from stored data D and substituted. Note that the data
of the preceding frame (stored data D2) is temporarily stored in a buffer etc. In
addition, when only video image data for the right eye has not been recovered,
processing is performed to display the video images for the left eye only.
[0043]
On the other hand, when correlating packets exist at step S3 8, the processing
advances to step S42. At step S42, processing is performed to identify the
correlating packets. Here, for example, it is identified, from among a plurality of
fi-ames, in which frames the left and right video images have a high correlation.
[0044]
Fig. 10 is a schematic diagram showing information of an FEC header provided in
the parity packet. Information that indicates whether or not correlation between the
left and right video images is high is included in an FEC correlation information field
that is stored in an FEC packet header shown in Fig. 10. The FEC correlation
information field stores information such as correlation information, a correlation
level, a priority, sequence numbers, time information etc. This field is used to
identify correlating packets at the time of FEC decoding, using a method in which
14

sequence numbers indicate a type of correlation, a correlation level, a priority, and which packet correlates to which packet, for example. [0045]
After step S42, the processing advances to step S44. At step S44, collective interpolation processing is performed by the packet processing portion 306. More specifically, when there is correlation between the left and right images, the correlating packets are identified fi"om the FEC correlation information field stored in the FEC packet header shown in Fig. 10, and interpolation processing is performed based on packet data correlating to a lost packet. That is, interpolation of, for example, a lost packet of the left image is based upon the correlated packet of the right image, and vice versa. [0046]
As described above, when data recovery is not possible according to the processing shown in Fig. 7, it is possible to perform data interpolation processing based on whether or not correlating packets exist. It is therefore possible to appropriately perform interpolation processing in accordance with a degree of correlation between the left and right video images. [0047]
Next, another processing procedure according to the transmission device 200 will be explained. Fig. 8 is a flow chart showing processing corresponding to the above-described priority-based FEC control and the CODEC-linked FEC control, and processing corresponding to dynamic QoS control. As described above, in the priority-based FEC control, a deterioration in stereoscopic video effect caused by the impact of packet loss during transmission is minimized by changing the FEC strength depending on whether data has a large parallax and a strong stereoscopic video effect or whether data has a small parallax and a weak stereoscopic video effect. When the priority-based FEC control is performed, in the 3D processing at step S12, the priority of the video image data is sorted. Then, in the FEC encoding processing at step SI8, the FEC encoding processing is performed in accordance with the priority. [0048] Therefore, at step SI8 shown in Fig. 8, the FEC encoding portion 208 performs
processing to change the number of parity packets in accordance with the packet
15

priority.
[0049]
As described above, the CODEC-linked FEC control links fluctuations in a
bandwidth for transmission to the encoding rate control of the CODEC, by changing
the FEC strength depending on whether data has a large parallax and a strong
stereoscopic video effect or whether data has a small parallax and a weak
stereoscopic video effect. In this way, it is possible to perform transmission in which
a rate of transmission of packets after FEC encoding is within a range of a fixed bit
rate. Thus, at step SI8 shown in Fig. 8, the FEC encoding portion 208 feeds back the
total amount of data after the FEC encoding to the encoding processing performed at
step SI2. Then, at step S12 shown in Fig. 8, the video image coding portion 204
performs "thinning out" or interpolation processing in accordance with the data
amount and performs control such that the total data amoimt does not exceed the
upper limit.
[0050]
Further, in dynamic QoS control according to the present embodiment, the FEC
encoding portion 208 performs processing for packet loss recovery by combining
retransmission control, such as ARQ (Auto Repeat Request) control, with the FEC.
The ARQ control can use the method disclosed in Patent Literature 2 described
above. More specitlcally, as the packet loss recovery method, rather than only FHC,
packet losses can be more reliably recovered by hybrid QoS control that concurrently
uses retransmission control (ARQ). When error correction by the FEC is possible,
the FEC encoding portion 208 performs error correction by FEC. Further, when
error correction is not possible by FEC, the FEC encoding portion 208 performs
ARQ processing as a retransmission request. At this time, the FEC encoding portion
208 generates a retransmission request RTCP packet, and transmits the RTCP packet
to the transmission device 200. Thus, at step SI8 shown in Fig. 8, the FEC encoding
portion 208 performs processing for packet loss recovery that combines the
retransmission control and the FEC, and when the retransmission request is received
from the reception device 300, the packet is retransmitted. Note that, when
retransmission control is performed, the reception portion 302 of the reception device
300 fiinctions as a transmission portion, and the transmission portion 210 of the
transmission device 200 functions as a reception portion, and the retransmission
16

request is transmitted from the reception device 300 to the transmission device 200
via the network 400. At this time, it is also possible to set up a system in which the
QoS control method for packet loss recovery or strength is changed in accordance
with the bandwidth of the transmission path, parallax information, liominance
information and color-difiference information etc.
[0051]
Fig. 9 is a flow chart showing another processing procedure of the reception device
300. Fig. 9 shows processing to change the FEC strength using luminance
information and color-difference information, for example, to raise the FEC strength
while taking particular account of the luminance information. In this case, in the
interpolation processing at step S40, in addition to the above-described method to
substitute data of the preceding frame, when the color-difference information is lost
and carmot be recovered, processing is performed to interpolate luminance (Y) only.
[0052]
In addition, in the dynamic QoS control corresponding to the transmission device
200 as illustrated by Fig. 8, as a result of determination of FEC recovery at step S34,
when recovery is not possible by FEC, processing is performed to transmit, to the
transmission device 200, a retransmission request by ARQ.
[0053]
(7) Handling of subtitle information and audio information
Next, processing of subtitle information will be explained. A degree of priority of
the subtitle information may be set to be higher than that of the video images, and
can be handled as special information with respect to video image information. As
one example, redundancy with respect to the subtitle information can be made higher
than the video image information. Further, a method is conceivable that uses SVC
(Scalable Video Codec) layers and the degree of priority. This is a method in which
an FEC amount is changed in accordance with a degree of importance of the layer,
and a processing method is possible that shortens delay time by performing
processing on each layer of a plurality of channels using a single encoding matrix. As
an example, more important images, such as low frequency images, are considered to
be more important layers and have a higher degree of priority.
[0054]
In addition, audio information can be handled in a similar manner, in which data of a
17

plurality of channels are multiplexed on a single FEC encoding matrix and
processed, and packets relating to surroimd etc. can be collectively processed when
packet loss occurs.
[0055]
As described above, according to the present embodiment, it is possible to construct
a 3D video image live transmission system with low delay and also high resilience to
packet loss. Then, even if packet loss recovery cannot be performed, video images
that do not lose the stereoscopic video effect can be provided by interpolation
processing. Furthermore, by performing FEC processing in accordance with a
degree of strength of the stereoscopic video effect, video images can be provided
which do not lose the stereoscopic video effect. In addition, it is possible to curb the
transmission bandwidth to be equal to or less than a certain bandwidth, even while
enhancing packet loss resilience.
[0056]
The exemplary embodiment is described in detail above with reference to the
appended drawings. However, the present invention is not limited to the above-described examples. It shoiild be understood by those skilled in the art that various modifications, combinations, sub-combinations and alterations may occiir depending on design requirements and other factors insofar as they are within the scope of the appended claims or the equivalents thereof.
[Reference Signs List] [0057]
100 Commimication system 200 Transmission device 202 3D processing portion 208 FEC encoding portion 300 Reception device 302 Reception portion 304 FEC decoding portion
18

We Claim:
[Claim 1]
A video image data encoder comprising:
an input configured to receive stereoscopic image data, including first and second image data having chronological correspondence;
an error correction encoding unit configured to combine portions of said first image data and corresponding portions of said second image data from said input unit into a common encoding matrix and adding to said encoding matrix an error correcting code derived from said combined portions.
[Claim 2]
The image data encoder of claim 1, wherein said first and second image data represent images sensed by left and right eyes, respectively. [Claim 3]
The image data encoder of claim 1, wherein said chronological correspondence is substantially simultaneously produced left eye and right eye image data.
[Claim 4]
The image data encoder of claim 1, wherein said portions of said first and second image data are data packets.
[Claim 5]
The image data encoder of claim 4, wherein said error correction encoding unit combines plural data packets of said first image data and plural data packets of said second image data into a common encoding matrix; and said error correcting code comprises at least one packet of parity data derived from said data packets in said common encoding matrix.
[Claim 6]
The image data encoder of claim 5, wherein said error correcting code comprises at least one additional packet of parity data derived from said data packets in said common encoding matrix. [Claim 7]
The image data encoder of claim 5, wherein said data packets of said first and
second image data that are combined into said common encoding matrix represent
19

images picked up in a common time period by stereoscopic image pickup apparatus; and wherein said input unit assigns priority levels to the data packets of image data picked up in respective time periods.
[Claim 8]
The image data encoder of claim 7, wherein said error correcting code comprises a greater number of packets of parity data for common encoding matrices containing data packets of higher priority levels than for common encoding matrices containing data packets of lesser priority levels.
[Claim 9]
The image data encoder of claim 8, wherein the priority levels assigned to data packets in a respective time period is a function of parallax and/or color difference between said first and second image data. [Claim 10]
The image data encoder of claim 9, wherein said error correction encoding unit is configured to identify those data packets of said first image data that correlate with data packets of said second image data in a common encoding matrix. [Claim 11]
A video image encoder that encodes stereoscopic image data in an error correcting code, comprising:
an image acquisition unit configured to acquire video image data representing left and right images picked up in successive time periods;
a packetizing unit configured to form left image data packets of left image data and right image data packets of right image data and to assign priority levels to the data packets in said time periods; and
a forward error correcting (FEC) unit configured to form FEC encoding matrices of the left and right image data packets in said successive time periods, each FEC encoding matrix containing the left and right image data packets in a respective time period and a number of parity packets determined by the priority level assigned to the data packets in said respective time period. [Claim 12]
An image transmitter, comprising:
an image acquisition unit configured to acquire video image data representing
left and right images picked up in successive time periods;
20

a packetizing unit configured to form left image data packets of left image data and right image data packets of right image data and to assign priority levels to the data packets in said time periods;
a forward error correcting (FEC) unit configured to form FEC encoding matrices of the left and right image data packets in said successive time periods, each FEC encoding matrix containing the left and right image data packets in a respective time period and a number of parity packets determined by the priority level assigned to the data packets in said respective time period; and
a transmission unit configured to transmit the FEC encoding matrices.
[Claim 13]
Image receiving apparatus, comprising:
a receiving unit configured to receive error corrected encoded matrices of image data, each matrix including left and right image data packets, representing left and right images acquired in a common time period, and a number of parity packets;
an error-correction decoding unit configured to recover the left and right image data packets in a received matrix, using said parity packets if an error is present in the received matrix; and
a packet processing unit configured to produce interpolated data packets if said error-correction decoding unit is unable to recover said left and right image data packets.
[Claim 14]
The image receiving apparatus of claim 13, wherein said matrix includes correlation information identifying correlating left and right image data packets in said matrix; and said packet processing unit performs collective interpolation processing using the left image data packet if the correlated right image data packet is lost and the right image data packet is used if the correlated left image data packet is lost.