SYSTEM AND METHOD FOR IMPLEMENTING LOW- COMPLEXITY MULTI- VIEW VIDEO CODING

FIELD OF THE INVENTION

[0001] The present invention relates generally to video coding. More particularly, the present invention relates to the enabling of multi-view video coding with a minimized level of complexity in terms of decoding processing and buffer requirements.

BACKGROUND OF THE INVENTION

[0002] This section is intended to provide a background or context to the invention that is recited in the claims. The description herein may include concepts that could be pursued, but are not necessarily ones that have been previously conceived or pursued. Therefore, unless otherwise indicated herein, what is described in this section is not prior art to the description and claims in this application and is not admitted to be prior art by inclusion in this section.

[0003] In multi-view video coding, video sequences output from different cameras, each corresponding to different views of a scene, are encoded into one bitstream. After decoding, to display a certain view, the decoded pictures belonging to that view are reconstructed and displayed. It is also possible for more than one view to be reconstructed and displayed.

[0004] Multiview video coding possesses a wide variety of applications, including free- viewpoint video/television, three dimensional (3D) TV and surveillance applications. Currently, the Joint Video Team (JVT) of the International Organization for Standardization (ISO)/International Engineering Consortium (IEC) Motion Picture Expert Group (MPEG) and International Telecommunication Union (ITU)-T Video Coding Expert Group is working to develop a multiview video coding (MVC) standard, which is becoming an extension of the ITU-T H.264 standard, also known as ISO/IEC MPEG-4 Part- 10. These draft standards as referred to herein as MVC and

AVC, respectively. The latest draft of the MVC standard is described in JVT-T208, "Joint Multiview Video Model (JMVM) 1.0", 20 ^th JVT meeting, Klagenfurt, Austria, July 2006, can be found at ftp3.itu.ch/av-arch/jvt-site/2006_07_Klagenfurt/JVT- T208.zip, and is incorporated herein by reference in its entirety. [0005] In JMVM 1.0, for each group of pictures (GOP), pictures of any view are contiguous in decoding order. This is depicted in Figure 1, where the horizontal direction denotes time (with each time instant being represented by Tm) and the vertical direction denotes view (with each view being represented by Sn). Pictures of each view are grouped into GOPs, e.g. pictures Tl to T8 in Figures 1 for each view form a GOP. This decoding order arrangement is referred to as view-first coding. It should be noted that, for the pictures in one view and in one GOP, although their decoding order is continuous without any other pictures to be inserted between any two of the pictures, internally their decoding order may change. [0006] It is also possible to have a different decoding order than that discussed for first- view coding. For example, pictures can be arranged such that pictures of any temporal location are contiguous in decoding order. This arrangement is shown in Figure 2. This decoding order arrangement is referred to as time-first coding. It should also be noted that the decoding order of columns (TO, Tl, etc.) may not be identical to the temporal order.

[0007] A typical prediction structure (including both inter-picture prediction within each view and inter-view prediction) for multi-view video coding is shown in Figure 3, where predictions are indicated by arrows, and the pointed-to object uses the pointed-from object for prediction reference. For views that share the same sequence parameter set (SPS), JMVM 1.0 provides the dependencies among views in a MVC SPS extension.

[0008] According to JMVM 1.0, given a MVC bitstream, for any view to be displayed, the pictures of the view and all other views the view directly or indirectly relies on, must be folly decoded and reconstructed. In this situation, "View A directly depends on view B" means that at least one picture in view B is used by a picture in view A for inter- view prediction. If "View A indirectly depends on view C," this means that no picture in view C is used by any picture in view A for inter- view

prediction, but View A cannot be correctly decoded without View C. For example, if view A directly depends on view B and view B directly depends on view C, then view A indirectly depends on view C. These relationships result significant decoding processing capability requirements, which therefore results in a high decoder implementation complexity and power consumption.

[0009J In addition to the above, when the number of views is large, both for time- first and view-first coding, the buffer size required for storing pictures used for interview prediction or temporal prediction becomes quite large. For example, when a hierarchical B GOP structure (the coding structure used in the time dimension in Fig. 3) is used in both the time dimension and the view dimension, for view-first coding, the required buffer size is equal to number_o reviews + GOP_Jength*(l +log2(number_of_views))+log2(GOP_length) [00101 In the above equation, "GOPJength" is the length of the GOP in number of pictures. When "GOP_length" is equal to 16 and "number_of_views" is equal to 17, the required buffer size is 101, in units of decoded frames.

[0011] As factors such as complexity, power consumption, and buffer size increase, an end result will ultimately involve a higher cost for devices capable of supporting multi-view decoding. These costs will become especially prohibitive for mobile devices, where space constraints inevitably result in still higher component costs. It would therefore be desirable to provide an arrangement where these complexities can be reduced efficiently.

SUMMARY OF THE INVENTION

[0012] The present invention provides a system and method for applying single- loop decoding to multi- view video coding. According to various embodiments of the present invention, for N coded views, where only random M of the N views are to be displayed, then only those M views arc required to be fully decoded and stored to a decoded picture buffer (DPB) when needed. Pictures of other views are only partially decoded or simply parsed and do not have to be stored into the DPB. The various embodiments also provide for an encoder that encodes multi-view video bitstreams in

-_>-

accordance with the single-loop decoding concept, as well as a decoder that utilizes single-loop decoding to decode and output on a subset of the encoded views from a multi-view bitstream. This system greatly reduces the complexity and buffer size requirements of a device, thereby reducing the cost of the device. [0013] These and other advantages and features of the invention, together with the organization and manner of operation thereof, will become apparent from the following detailed description when taken in conjunction with the accompanying drawings, wherein like elements have like numerals throughout the several drawings described below.

BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 is an arrangement of pictures in a first-view coding arrangement;

[0015] Figure 2 is an arrangement of pictures in a time-first coding arrangement;

[0016] Figure 3 is a depiction of an example MVC temporal and inter-view prediction structure;

[0017] Figure 4 is an overview diagram of a system within which the present invention may be implemented;

[0018] Figure 5 is a perspective view of a mobile device that can be used in the implementation of the present invention; and

[0019] Figure 6 is a schematic representation of the circuitry of the mobile device of

Figure 5.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0020] Figure 4 shows a generic multimedia communications system for use with the present invention. As shown in Figure 4, a data source 100 provides a source signal in an analog, uncompressed digital, or compressed digital format, or any combination of these formats. An encoder 110 encodes the source signal into a coded media bitstream. The encoder 110 may be capable of encoding more than one media type, such as audio and video, or more than one encoder 110 may be required to code different media types of the source signal. The encoder 110 may also get

synthetically produced input, such as graphics and text, or it may be capable of producing coded bitstreams of synthetic media. In the following, only processing of one coded media bitstream of one media type is considered to simplify the description. It should be noted, however, that typically real-time broadcast services comprise several streams (typically at least one audio, video and text sub-titling stream). It should also be noted that the system may include many encoders, but in the following only one encoder 110 is considered to simplify the description without a lack of generality.

[0021] The coded media bitstream is transferred to a storage 120. The storage 120 may comprise any type of mass memory to store the coded media bitstream. The format of the coded media bitstream in the storage 120 may be an elementary self- contained bitstream format, or one or more coded media bitstreams may be encapsulated into a container file. Some systems operate "live", i.e. omit storage and transfer coded media bitstream from the encoder 110 directly to the sender 130, The coded media bitstream is then transferred to the sender 130, also referred to as the server, on a need basis. The format used in the transmission may be an elementary self-contained bitstream format, a packet stream format, or one or more coded media bitstreams may be encapsulated into a container file. The encoder 110, the storage 120, and the sender 130 may reside in the same physical device or they may be included in separate devices. The encoder 110 and sender 130 may operate with live real-time content, in which case the coded media bitstream is typically not stored permanently, but rather buffered for small periods of time in the content encoder 110 and/or in the sender 130 to smooth out variations in processing delay, transfer delay, and coded media bitrate.

[0022] The sender 130 sends the coded media bitstream using a communication protocol stack. The stack may include but is not limited to Real-Time Transport Protocol (RTP), User Datagram Protocol (UDP), and Internet Protocol (IP). When the communication protocol stack is packet-oriented, the sender 130 encapsulates the coded media bitstream into packets. For example, when RTP is used, the sender 130 encapsulates the coded media bitstream into RTP packets according to an RTF payload format. Typically, each media type has a dedicated RTP payload format. It

should be again noted that a system may contain more than one sender 130, but for the sake of simplicity, the following description only considers one sender 130. [0023] The sender 130 may or may not be connected to a gateway 140 through a communication network. The gateway 140 may perform different types of functions, such as translation of a packet stream according to one communication protocol stack to another communication protocol stack, merging and forking of data streams, and manipulation of data stream according to the downlink and/or receiver capabilities, such as controlling the bit rate of the forwarded stream according to prevailing downlink network conditions. Examples of gateways 140 include multipoint conference control units (MCUs), gateways between circuit-switched and packet- switched video telephony, Push-to-talk over Cellular (PoC) servers, IP encapsulators in digital video broadcasting-handheld (DVB-H) systems, or set-top boxes that forward broadcast transmissions locally to home wireless networks. When RTP is used, the gateway 140 is called an RTP mixer and acts as an endpoint of an RTP connection.

[0024] The system includes one or more receivers 150, typically capable of receiving, de-modulating, and de-capsulating the transmitted signal into a coded media bitstream. The coded media bitstream is typically processed further by a decoder 160, whose output is one or more uncompressed media streams. It should be noted that the bitstream to be decoded can be received from a remote device located within virtually any type of network. Additionally, the bitstream can be received from local hardware or software. Finally, a Tenderer 170 may reproduce the uncompressed media streams with a loudspeaker or a display, for example. The receiver 150, decoder 160, and renderer 170 may reside in the same physical device or they may be included in separate devices.

[0025] Scalability in terms of bitrate, decoding complexity, and picture size is a desirable property for heterogeneous and error prone environments. This property is desirable in order to counter limitations such as constraints on bit rate, display resolution, network throughput, and computational power in a receiving device. [0026} It should be understood that, although text and examples contained herein may specifically describe an encoding process, one skilled in the art would readily

understand that the same concepts and principles also apply to the corresponding decoding process and vice versa. It should be noted that the bitstream to be decoded can be received from a remote device located within virtually any type of network. Additionally, the bitstream can be received from local hardware or software. [0027] Communication devices of the present invention may communicate using various transmission technologies including, but not limited to, Code Division Multiple Access (CDMA), Global System for Mobile Communications (GSM), Universal Mobile Telecommunications System (UMTS), Time Division Multiple Access (TDMA), Frequency Division Multiple Access (FDMA), Transmission Control Protocol/Internet Protocol (TCP/IP), Short Messaging Service (SMS), Multimedia Messaging Service (MMS), e-mail, Instant Messaging Service (IMS), Bluetooth, IEEE 802.11, etc. A communication device may communicate using various media including, but not limited to, radio, infrared, laser, cable connection, and the like.

[0028] Figures 5 and 6 show one representative mobile device 12 within which the present invention may be implemented. It should be understood, however, that the present invention is not intended to be limited to one particular type of mobile device 12 or other electronic device. Some or all of the features depicted in Figures 5 and 6 could be incorporated into any or all devices that may be utilized in the system shown in Figure 4.

[0029] The mobile device 12 of Figures 5 and 6 includes a housing 30, a display 32 in the form of a liquid crystal display, a keypad 34, a microphone 36, an ear-piece 38, a battery 40, an infrared port 42, an antenna 44, a smart card 46 in the form of a UICC according to one embodiment of the invention, a card reader 48, radio interface circuitry 52, codec circuitry 54, a controller 56 and a memory 58. Individual circuits and elements are all of a type well known in the art, for example in the Nokia range of mobile devices.

{0030J Various embodiments of the present invention provide a system and method for applying single-loop decoding to multi-view video coding. According to various embodiments of the present invention, for N coded views, where only random M of the N views are to be displayed, then only those M views are required to be fully

decoded and stored to a decoded picture buffer (DPB) when needed. Pictures of other views are only partially decoded or simply parsed and do not have to be stored into the DPB. The various embodiments also provide for an encoder that encodes multi- view video bitstreams in accordance with the single-loop decoding concept, as well as a decoder that utilizes single-loop decoding to decode and output on a subset of the encoded views from a multi-view bitstream.

[0031] Single-loop decoding is conventionally used in Scalable Video Coding and is enabled by using a constrained intra-texture prediction mode, hi scalable video coding, a video signal can be encoded into a base layer and one or more enhancement layers constructed in a pyramidal fashion. An enhancement layer enhances the temporal resolution (i.e., the frame rate), the spatial resolution, or simply the quality of the video content represented by another layer or part thereof. Each layer, together with all its dependent layers, is one representation of the video signal at a certain spatial resolution, temporal resolution and quality level, As discussed herein, a scalable layer together with all of its dependent layers is referred to as a "scalable layer representation". The portion of a scalable bitstream corresponding to a scalable layer representation can be extracted and decoded to produce a representation of the original signal at certain fidelity.

[0032] With a constrained intra-texture prediction mode, which is used for single- loop decoding in SVC, the inter-layer intra-texture prediction can be applied to macroblocks (MBs) for which the corresponding block of the base layer is located inside intra-MBs. At the same time, those intra-MBs in the base layer use constrained intra-prediction. In single-loop decoding, the decoder needs to perform motion compensation and full picture reconstruction only for the scalable layer desired for playback (referred to herein as the desired layer), thereby greatly reducing decoding complexity. All of the layers other than the desired layer do not need to be fully decoded because all or part of the data of the MBs not used for inter-layer prediction (whether for inter-layer intra-texture prediction, inter-layer motion prediction or inter- layer residual prediction) are not needed for reconstruction of the desired layer. A single decoding loop is needed for decoding of most pictures, while a second decoding loop is applied to reconstruct the base representations, which are needed for

prediction reference but not for output or display, and are reconstructed only for "key" pictures.

[0033] A first exemplary method of the present invention is based on JMVM 1.0 and is as follows. According to this method, a particular flag is signaled in the bitstream, e.g. in the MVC SPS extension. This flag indicates that, to display any M of the N coded views, only pictures of the M views need to be fully decoded. [0034] A pair of constraints are applied during encoding of the MVC bitstream. In one particular embodiment of the present invention, for any block in a coded picture, when inter-view prediction is used, the inter-view prediction signal only comes from intra-coded blocks for which constrained intra-prediction were used. If constrained intra-prediction is in use for an intra-coded block, then any data from inter-coded blocks is not used for intra-prediction when encoding the intra-coded block. Second, the decoded signal of the intra-coded blocks used for inter- view prediction is the signal before loop filtering is applied.

[0035] In decoding, only the pictures in the views to be displayed are fully decoded. These views are referred to as the target views. For other views depended upon by the target views, only those constrainedly intra-coded blocks are decoded for interview prediction. Other blocks are only needed to be parsed to the extent that it is known that they are not needed for inter- view prediction. The decoded data of the constrainedly intra-coded blocks of the views the target views rely on are stored in a temporal buffer, which is typically not very large. Alternatively, the picture in a target view and its inter-view reference pictures can be decoded synchronously, i.e. the macroblocks in address A in each one of the inter- view reference pictures is decoded, provided that they are constrainedly intra-coded, just before the macroblock in address A in the target view picture, provided that the motion vector used for the inter- view prediction of the macroblock in address A in the target view picture does not point to the right or bottom direction indicated by the macroblocks in address A in each of the inter- view reference pictures .

(0036J The fully decoded pictures are stored in the DPB when they are required either for output or prediction reference. The pictures of any view that is neither a target view nor a view depended on by any target view can be simply discarded. Due

to the fact that only the pictures of the target views needs to be stored in the DPB and managed by the DPB management process, a much smaller DPB size is needed when compared to the arrangement described in JMVM 1.0. Furthermore, due to the fact that motion compensation (including sub-pixel interpolation) and loop filtering etc. operations do not need to be performed for pictures in the views depended on by the target views, the decoding complexity is greatly simplified.

[0037] In one particular embodiment of the present invention, signal elements for different information can signaled. For example, for each value of M (in the range of 1 to N, inclusive), the maximum number of reference frames and the required DPB size when only M of the N coded views are randomly selected and decoded for playback. These signal elements may be signaled in the bitstream, e.g., in the MVC SPS extension or in a supplemental enhancement information (SEI) message. [0038] According to another embodiment, signal elements for each value of M (in the range of 1 to N, inclusive) are signaled to indicate the maximum decoding level, i.e. the maximum quantized resource consumption for decoding similar to the levels specified in Annex A of the H.264/ Advanced Video Coding (AVC) standard, required when any M of the N coded view are selected for decoding. These signal elements may also be signaled in the bitstream, e.g., in the MVC SPS extension or in a supplemental enhancement information (SEI) message. For this particular embodiment, it is assumed that level definitions can consider intra-blocks in depend- on views as reconstructed pixels, and therefore levels would constrain how deep and extensive inter- view prediction can be performed.

[0039] In accordance with still another embodiment, inter- view prediction of block coding mode, motion information or residual information is used to improve coding efficiency. This information is used in a manner similar to that specified in SVC for inter-layer prediction of block coding mode, motion information or residual information. These inter-view predictions may be further extended to involve motion compensation such that a prediction of block coding mode, motion information, or residual information may come from non-collocated blocks of the current block being encoded/decoded. This involving of motion compensation is similar to the inter- view

prediction of intra- sample values from constrainedly intra-coded blocks. These signal elements may also be signaled in the bitstream.

[0040] Another exemplary method for implementation is based on SVC and is as follows. In some cases of SVC, data in an enhancement layer can be truncated after a certain location, or at arbitrary positions, where each truncation position may include additional data representing increasingly enhanced visual quality. Such scalability is referred to as fine-grained (granularity) scalability (FGS). In contrast to FGS, the scalability provided by those enhancement layers that cannot be truncated is referred to as coarse-grained (granularity) scalability (CGS). It collectively includes the traditional quality (SNR) scalability and spatial scalability. In this particular method, the CGS quality scalable coding technique, i.e., CGS excluding spatial scalability, specified in SVC is applied for coding of multi-view video. In this situation, each CGS quality scalable layer corresponds to a view. Each value of dependency_id corresponds to the view_id of the view associated with the dependency_id value. The decoding process and other aspects are the same as what is specified in the SVC specification, except that more than one view, each coded as a CGS quality scalable layer, may be full decoded and outputted .

[0041] To allow for more views, the coded length of dependency_id can be extended, e.g. to 10 bits as view_id is specified in JMVM 1.0. In addition, a particular flag can be signaled, e.g. in the SVC SPS extension, to indicate whether the bitstream is a multi-view bitstream or SVC bitstream.

[0042] For inter-layer prediction (i.e. inter-view prediction in a multi-view context) of intra-sample values, block coding modes, motion information or residual information, motion compensation may be involved such that a prediction of any of this information may come from non-collocated blocks of the current block being encoded/decoded. This means that additional motion information needs to be signaled for inter-layer prediction (i.e. inter-view prediction in multi-view context). [0043J The present invention is described in the general context of method steps. which may be implemented in one embodiment by a program product including computer-executable instructions, such as program code, executed by computers in networked environments. Generally, program modules include routines, programs,

objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. Computer-executable instructions, associated data structures, and program modules represent examples of program code for executing steps of the methods disclosed herein. The particular sequence of such executable instructions or associated data structures represents examples of corresponding acts for implementing the functions described in such steps. [0044] Software and web implementations of the present invention could be accomplished with standard programming techniques with rule based logic and other logic to accomplish the various database searching steps, correlation steps, comparison steps and decision steps. It should also be noted that the words "component" and "module," as used herein and in the claims, is intended to encompass implementations using one or more lines of software code, and/or hardware implementations, and/or equipment for receiving manual inputs. [0045] The foregoing description of embodiments of the present invention have been presented for purposes of illustration and description. It is not intended to be exhaustive or to limit the present invention to the precise form disclosed, and modifications and variations are possible in light of the above teachings or may be acquired from practice of the present invention. The embodiments were chosen and described in order to explain the principles of the present invention and its practical application to enable one skilled in the art to utilize the present invention in various embodiments and with various modifications as are suited to the particular use contemplated.