JP4861371B2 - Video quality estimation apparatus, method, and program - Google Patents

Description

  The present invention relates to a communication quality estimation technique, and more particularly to a technique for estimating quality related to video communication such as an IPTV service, a video distribution service, and a video communication service performed via an IP network such as the Internet.

Along with the increase in speed and bandwidth of access lines connected to the Internet, video communication services using audio and video are expected. Since the Internet is not necessarily a network where communication quality is guaranteed, when performing voice and video communications, if the bandwidth of the network connecting users is narrow or the line is congested, users will be able to respond to voice and video. Perceived by the user, that is, the quality of experience (QoE) of the user deteriorates. Specifically, when quality degradation is added to a video, it is perceived as blurring, blurring, mosaic distortion, jerky feeling, or the like.
Therefore, in order to provide such a video communication service with high quality, quality design prior to service provision and quality management after service start are important. For this purpose, the quality enjoyed by the user can be expressed appropriately. In addition, a simple and efficient video quality evaluation technique is required.

As the video quality evaluation technique, there has been proposed a video quality estimation technique for estimating a video degradation amount based on the total amount of invalid frames including a packet loss frame and a frame that refers to the frame (see Patent Document 1, etc.). ). However, since the above technique does not consider the dependency of the video content and estimates the video quality only from the frame structure, the video quality cannot be estimated appropriately.
In general, even if the number of invalid frames is the same, the greater the amount of motion of the video, the greater the quality degradation perceived by the user. Therefore, when the amount of motion of the video is not taken into account, for example, when the same number of frames of the video A and the video B are lost, the video A is less likely to be detected as quality degradation because the motion is small, but the video B has a large motion Therefore, the phenomenon that it is easily detected as quality degradation cannot be reflected in the video quality.

  Conventionally, with respect to such a video quality evaluation technique, a technique has been proposed in which a motion amount of a video to be subjected to quality estimation is calculated in advance, and video quality is estimated from the motion amount and an invalid frame ( (See Patent Document 2).

JP 2006-033722 A JP 2008-005108 A ITU-T Rec. P.910, "Subjective video quality assessment methods For multimedia applications".

However, in such a conventional technique, an index such as Temporal Index (for example, refer to Non-Patent Document 1 or the like), which is calculated by comparing pixel values of two adjacent frames as the motion amount of the video used for video quality estimation. Since the value is used, a huge processing load is required to calculate the index value. Therefore, only when the motion amount of the video is calculated in advance, the video quality can be estimated in consideration of the influence of the motion amount of the video. There has been a problem that the video quality considering the influence of the amount of motion cannot be estimated in real time.
An object of the present invention is to provide a video quality estimation apparatus capable of calculating a video quality value in consideration of a video motion amount with a small processing load.

  In order to achieve such an object, the video quality estimation apparatus according to the present invention compresses and encodes a target video into a plurality of frames based on an intra-frame and inter-frame predictive encoding scheme and transmits the target video as IP packets. A video quality estimation apparatus for estimating video quality of a video, analyzing an input IP packet, and obtaining various quality parameters related to video quality of target video communication and quantization parameters related to a quantization step width of compression coding. An IP packet analysis unit to be extracted, and an invalid frame number calculation unit that calculates the number of invalid frames including a frame having a packet loss and another frame that is affected by the packet loss of the frame at the time of decoding based on the quality parameter And the target video communication code considering only video degradation due to compression coding based on the quality parameter. An encoded video quality calculation unit that calculates a video quality value, an average quantization parameter calculation unit that calculates an average value of quantization parameters for each frame of the target video communication corresponding to the quality estimation interval, and quantization of the target video communication Based on the representative quantization parameter calculation unit that calculates the representative quantization parameter that represents the parameter, and the average quantization parameter and the representative quantization parameter, a packet loss change index that indicates the amount of motion of the video between the target frames is calculated. And a video quality calculation unit that calculates the video quality value of the target video communication based on the encoded video quality value, the number of invalid frames, and the packet loss change index.

  At this time, in the packet loss change index calculation unit, based on the function indicating the characteristic that the packet loss change index monotonously decreases in accordance with the increase of the difference quantization parameter indicating the difference between the average quantization parameter and the representative quantization parameter, A packet loss change index may be calculated.

  Further, the representative quantization parameter calculation unit calculates the quantization parameter corresponding to the coding rate included in the quality parameter based on a function indicating the characteristic that the quantization parameter monotonously decreases as the coding rate increases. It may be calculated as a parameter.

Further, the video quality calculation unit sets the encoded video quality value as the maximum video quality value when the number of invalid frames is 0, and is based on a function indicating a characteristic that the video quality value monotonously decreases as the number of invalid frames increases. Thus, the video quality value corresponding to the encoded video quality value, the number of invalid frames, and the packet loss change index may be calculated.

  In addition, the video quality estimation method according to the present invention is a video quality estimation method for video quality of target video communication in which a target video is compressed and encoded into a plurality of frames based on an intra-frame and inter-frame predictive coding scheme and transmitted in IP packets A video quality estimation method used in an estimation apparatus, in which an IP packet analysis unit analyzes an input IP packet and relates to various quality parameters related to video quality of target video communication and a quantization step width of compression coding The IP packet analyzing step for extracting the quantization parameter, and the invalid frame number calculation unit, based on the quality parameter, are composed of a frame having a packet loss and another frame that is affected by the packet loss of the frame at the time of decoding. An invalid frame number calculating step for calculating the number of invalid frames, and an encoded video quality calculating unit, An encoded video quality calculation step for calculating the encoded video quality value of the target video communication considering only video degradation due to compression encoding based on the quality parameter, and an average quantization parameter calculation unit correspond to the quality estimation section An average quantization parameter calculating step for calculating an average value of quantization parameters for each frame of the target video communication, and a representative quantization parameter calculation unit calculates a representative quantization parameter representative of the quantization parameter of the target video communication Packet loss for which a representative quantization parameter calculation step and a packet loss change index calculation unit calculate a packet loss change index indicating a motion amount of video between target frames based on the average quantization parameter and the representative quantization parameter The change index calculation step and the video quality calculation unit include an encoded video quality value, the number of invalid frames, And based on the packet loss variation index, and a video quality calculation step of calculating an image quality value of the target video communication.

  In addition, a program according to the present invention is a video quality estimation apparatus that estimates video quality of target video communication that is compressed and encoded into a plurality of frames based on an intra-frame and inter-frame predictive coding scheme and is transmitted as an IP packet. An IP packet analysis unit that analyzes an input IP packet and extracts various quality parameters related to video quality of target video communication and quantization parameters related to a quantization step width of compression coding in a computer The number of invalid frames for which the invalid frame number calculation unit calculates the number of invalid frames including a frame having a packet loss and another frame that is affected by the packet loss of the frame at the time of decoding based on the quality parameter. The calculation step and the encoded video quality calculation unit, based on the quality parameter, An encoded video quality calculation step for calculating the encoded video quality value of the target video communication considering only video degradation due to the reduced encoding, and an average quantization parameter calculation unit for each of the target video communication corresponding to the quality estimation section Average quantization parameter calculation step for calculating an average value of quantization parameters related to the frame, and representative quantization parameter calculation step for the representative quantization parameter calculation unit to calculate a representative quantization parameter representing the quantization parameter of the target video communication A packet loss change index calculating unit that calculates a packet loss change index indicating the amount of motion of video between the target frames based on the average quantization parameter and the representative quantization parameter; and The video quality calculation unit displays the encoded video quality value, the number of invalid frames, and the packet loss change index. Based on, to perform the image quality calculation step of calculating an image quality value of the target video communication.

  According to the present invention, the packet loss change index indicating the motion of the video is calculated from the quantization parameter described in each frame of the target video communication without calculating the pixel values of the two adjacent frames. Can do. Therefore, it is possible to calculate the video quality value considering the motion amount of the video with a small processing load, and even for the majority of the videos whose motion amount of the video is not known in advance, the video quality considering the effect of the motion amount of the video is estimated in real time. be able to.

Next, embodiments of the present invention will be described with reference to the drawings.
[First Embodiment]
First, the video quality estimation apparatus according to the first embodiment of the present invention will be described with reference to FIG. FIG. 1 is a block diagram showing the configuration of the video quality estimation apparatus according to the first embodiment of the present invention.

  The video quality estimation apparatus 1 is composed of an information processing apparatus such as a computer that performs arithmetic processing on input information, and compresses and encodes a video signal encoded into a plurality of frames based on intra-frame and inter-frame predictive encoding methods. Thereafter, the video communication transmitted by the IP packet from the communication terminal 2A to the communication terminal 2B via the IP network 3 such as the Internet was decoded and reproduced by the communication terminal 2B with the IP packet of the video communication as an input. This is an apparatus for calculating an estimated value of subjective video quality perceived by the viewer from the video as the individual video quality.

  The video quality estimation apparatus 1 includes an arithmetic processing unit 10, a storage unit 20, a communication interface unit (hereinafter referred to as a communication I / F unit) 31, an operation input unit 32, and a screen display unit 33 as main functional units. It has been. The arithmetic processing unit 10 includes, as main processing units, an IP packet analysis unit 11, an invalid frame number calculation unit 12, an encoded video quality calculation unit 13, a representative quantization parameter calculation unit 14, and an average quantization parameter calculation unit. 15, a packet loss change index calculation unit 16 and a video quality estimation unit 17 are provided.

[Principle of the Invention]
In general, in video communication in which video is compressed and encoded into a plurality of frames based on intra-frame and inter-frame predictive coding and transmitted as IP packets, when an IP packet is lost, the quality of the frame including the lost IP packet is degraded. At the same time, the quality degradation propagates to other frames encoded by the inter-frame predictive encoding method using the frame.
At this time, image quality degradation includes degradation that is easy for the user to perceive quality degradation and degradation that is difficult for the user to perceive, and the degree of influence of the video quality on the subjective evaluation by the user varies according to the amount of motion of the video. To do.

  FIG. 2 is an explanatory diagram showing the relationship between the amount of motion of video and the number of invalid frames-subjective evaluation value characteristics. As shown in FIG. 2, the relationship between the number of invalid frames IF and the subjective evaluation value V (MOS) is, as shown by the characteristic 51, that the deterioration in the subjective evaluation value V with respect to the increase in the number of invalid frames IF decreases as the amount of motion decreases. As shown in the characteristic 52, the deterioration of the subjective evaluation value V with respect to the increase in the number of invalid frames IF increases as the motion amount increases.

  When the amount of motion of the video is not taken into account in the video quality estimation, all the plots indicating the relationship between the number of invalid frames IF and the subjective evaluation value V are grasped in a unified manner as shown in the characteristic 53. Therefore, the estimated video quality The error between the value V and the actual subjective evaluation value V is large. Therefore, if the motion amount of the target video is calculated from the target video communication to be the quality estimation target and the invalid frame number-subjective evaluation value characteristic is specified, the video quality considering the motion amount of the video from the invalid frame number IF The value V can be estimated with high accuracy.

On the other hand, MPEG2 and H.264. In video communication using intra-frame and inter-frame predictive encoding schemes such as H.264, video information is compressed by encoding the video into frames such as I, B, and P frames.
At this time, for a video with a large amount of motion, instead of reducing the definition of the I frame, the definition of the B frame or the P frame transmitted between the I frames is increased to improve the quality of the entire video. The compression encoding process is controlled.

In addition, for a video with a small amount of motion, instead of increasing the definition of the I frame, the definition of the B frame or the P frame transmitted during the I frame is controlled to be low.
Therefore, as the amount of motion increases, the information amount of the B frame and the P frame increases instead of the decrease of the information amount of the I frame. In addition, the smaller the amount of motion, the lower the information amount of the B frame and the P frame, instead of the increase of the information amount of the I frame.

  In the present invention, paying attention to the relationship between the information amount of the I frame and the motion amount of the video, the packet loss change index D indicating the change of the packet loss corresponding to the motion amount of the video is derived from the information amount of the I frame. Then, based on this index and the number of invalid frames, the video quality is estimated in consideration of the video motion amount.

Also, in the present invention, focusing on the fact that the quantization parameter related to the quantization step width used in the compression encoding of the target video has a correlation with the information amount of each frame, the packet loss change index D is determined from this quantization parameter. Is derived.
At this time, since the quantization parameter varies in value depending on the type of video communication, the representative quantization parameter QPt related to the same type of video communication as the target video communication and the average quantization parameter calculated from the target video communication A differential quantization parameter ΔQP indicating a difference from QP is used as an intermediate parameter.

  FIG. 3 is an explanatory diagram showing packet loss change index calculation characteristics. The packet loss change index characteristic is a characteristic indicating the relationship between the differential quantization parameter ΔQP and the packet loss change index D, and can be represented by an approximate function such as a linear function as shown in FIG. In the present invention, a coefficient for specifying an approximate function is calculated in advance for each type of video communication, and the packet loss change index D is calculated based on the differential quantization parameter ΔQP calculated from the target video communication.

[Configuration of First Embodiment]
Next, the configuration of the video quality estimation apparatus according to the first embodiment of the present invention will be described in detail with reference to FIG.
In the video quality estimation apparatus 1 in FIG. 1, the communication I / F unit 31 includes a dedicated communication circuit unit, and has a function of performing data communication with an arbitrary communication terminal by transmitting and receiving IP packets via the IP network 3. And a function of receiving IP packets for video communication exchanged between the communication terminals 2A and 2B via the IP network 3.

The operation input unit 32 includes an operation input device such as a keyboard and a mouse, and has a function of detecting an operation of the operator and outputting the operation to the arithmetic processing unit 10.
The screen display unit 33 includes a screen display device such as an LCD or a PDP, and has a function of displaying an operation menu, an estimated video quality value, and the like according to an instruction from the arithmetic processing unit 10.

The storage unit 20 includes a storage device such as a hard disk or a memory, and has a function of storing various processing information and a program 29 used for video quality estimation processing in the arithmetic processing unit 10.
The program 29 is a program that realizes various processing units necessary for the video quality estimation processing by being read and executed by the arithmetic processing unit 10, and is connected via an input / output interface unit such as the communication I / F unit 31. And read from an external device (not shown) or a recording medium and stored in the storage unit 20 in advance.

  Main processing information stored in the storage unit 20 includes a video communication specific parameter 21, encoded video quality information 22, representative quantization parameter information 23, packet loss change index characteristic information 24, and video quality information 25.

  The video communication specifying parameter 21 is a parameter related to video quality among various parameters related to video communication to be estimated for video quality. Examples of the video communication specific parameter 21 include H.264. 264, codec type information such as MPEG, video format information such as Full HD, HD, and SD, and service type information such as IPTV, VoD, and TV phone are used. The video communication specific parameter 21 is previously input from the outside using the operation input unit 32 or the like and stored in the storage unit 20.

  Regarding the video communication specific parameter 21, in addition to the above, frame information such as intra refresh rate and GOP, information on error signals after motion compensation such as DCT coefficients, integer transform coefficients, and wavelet coefficients, packet loss rate, packet loss pattern, Packet loss information such as packet delay time may be used.

  The encoded video quality information 22 is information indicating a relationship between a combination of the video communication specific parameter 21 and various quality parameters and an encoded video quality value considering only video degradation due to compression encoding. FIG. 4 is a configuration example of encoded video quality information. Here, the encoded video quality Vc corresponding to the frame rate FR and the encoding rate BR of the video communication is registered for each combination of the CODEC, the video format, and the video communication type of the video communication specific parameter 21. In the encoded video quality information 22, the correspondence relationship between the video communication specific parameter 21 and the combination of various quality parameters and the encoded video quality Vc may be calculated in advance from various video communications.

  The representative quantization parameter information 23 indicates a relationship between a combination of the video communication specific parameter 21 and various quality parameters of video communication and a representative quantization parameter that is a representative value of a quantization parameter related to a quantization step width of compression coding. Information. FIG. 5 is a configuration example of representative quantization parameter information. Here, the representative quantization parameter QPt corresponding to the video communication coding rate BR is registered for each combination of the CODEC, video format, and video communication type of the video communication specific parameter 21.

  In the representative quantization parameter information 23, the correspondence relationship between the combination of the video communication specific parameter 21 and various quality parameters and the representative quantization parameter may be calculated in advance from various video communications. At this time, a plurality of typical videos v having different motion amounts are prepared, and an average quantization parameter QP is calculated for each of these videos v using Equation (1) to be described later, and video communication specific parameters for each video v are calculated. For each combination of 21 and various quality parameters, an average value of the average quantization parameter QP may be obtained and used as the representative quantization parameter QPt.

The packet loss change index characteristic information 24 is information indicating a correspondence relationship between the video communication specific parameter 21 of video communication and the coefficients a and b specifying the packet loss change index calculation characteristic. FIG. 6 is a configuration example of packet loss change index characteristic information.
The packet loss change index characteristic is a characteristic indicating the relationship between the differential quantization parameter ΔQP and the packet loss change index D, and can be represented by an approximate function such as a linear function as shown in FIG. In the packet loss change index characteristic information 24, as shown in FIG. 6, a coefficient for specifying such a linear function is registered for each combination of CODEC, video format, and video communication type of the video communication specifying parameter 21. . The correspondence between the video communication specific parameter 21 and the coefficients a and b in the packet loss change index characteristic information 24 may be calculated in advance from the packet loss change index characteristics derived for each type of video.

The video quality information 25 is information indicating a correspondence relationship between the video communication specific parameter 21 of video communication, the combination of the encoded video quality value Vc, the number of invalid frames IF, and the packet loss change index D and the video quality value V. . FIG. 7 is a configuration example of video quality information. Here, the encoded video quality value Vc, the number of invalid frames IF, and the video quality value V corresponding to the packet loss change index D are registered for each combination of the CODEC, the video format, and the video communication type of the video communication specific parameter 21. ing.
In the video quality information 25, the correspondence between the video communication specific parameter 21, the encoded video quality value Vc, the number of invalid frames IF, the packet loss change index D, and the video quality value V is calculated in advance from various video communications. You just have to.

The arithmetic processing unit 10 includes a microprocessor such as a CPU and its peripheral circuits, and reads and executes the program 29 in the storage unit 20 to cause the hardware and the program 29 to cooperate with each other, thereby improving video quality. It has a function of realizing various processing units necessary for estimation processing.
The main processing units realized by the arithmetic processing unit 10 include an IP packet analysis unit 11, an invalid frame number calculation unit 12, an encoded video quality calculation unit 13, a representative quantization parameter calculation unit 14, and an average quantization parameter calculation unit 15. A packet loss change index calculation unit 16 and a video quality estimation unit 17.

The IP packet analysis unit 11 receives various quality parameters related to video quality and quantization parameters related to the quantization step width of the compression coding of the video from the IP packet of the target video communication received by the communication I / F unit 31. It has a function to extract.
Based on the quality parameter frame type information ft, the frame number fn, and the lost frame information fd extracted by the IP packet analysis unit 11, the invalid frame count calculation unit 12 It has a function of calculating the number of invalid frames made up of other frames affected by packet loss.

The encoded video quality calculation unit 13 is based on the frame rate FR and encoding rate BR of the quality parameters extracted by the IP packet analysis unit 11, the video communication specifying parameter 21 and the encoded video quality information 22 in the storage unit 20. And a function of calculating an encoded video quality value Vc considering only video deterioration due to compression encoding.
The representative quantization parameter calculation unit 14 performs quality estimation based on the coding rate BR of the quality parameter extracted by the IP packet analysis unit 11, the video communication specific parameter 21 and the representative quantization parameter information 23 in the storage unit 20. It has a function of calculating a representative quantization parameter QPt that represents the same type of video communication as the target video communication.

The average quantization parameter calculation unit 15 has a function of calculating an average quantization parameter QP indicating an average value based on the quantization parameter of each frame corresponding to the quality estimation section extracted by the IP packet analysis unit 11. is doing.
The packet loss change index calculation unit 16 calculates a packet loss change index D indicating the amount of motion of the target video based on the video communication specific parameter 21, the average quantization parameter QP, and the representative quantization parameter QPt in the storage unit 20. It has a function.
The video quality estimator 17 calculates the video quality value V of the target video communication based on the encoded video quality value Vc, the number of invalid frames IF, and the packet loss change index D, and stores this video quality value V. A function of outputting to the unit 20 and the screen display unit 33.

[Operation of First Embodiment]
Next, the operation of the video quality estimation apparatus according to the first embodiment of the present invention will be described with reference to FIGS. FIG. 8 is a flowchart showing the video quality estimation operation of the video quality estimation apparatus according to the first embodiment of the present invention. FIG. 9 is a flowchart showing video quality estimation processing of the video quality estimation apparatus according to the first embodiment of the present invention.
The arithmetic processing unit 10 of the video quality estimation apparatus 1 starts the video quality estimation process of FIG. 9 when an operator operation indicating the start of the video quality estimation process is detected by the operation input unit 32.

First, the arithmetic processing unit 10 receives an IP packet of target video communication that is a target for video quality estimation by the communication I / F unit 31, and the IP packet analysis unit 11 uses the IP packet to perform various quality parameters related to video quality. And a quantization parameter relating to a quantization step width of compression encoding of the video (step 100).
In an IP packet, there are information such as an IP header, UDP (User Datagram Protocol) header, RTP (Real-time Transport Protocol) header, TS (Transport stream), ES (Elementaly stream), etc. To extract.

  For the coding rate (bit rate) BR, the code amount (number of bits) required for encoding the target video included in the quality estimation interval is counted, and the code amount per unit time calculated from the code amount is calculated. The encoding rate BR may be used. In this embodiment, the coding rate BR in the quality estimation interval is used, but other statistical information indicating the code amount, such as the number of IP packets in the quality estimation interval, may be used. In addition, when the service type of the target video communication is determined, since it is often transmitted at a fixed encoding rate, the encoding rate BR corresponding to the target video communication is not calculated from the IP packet. The video communication specific parameter 21 may be externally input.

  The frame rate FR may be obtained from the description in the IP packet. Specifically, information on the frame rate is directly described in the portion where the video encoding information in the IP packet is described, or information that can be converted into the frame rate is described. The rate can be specified. Compression encoding method In the case of H.264, it is specified by timing_info_present_flag, Num_units_in_tick, Time_scale, and Fixed_frame_rate_flag. When these parameters are 1,1001, 60000, and 1, respectively, the frame rate FR is 29.97 fps.

The frame type ft may be obtained from the description in the IP packet. Specifically, the compression encoding method H.264 is used. In the case of H.264, there is a bit string called primary_pic_type. The primary_pic_type indicates the type of frame that the frame can take. When this parameter is 0, 1, or 2, it may be determined that the frame is an I, P, or B frame, respectively.
For the frame number fn, a convenient number may be assigned to each frame in the quality estimation section. Compression encoding method In the case of H.264 IP packets and TS packets, since the frame start position can be determined by the Access unit delimiter in Nal_unit_type, each time the Access unit delimiter is identified, the frame number fn may be counted and added one by one. However, in the case of having a slice structure, there are as many Access unit delimiters as the number of slices. For example, when the number of slices is 6, there are also 6 Access unit delimiters. In such a case, the frame may be counted in consideration of the number of slices.

The lost frame number fd is composed of the frame number fn of the frame in which the packet loss has occurred among the frames in the quality estimation section.
As for the quantization parameter qp, the encoding information in the IP packet describes the quantization parameter qp (Quantization parameter) for each block constituting the frame.

The arithmetic processing unit 10 thus extracts various quality parameters and quantization parameters by the packet analysis unit 11, and then uses the invalid frame number calculation unit 12 to determine a frame with a packet loss and a packet of the frame at the time of decoding. The number of invalid frames consisting of other frames affected by the loss is calculated (step 101).
The invalid frame number calculation unit 12 identifies an invalid frame based on the frame type information ft, the frame number fn, and the lost frame information fd among the various quality parameters extracted by the packet analysis unit 11.

  FIG. 10 is an explanatory diagram showing the relationship between the frame type and the deterioration perception range. When the inter-frame predictive coding method is used, the propagation of the frame deterioration differs depending on the frame type of the frame in which the packet loss has occurred, and as a result, the perceived deterioration perception range differs. For example, as shown in FIG. 10, when a packet loss occurs in the I frame (f3), the frame degradation is caused by the I frame (f1) received immediately after the last received P frame (f0) ( Deterioration propagates to the frame (f17) immediately before f18).

  In addition, when packet loss occurs in the P frame (f12), the frame degradation is caused by the frame (f17) immediately before the I frame (f18) received next from the frame (f10) immediately after the last received P frame (f9). ) Until the deterioration propagates. On the other hand, when a packet loss occurs in the B frame (f15), the frame degradation is only the own frame (f15). Based on the deterioration perception range for each frame type, the deteriorated frame and the frame in which the deterioration has propagated are specified as invalid frames, and the number of invalid frames is counted.

Subsequently, the arithmetic processing unit 10 uses the encoded video quality calculation unit 13 to calculate an encoded video quality value Vc considering only video degradation due to compression encoding (step 102). This encoded video quality value Vc corresponds to the video quality value of the video obtained by decoding each frame immediately after compression encoding.
The encoded video quality calculation unit 13 refers to the encoded video quality information 22 in the storage unit 20, and among the various quality parameters extracted by the packet analysis unit 11, the frame rate FR and the encoding rate BR and the video in the storage unit 20. An encoded video quality value Vc corresponding to the communication specific parameter 21 is acquired.

Next, the arithmetic processing unit 10 uses the representative quantization parameter calculation unit 14 to calculate a representative quantization parameter QPt that represents the same type of video communication as the target video communication (step 103).
The representative quantization parameter calculation unit 14 refers to the representative quantization parameter information 23 in the storage unit 20, determines the encoding rate BR of various quality parameters extracted by the packet analysis unit 11 and the video communication specific parameter 21 in the storage unit 20. A corresponding representative quantization parameter QPt is calculated.

Subsequently, the arithmetic processing unit 10 uses the average quantization parameter calculation unit 15 to calculate the average quantization parameter QP of each frame in the quality estimation section (step 104).
The quality estimation section has a time length of, for example, 10 seconds or 300 frames. For each frame in the quality estimation section, the quantization parameter is extracted by the IP packet analysis unit 11 for each block constituting the frame. A block is a processing unit of compression encoding in each frame, and is arbitrarily set for each frame.

FIG. 11 is an explanatory diagram showing the quantization parameter of each block. F is the number of frames in the quality estimation interval, N is the number of horizontal blocks constituting each frame, M is the number of vertical blocks constituting each frame, f is the frame number of a frame existing in the quality estimation interval, The horizontal block number of an arbitrary block in the frame f is n, the vertical block number of an arbitrary block in the frame f is m, and the quantization parameter of a block having the horizontal block number n and the vertical block number m is qp _mnf . In this case, the quantization parameter QP of the frame f can be calculated by the following equation (1).

Thereafter, the arithmetic processing unit 10 uses the packet loss change index calculation unit 16 to calculate a packet loss change index D indicating the amount of motion of the target video (step 105).
The packet loss change index calculation unit 16 first calculates the following equation based on the difference between the representative quantization parameter QPt calculated by the representative quantization parameter calculation unit 14 and the average quantization parameter QP calculated by the average quantization parameter calculation unit 15. Based on (2), the difference quantization parameter ΔQP is calculated.

Subsequently, the packet loss change index calculation unit 16 refers to the packet loss change index characteristic information 24 in the storage unit 20 and corresponds to the coefficient a, the packet loss change index characteristic corresponding to the video communication specific parameter 21 in the storage unit 20. Get b. Then, a packet loss change index D is calculated from the coefficients a and b and the difference quantization parameter ΔQP based on the following equation (3).

Thereafter, the arithmetic processing unit 10 uses the video quality estimation unit 17 to calculate the video quality value V of the target video communication (step 106).
The video quality estimation unit 17 refers to the video quality information 25 in the storage unit 20 and refers to the video communication specific parameter 21 in the storage unit 20, the invalid frame number IF calculated by the invalid frame number calculation unit 12, and the encoded video quality calculation unit. The video quality value V corresponding to the encoded video quality value Vc calculated at 13 and the packet loss change index D calculated by the packet loss change index calculation unit 16 is acquired.

[Effect of the first embodiment]
Thus, in this embodiment, the average quantization parameter indicating the average value of the quantization parameter for each frame of the target video communication corresponding to the quality estimation section, and the representative quantization representing the quantization parameter of the target video communication Based on the parameters, a packet loss change index indicating the amount of motion of the video between the target frames is calculated, and based on the encoded video quality value, the number of invalid frames, and the packet loss change index related to the target video communication, The video quality value of communication is calculated.

  Thereby, the packet loss change index indicating the motion of the video can be calculated from the quantization parameter described in each frame of the target video communication without calculating the pixel values of the two adjacent frames. Therefore, it is possible to calculate the video quality value considering the motion amount of the video with a small processing load, and even for the majority of the videos whose motion amount of the video is not known in advance, the video quality considering the effect of the motion amount of the video is estimated in real time. be able to.

[Second Embodiment]
Next, with reference to FIGS. 12 and 13, a video quality estimation apparatus according to the second embodiment of the present invention will be described. FIG. 12 is an explanatory diagram showing representative quantization parameter characteristics. FIG. 13 is an explanatory diagram showing representative quantization parameter characteristic information.

  In the first embodiment, when the representative quantization parameter QPt is calculated, the representative quantization parameter calculation unit 14 refers to the representative quantization parameter information 23 in the storage unit 20 and extracts various qualities extracted by the packet analysis unit 11. The case where the representative quantization parameter QPt corresponding to the parameter coding rate BR and the video communication specific parameter 21 of the storage unit 20 is calculated has been described.

In the present embodiment, the representative quantization parameter QPt is calculated based on a function indicating the representative quantization parameter characteristic.
As shown in FIG. 12, the coding rate-representative quantization parameter characteristic has a characteristic that the representative quantization parameter QPt decreases as the coding rate BR of video communication increases. When the video communication coding rate is BR and the coefficients for specifying the representative quantization parameter characteristic function are t1, t2, and t3, the representative quantization parameter QPt is expressed by the following equation (4).

  This representative quantization parameter characteristic function varies depending on the type of video communication. In the present embodiment, the coefficients t1, t2, and t3 are derived in advance for each video communication specific parameter 21, and the correspondence between the video communication specific parameter 21 and the coefficients t1, t2, and t3 as shown in FIG. The representative quantization parameter characteristic information 26 shown is stored in the storage unit 20.

  Accordingly, in step 103 of FIG. 9 described above, the representative quantization parameter calculation unit 14 refers to the representative quantization parameter characteristic information 26 in the storage unit 20 and the coefficient t1 corresponding to the video communication specific parameter 21 in the storage unit 20. , T2, t3, and from the coefficients t1, t2, t3 and the coding rate BR of various quality parameters extracted by the packet analysis unit 11, based on the above-described equation (4), A representative quantization parameter QPt representative of video communication is calculated.

[Effect of the second embodiment]
As described above, according to the present embodiment, the representative quantization parameter information 23 having a large amount of information is prepared in advance as compared with the case where the representative quantization parameter information 23 described in the first embodiment is used. Even when the encoding rate BR of the target video communication changes finely, it is possible to calculate the video quality value considering the motion amount of the video with a small processing load.

[Third Embodiment]
Next, a video quality estimation apparatus according to the third embodiment of the present invention will be described.
In the first embodiment, when calculating the video quality value, the video quality estimation unit 17 refers to the video quality information 25 in the storage unit 20 to calculate the video communication specific parameter 21 in the storage unit 20 and the number of invalid frames. Video quality value corresponding to the number of invalid frames IF calculated by the unit 12, the encoded video quality value Vc calculated by the encoded video quality calculation unit 13, and the packet loss change index D calculated by the packet loss change index calculation unit 16 The case where V is acquired has been described.

In the present embodiment, the video quality value V is calculated based on a function indicating the video quality characteristic.
As shown in FIG. 2, the video quality characteristic indicating the relationship between the number of invalid frames IF and the video quality value V in consideration of the packet loss change index D is the number of invalid frames IF shown in FIG. When = 0, the encoded video quality value Vc is set to the maximum video quality value, and the video quality value decreases exponentially as the number of invalid frames IF increases. This video quality characteristic can be approximated by the following equation (5) using an exponential function.

  The video quality estimation unit 17 uses the invalid frame number IF calculated by the invalid frame number calculation unit 12, the encoded video quality value Vc calculated by the encoded video quality calculation unit 13, and the packet calculated by the packet loss change index calculation unit 16. From the loss change index D, the video quality value V of the target video communication is calculated based on the above formula (5).

[Effect of the third embodiment]
As described above, according to the present embodiment, compared with the case where the video quality information 25 described in the first embodiment is used, even when the packet loss change index D of the target video communication changes finely, it is large. Without preparing the video quality information 25 of the information amount in advance, it is possible to calculate the video quality value considering the motion amount of the video with a small processing burden.

It is a block diagram which shows the structure of the video quality estimation apparatus concerning the 1st Embodiment of this invention. It is explanatory drawing which shows the relationship between the amount of motion of an image | video, and the number of invalid frames-subjective evaluation value characteristics. It is explanatory drawing which shows a packet loss change parameter | index calculation characteristic. It is an example of a structure of encoding video quality information. It is an example of a structure of representative quantization parameter information. It is a structural example of packet loss change index characteristic information. It is a structural example of video quality information. It is a flowchart which shows the video quality estimation operation | movement of the video quality estimation apparatus concerning the 1st Embodiment of this invention. It is a flowchart which shows the video quality estimation process of the video quality estimation apparatus concerning the 1st Embodiment of this invention. It is explanatory drawing which shows the relationship between a frame classification and a degradation perception range. It is explanatory drawing which shows the quantization parameter of each block. It is explanatory drawing which shows a representative quantization parameter characteristic. It is explanatory drawing which shows representative quantization parameter characteristic information.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Video quality estimation apparatus, 10 ... Operation processing part, 11 ... Packet analysis part, 12 ... Invalid frame number calculation part, 13 ... Encoded video quality calculation part, 14 ... Representative quantization parameter calculation part, 15 ... Average quantization Parameter calculation unit 16 ... Packet loss change index calculation unit, 17 ... Video quality estimation unit, 20 ... Storage unit, 21 ... Video communication specific parameter, 22 ... Encoded video quality information, 23 ... Representative quantization parameter information, 24 ... Packet loss change index characteristic information, 25 ... Video quality information, 26 ... Representative quantization parameter characteristic information, 29 ... Program, 31 ... Communication I / F unit, 32 ... Operation input unit, 33 ... Screen display unit, 2A, 2B ... Communication terminal, 3 ... IP network.

Claims (6)

A video quality estimation device that estimates video quality of target video communication in which a target video is compressed and encoded into a plurality of frames based on an intra-frame and inter-frame predictive encoding scheme and is transmitted as an IP packet,
An IP packet analysis unit that analyzes the input IP packet and extracts various quality parameters related to video quality of the target video communication and a quantization parameter related to a quantization step width of the compression encoding;
Based on the quality parameter, an invalid frame number calculating unit that calculates the number of invalid frames including a frame having a packet loss and another frame that is affected by the packet loss of the frame at the time of decoding;
An encoded video quality calculation unit that calculates an encoded video quality value of the target video communication in consideration of only video degradation due to the compression encoding based on the quality parameter;
An average quantization parameter calculation unit for calculating an average value of quantization parameters for each frame of the target video communication corresponding to a quality estimation section;
A representative quantization parameter calculation unit for calculating a representative quantization parameter representing the quantization parameter of the target video communication;
Based on the average quantization parameter and the representative quantization parameter, a packet loss change index calculation unit that calculates a packet loss change index indicating the amount of motion of video between the target frames;
A video quality estimation device comprising: a video quality calculation unit that calculates a video quality value of the target video communication based on the encoded video quality value, the number of invalid frames, and the packet loss change index. The video quality estimation apparatus according to claim 1,
The packet loss change index calculation unit is based on a function indicating a characteristic that the packet loss change index monotonously decreases in accordance with an increase in a differential quantization parameter indicating a difference between the average quantization parameter and the representative quantization parameter. A video quality estimation apparatus that calculates the packet loss change index. The video quality estimation apparatus according to claim 1,
The representative quantization parameter calculation unit calculates a quantization parameter corresponding to a coding rate included in the quality parameter based on a function indicating a characteristic that the quantization parameter monotonously decreases as the coding rate increases. A video quality estimation apparatus characterized by being calculated as a quantization parameter. The video quality estimation apparatus according to claim 1,
The video quality calculation unit sets the encoded video quality value as the maximum value of the video quality value when the number of invalid frames is 0, and is based on a function indicating a characteristic that the video quality value monotonously decreases as the number of invalid frames increases. A video quality estimation apparatus that calculates a video quality value corresponding to the encoded video quality value, the number of invalid frames, and the packet loss change index. A video quality estimation method used in a video quality estimation apparatus for estimating video quality of target video communication in which a target video is compressed and encoded into a plurality of frames based on an intra-frame and inter-frame predictive coding scheme and transmitted as an IP packet. ,
An IP packet analysis unit that analyzes the input IP packet and extracts various quality parameters related to the video quality of the target video communication and quantization parameters related to the quantization step width of the compression coding Steps,
The invalid frame number calculation unit calculates the number of invalid frames based on the quality parameter and calculates the number of invalid frames including a frame having a packet loss and another frame affected by the packet loss of the frame at the time of decoding. When,
An encoded video quality calculating unit that calculates an encoded video quality value of the target video communication in consideration of only video degradation due to the compression encoding based on the quality parameter;
An average quantization parameter calculating unit that calculates an average value of the quantization parameter for each frame of the target video communication corresponding to the quality estimation section; and
A representative quantization parameter calculating unit that calculates a representative quantization parameter that represents the quantization parameter of the target video communication; and
A packet loss change index calculating unit that calculates a packet loss change index indicating a motion amount of the video between the target frames based on the average quantization parameter and the representative quantization parameter; ,
A video quality calculation unit comprising: a video quality calculation step of calculating a video quality value of the target video communication based on the encoded video quality value, the number of invalid frames, and the packet loss change index. Video quality estimation method. In a computer of a video quality estimation device that estimates video quality of target video communication that is compressed and encoded into a plurality of frames based on an intra-frame and inter-frame predictive coding scheme and is transmitted as an IP packet,
An IP packet analysis unit that analyzes the input IP packet and extracts various quality parameters related to the video quality of the target video communication and quantization parameters related to the quantization step width of the compression coding Steps,
The invalid frame number calculation unit calculates the number of invalid frames based on the quality parameter and calculates the number of invalid frames including a frame having a packet loss and another frame affected by the packet loss of the frame at the time of decoding. When,
An encoded video quality calculating unit that calculates an encoded video quality value of the target video communication in consideration of only video degradation due to the compression encoding based on the quality parameter;
An average quantization parameter calculating unit that calculates an average value of the quantization parameter for each frame of the target video communication corresponding to the quality estimation section; and
A representative quantization parameter calculating unit that calculates a representative quantization parameter that represents the quantization parameter of the target video communication; and
A packet loss change index calculating unit that calculates a packet loss change index indicating a motion amount of the video between the target frames based on the average quantization parameter and the representative quantization parameter; ,
A video quality calculation step in which a video quality calculation unit calculates a video quality value of the target video communication based on the encoded video quality value, the number of invalid frames, and the packet loss change index.

Images