JP5394991B2 - Video frame type estimation adjustment coefficient calculation method, apparatus, and program - Google Patents

Description

  The present invention relates to a video frame type estimation technique, and more particularly to a video frame type estimation adjustment coefficient calculation technique used when estimating the frame type of an encrypted video frame.

  MPEG-2 standard (ISO / IEC 13818) and MPEG-4 standard (ISO / IEC 14496, ITU-) for video distribution such as digital broadcasting using radio waves and IP retransmission of digital broadcasting using communication networks. When a compression encoding method compliant with MPEG standards such as T H.264) is used, the original video is compressed and encoded into frames (pictures) such as I, P, B, etc., and an elementary stream (Elementary Stream), etc. Video streams are generated and divided into packets called TS packets (Transport Stream Packets) for distribution.

  In such video distribution, when copyright considerations are necessary, the payload portion in which video data and audio data are stored in the TS packet is encrypted. The encrypted TS packet is decrypted after being received by the receiving terminal, finally outputted as a video, and provided to the viewer.

  On the other hand, in these video applications, if the air quality deteriorates if the video is distributed using radio waves, or if packet loss or delay occurs in the network or the terminal if the video is distributed using the network, either However, the video quality deteriorates and the service quality deteriorates. Therefore, in order to provide a stable service, it is necessary to manage the quality of the communication network and the terminal so as to ensure a certain video quality.

In such quality management, it is necessary to measure the influence of atmospheric conditions and network quality (for example, packet loss and delay) on the video quality viewed by the user, and manage the quality state.
When TS packet loss occurs in an atmospheric state, a network, or a terminal, a video signal transmitted by compressing video by high-efficiency coding uses motion prediction coding between frames. For this reason, when decoding an arbitrary video frame, the frame information of the previous and subsequent video frames is used, so that the video quality deterioration due to the loss of one packet is not limited to one video frame, and may extend to a plurality of frames.

  In such a case, the quality of the decoded video is degraded even though there is no packet loss in the video frame decoded with reference to the degraded video frame. For this reason, the atmospheric state or the packet loss occurrence state on the network does not necessarily correspond to the deterioration occurrence state in the video application. Therefore, in order to accurately obtain the correspondence relationship between the atmospheric state, network quality, and application quality, it is necessary to accurately grasp the influence range of quality degradation based on the encoded information.

Conventionally, as a technique for grasping the range of influence of quality degradation based on encoded information in a non-encrypted video stream, a method for calculating the number of frames in which degradation spreads from the type of frame where degradation has occurred and the status of previous and subsequent frames has been proposed. (See, for example, Patent Document 1).
In addition, in the encrypted video stream, the determination threshold value calculated for each 1 GOP (Group Of Pictures) and the intra-frame data amount of each video frame using the fact that the intra-frame data amount differs for each video frame type. And a method for estimating the types of video frames, that is, I, P, and B frames have been proposed (see, for example, Patent Document 2).

JP 2007-060475 A International Publication WO2009 / 025357 Pamphlet JP 2006-033722 A Japanese Patent No. 3764394

  In such a conventional technique, when the frame type is estimated by comparing the amount of data in each video frame with the determination threshold, the estimation accuracy depends on the determination threshold. Must be calculated accurately. In particular, in Patent Document 2, the determination threshold value is calculated from the average value of the in-frame data amount of the target frame calculated from the target encrypted video stream and the adjustment coefficient calculated in advance. Yes.

Here, as the adjustment coefficient for obtaining the determination threshold, the optimum adjustment coefficient is selected from the temporary adjustment coefficients that are candidates for the optimum adjustment coefficient based on the frame type estimation accuracy. For this reason, it is necessary to calculate the adjustment coefficient in advance using an unencrypted video stream whose frame type is known in advance.
However, when the above-described encrypted video stream is a frame type determination target, it is often difficult to obtain a non-encrypted video stream. Under such circumstances, the adjustment coefficient cannot be calculated. There was a problem.

  The present invention is for solving such a problem, and an adjustment coefficient calculation for video frame type estimation capable of calculating an adjustment coefficient for obtaining a determination threshold without requiring an unencrypted video stream. The purpose is to provide technology.

In order to achieve such an object, the video frame type estimation adjustment coefficient calculation method according to the present invention includes an intra-frame data of each target frame that is an estimation target of a frame type among video frames constituting a video stream. A video frame type estimation device that estimates the frame type of these target frames by comparing the amount with a determination threshold value calculated by adjusting the average value of the intra-frame data amount of these target frames with a predetermined adjustment coefficient A video frame type estimation adjustment coefficient calculation method used in a video frame type estimation adjustment coefficient calculation apparatus for calculating an optimal adjustment coefficient that is an optimal value of an adjustment coefficient, wherein the frame type estimation unit In-frame data of each target frame that is the target of estimation of the frame type among each video frame A frame type estimation step for estimating the frame type of these target frames by comparing the average value of the in-frame data amount of these target frames with a determination threshold value calculated by adjusting with the designated temporary adjustment coefficient The feature value calculation unit estimates the frame type of the target frame by the frame type estimation unit using the temporary adjustment coefficient for each of a plurality of temporary adjustment coefficients that are candidates for the optimal adjustment coefficient, and the obtained temporary estimation result A characteristic value calculating step for calculating a temporary characteristic value indicating a characteristic specific to the video stream from the storage unit, and a characteristic that is specific to the video stream obtained by the storage unit when the frame type of the video stream is legitimately determined. A storage step that prestores the indicated feature value as a reference feature value, and an optimum adjustment coefficient calculation unit compares each of the temporary feature values with the reference feature value, And a optimum adjustment coefficient calculation step of calculating the optimum adjustment coefficient based on the result, the temporary characteristic value was calculated on the basis of the temporary estimated result, from the temporary frame ratio showing a component ratio of any frame type relating to the video stream The reference feature value includes a reference frame ratio indicating a composition ratio of an arbitrary frame type related to the video stream, calculated based on the video stream configuration information of the video stream, and the optimal adjustment coefficient calculating step includes: The method includes a step of selecting an optimum adjustment coefficient that has the smallest difference between the temporary frame ratio and the reference frame ratio.

In addition, another video frame type estimation adjustment coefficient calculation method according to the present invention is a method of calculating the intra-frame data amount of each target frame, which is a frame type estimation target, among the video frames constituting the video stream. For the video frame type estimation device for estimating the frame type of these target frames by comparing the average value of the in-frame data amount with a determination threshold value adjusted by a predetermined adjustment coefficient, the optimum value of the adjustment coefficient A video frame type estimation adjustment coefficient calculation method used in a video frame type estimation adjustment coefficient calculation apparatus for calculating an optimal adjustment coefficient, wherein the frame type estimation unit is configured to calculate each of the video frames constituting the video stream. Of these, the amount of in-frame data of each target frame for which the frame type is to be estimated A frame type estimation step for estimating the frame type of these target frames by comparing the average value of the in-frame data amount with a determination threshold value calculated by adjusting with a specified temporary adjustment coefficient, and a feature value calculation For each of a plurality of temporary adjustment coefficients that are candidates for the optimum adjustment coefficient, the frame type estimation unit estimates the frame type of the target frame using the temporary adjustment coefficient, and from the obtained temporary estimation result to the video stream. A feature value calculating step for calculating a temporary feature value indicating a unique feature, and a storage unit based on a feature value indicating a feature unique to the video stream, which is obtained when the frame type of the video stream is legitimately determined. A storage step that stores in advance as feature values and an optimum adjustment coefficient calculation unit compare each of the temporary feature values with the reference feature value, and perform an optimum adjustment based on the comparison result. And a optimum adjustment coefficient calculation step of calculating the coefficients, the temporary characteristic value, at least every various video communication conditions, the obtained from the video stream delivered by the video communication conditions, the provisional estimation result and each video frame frame It consists of provisional video quality characteristics indicating the relationship between the invalid frame rate and the video quality value, which is specified by the invalid frame rate calculated from the loss situation, and the reference feature value is at least for each video communication condition. It consists of reference video quality characteristics indicating the relationship between the invalid frame rate related to the video stream and the video quality value obtained by evaluating the video stream distributed below in advance. Select a correlation coefficient between the provisional video quality characteristic and the reference video quality characteristic within the range from the maximum value to a predetermined correlation threshold value. In addition, a step of calculating a representative value of these temporary adjustment coefficients as an optimal adjustment coefficient is included .

Also, the video frame type estimation adjustment coefficient calculating apparatus according to the present invention calculates the intra-frame data amount of each target frame to be estimated for the frame type among the video frames constituting the video stream. The optimum value of the adjustment coefficient for the video frame type estimation device that estimates the frame type of these target frames by comparing the average value of the internal data amount with the determination threshold value adjusted by adjusting with a predetermined adjustment coefficient An adjustment coefficient calculating apparatus for estimating a video frame type for calculating an optimal adjustment coefficient, and the amount of data in a frame of each target frame to be estimated for the frame type among the video frames constituting the video stream, Judgment calculated by adjusting the average value of the amount of data in the target frame with the specified temporary adjustment coefficient A frame type estimation unit that estimates the frame type of these target frames by comparing with the threshold value, and a frame type estimation unit that uses the temporary adjustment coefficient for each of a plurality of temporary adjustment coefficients that are candidates for the optimal adjustment coefficient. A feature value calculation unit that estimates a frame type of a target frame, calculates a temporary feature value indicating a characteristic unique to the video stream from the obtained temporary estimation result, and when a frame type of the video stream is properly determined A storage unit that preliminarily stores feature values indicating features unique to the video stream as reference feature values, and compares each of the temporary feature values with the reference feature values, and calculates an optimum adjustment coefficient based on the comparison results and a optimal adjustment coefficient calculation unit for temporary feature value indicates was calculated based on the provisional estimation result, the composition ratio of any frame type relating to the video stream The reference feature value is composed of a reference frame ratio indicating a composition ratio of an arbitrary frame type related to the video stream, which is calculated based on the video stream configuration information of the video stream, and is temporarily adjusted by the optimum adjustment coefficient calculation unit. Among the coefficients, the coefficient with the smallest difference between the temporary frame ratio and the reference frame ratio is selected as the optimum adjustment coefficient.

In addition, another video frame type estimation adjustment coefficient calculation apparatus according to the present invention is configured to calculate the amount of data in each frame of each target frame to be estimated for the frame type among the video frames constituting the video stream. For the video frame type estimation device for estimating the frame type of these target frames by comparing the average value of the in-frame data amount with a determination threshold value adjusted by a predetermined adjustment coefficient, the optimum value of the adjustment coefficient An adjustment coefficient calculating apparatus for estimating a video frame type for calculating an optimal adjustment coefficient, wherein an amount of intra-frame data of each target frame to be estimated for a frame type is calculated from each video frame constituting a video stream. , Calculated the average value of the amount of data in the target frame by adjusting the specified temporary adjustment coefficient A frame type estimation unit that estimates the frame type of these target frames by comparing with a constant threshold, and a frame type estimation using the temporary adjustment coefficient for each of a plurality of temporary adjustment coefficients that are candidates for the optimal adjustment coefficient The frame type of the target frame is estimated by the unit, a feature value calculation unit that calculates a temporary feature value indicating a characteristic unique to the video stream from the obtained temporary estimation result, and the frame type of the video stream is correctly determined A storage unit that preliminarily stores characteristic values indicating characteristics unique to the video stream obtained as reference characteristic values, and compares each of the temporary characteristic values with the reference characteristic values, and based on the comparison results, an optimum adjustment coefficient and a optimal adjustment coefficient calculation section that calculates a temporary feature values, at least every various video communication conditions, obtained from a video stream delivered by the video communication conditions Is composed of temporary video quality characteristics indicating the relationship between the invalid frame rate and the video quality value specified by the invalid frame rate calculated from the temporary estimation result and the frame loss situation of each video frame. At least for each video communication condition, comprising a reference video quality characteristic indicating a relationship between an invalid frame rate and a video quality value related to the video stream, obtained by evaluating a video stream distributed under the video communication condition in advance, The optimum adjustment coefficient calculation unit selects a temporary adjustment coefficient that includes a correlation coefficient between the temporary video quality characteristic and the reference video quality characteristic within a range from the maximum value to a predetermined correlation threshold value. , in which to calculate the representative value of provisional adjustment coefficient as the optimum adjustment factor.

  In addition, the program according to the present invention is a program for causing a computer to function as each of the components that constitute any one of the above-described video frame type estimation adjustment coefficient calculation apparatuses.

  According to the present invention, it is possible to evaluate the estimation result of the frame type when using each temporary adjustment coefficient without using the frame type of the video stream used for calculating the optimum adjustment coefficient. Therefore, even in a video stream in which the payload portion of the TS packet is encrypted and the video data itself cannot be processed, the adjustment coefficient used for identifying the frame type can be automatically calculated.

It is a block diagram which shows the structure of the adjustment coefficient calculation apparatus for video frame classification estimation concerning 1st Embodiment. It is a flowchart which shows a frame classification estimation process. It is explanatory drawing which shows a frame classification estimation process. It is a flowchart which shows the optimal adjustment coefficient calculation process concerning 1st Embodiment. It is a block diagram which shows the structure of the adjustment coefficient calculation apparatus for video frame classification estimation concerning 2nd Embodiment. It is a flowchart which shows the optimal adjustment coefficient calculation process concerning 2nd Embodiment. It is a pseudo code example of the optimal adjustment coefficient calculation process concerning 2nd Embodiment. It is explanatory drawing which shows the optimal adjustment coefficient calculation process concerning 2nd Embodiment. It is a block diagram which shows the structure of the adjustment coefficient calculation apparatus for video frame classification estimation concerning 3rd Embodiment. It is a flowchart which shows the optimal adjustment coefficient calculation process concerning 3rd Embodiment. It is a pseudo code example of the optimal adjustment coefficient calculation process concerning 3rd Embodiment. It is an example of calculation of a temporary video quality characteristic concerning a 3rd embodiment. It is explanatory drawing which shows the optimal adjustment coefficient calculation process concerning 3rd Embodiment.

Next, embodiments of the present invention will be described with reference to the drawings.
[First Embodiment]
First, a video frame type estimation adjustment coefficient calculating apparatus according to a first embodiment of the present invention will be described with reference to FIG. FIG. 1 is a block diagram illustrating a configuration of a video frame type estimation adjustment coefficient calculation apparatus according to the first embodiment.

  In the above-described video frame type estimation technique (see, for example, Non-Patent Document 2), an encrypted MPEG standard video stream distributed in TS packets via a packet communication network such as the Internet or digital broadcasting Using the fact that the amount of data in a frame differs for each video frame type, out of the video frames constituting the video stream, the amount of data in the frame of each target frame that is the target of estimation of the frame type The frame type of these target frames is estimated by comparing the average value of the internal data amount with a determination threshold value calculated by adjusting with a predetermined adjustment coefficient.

At this time, the adjustment coefficient is a very important coefficient for improving the estimation accuracy of the frame type, and needs to be calculated in advance before the actual frame type process is executed.
A video frame type estimation adjustment coefficient calculation device (hereinafter referred to as an adjustment coefficient calculation device) 10 is composed of an information processing device such as a computer for processing input data, such as a server device or a personal computer, as a whole. It has a function of calculating an adjustment coefficient used when estimating the frame type by the estimation apparatus from an encrypted MPEG standard video stream prepared in advance.

  In the present embodiment, a frame type estimation unit that estimates the frame type of the target frame is provided in the same manner as the video frame type estimation device, and a plurality of temporary adjustment coefficients that are candidates for the optimal adjustment coefficient that is the optimal value of the adjustment coefficient In addition, the frame type estimation unit estimates the frame type of the target frame using the temporary adjustment coefficient, calculates temporary feature values indicating features unique to the video stream from the obtained temporary estimation results, and sets the temporary feature values. Each of the above is compared with a reference feature value, which is obtained when the frame type of the video stream is legitimately determined, and indicates a characteristic characteristic of the video stream, and an optimum adjustment coefficient is calculated based on the comparison result. It is what you do.

  Note that the optimum value of the adjustment coefficient, that is, the optimum adjustment coefficient, varies depending on conditions when the video stream is actually distributed, for example, encoder conditions such as the type of codec used for video communication and the encoding bit rate. Accordingly, an MPEG standard video stream is prepared in advance for each different encoder condition, and the optimum adjustment coefficient corresponding to each encoder condition is calculated by the adjustment coefficient calculation device 10. At this time, the input video stream may be encrypted, but the optimum adjustment coefficient can be calculated even if it is a non-encrypted video stream. Note that an input video stream uses an undegraded stream without packet loss.

[Frame type estimation device]
Next, the configuration of the adjustment coefficient calculation apparatus 10 according to the present embodiment will be described in detail with reference to FIG.
The adjustment coefficient calculation apparatus 10 includes a communication interface unit (hereinafter referred to as a communication I / F unit) 11, an operation input unit 12, a screen display unit 13, a storage unit 14, and an arithmetic processing unit 15 as main functional units. It has been. In addition, the arithmetic processing unit 15 includes a frame type estimation unit 15A, a feature value calculation unit 15B, and an optimum adjustment coefficient calculation unit 15C as main processing units.

The communication I / F unit 11 includes a dedicated communication circuit, and has a function of transmitting and receiving various IP packets such as TS packets of a video stream via a packet communication network or broadcast radio waves.
The operation input unit 12 includes an operation input device such as a keyboard and a mouse, and has a function of detecting an operator operation and outputting the operation to the arithmetic processing unit 15.
The screen display unit 13 includes a screen display device such as an LCD or a PDP, and has a function of displaying various information such as an operation menu and a video quality estimation result on the screen in response to an instruction from the arithmetic processing unit 15.

The storage unit 14 includes a storage device such as a hard disk and a memory, and has a function of storing processing information and programs 14P used for various types of processing in the arithmetic processing unit 15. The program 14P is read in advance from an external device or a recording medium (not shown) via a data input / output function such as the communication I / F unit 11 provided in the adjustment coefficient calculation device 10, and is stored in the storage unit 14. ing.
Main processing information stored in the storage unit 14 includes video stream configuration information 14A, a reference feature value 14B, and an optimum adjustment coefficient 14C.

The video stream configuration information 14A is configuration information indicating the GOP structure in the video stream, such as GOP information, and the content corresponding to the target video communication is input in advance via the operation input unit 12 or the communication I / F unit 11. And stored in the storage unit 14.
The reference feature value 14B is a feature value indicating a feature unique to the video stream, which is obtained when the frame type of the video stream is properly determined, and is calculated in advance and stored in the storage unit 14.

  The optimum adjustment coefficient 14C is an optimum value of the adjustment coefficient. The adjustment coefficient is a coefficient for adjusting the average value when calculating the determination threshold value used for determining the frame type from the average value of the in-frame data amount of each target frame to be estimated for the frame type. is there. The optimum adjustment coefficient 14C is selected by the optimum adjustment coefficient calculation unit 15C from the candidate temporary adjustment coefficients and stored in the storage unit 14.

  The arithmetic processing unit 15 has a microprocessor such as a CPU and its peripheral circuits, and reads the program 14P of the storage unit 14 and executes it by the microprocessor, thereby causing the hardware and the program 14P to cooperate with each other, It has a function of realizing various processing units such as a type estimation unit 15A, a feature value calculation unit 15B, and an optimum adjustment coefficient calculation unit 15C.

  The frame type estimation unit 15A has a function of calculating the amount of intra-frame data of each video frame constituting the video stream, and an average value of the intra-frame data amount of the target frame to be estimated for the frame type among these video frames. For each target frame, a function for calculating a determination threshold value calculated by adjusting the average value by the temporary adjustment coefficient a by multiplying the average value by a temporary adjustment coefficient a described later, It has a function of estimating the frame type of these target frames, that is, I, P, and B frames, by comparing the amount of data in the frame with a determination threshold value.

  For each of the plurality of temporary adjustment coefficients a that are candidates for the optimum adjustment coefficient A (14C), the feature value calculation unit 15B uses the temporary adjustment coefficient a for all the video frames constituting the video stream by the frame type estimation unit 15A. It has a function of estimating the frame type and a function of calculating a temporary feature value Ra indicating a characteristic unique to the video stream from the obtained temporary estimation result. As a specific example of the temporary feature value Ra, a ratio of B frames included in the video stream and a video quality characteristic of the video stream, which will be described later, may be used.

  The optimum adjustment coefficient calculation unit 15C compares each of the temporary feature values Ra obtained for each temporary adjustment coefficient a with the reference feature value Rs (14C) in the storage unit 14, and based on the comparison result, the optimum adjustment coefficient A (14C) is calculated and stored in the storage unit 14.

[Operation of First Embodiment]
Next, as the operation of the adjustment coefficient calculation apparatus 10 according to the present embodiment, the optimum value of the adjustment coefficient for estimating the frame type is calculated based on the MPEG standard video stream corresponding to an arbitrary encoder condition prepared in advance. The case of calculating will be described.
The operation of the adjustment coefficient calculation apparatus 10 includes a frame type process for estimating the frame type of the target frame included in the input video stream using the temporary adjustment coefficient a, and a frame type obtained for each temporary adjustment coefficient a. And an optimum adjustment coefficient calculation process for calculating an optimum adjustment coefficient based on the estimation result.

[Frame type estimation process]
First, with reference to FIG. 2 and FIG. 3, the frame type estimation process in the adjustment coefficient calculation apparatus 10 according to the present embodiment will be described. FIG. 2 is a flowchart showing the frame type estimation process. FIG. 3 is an explanatory diagram showing the frame type estimation process.

  This frame type estimation process is the same as that of Patent Document 2 described above, and here, the most basic frame type estimation process will be described. Patent Document 2 describes another frame type estimation process different from that in FIG. 2, and is specified by the frame type estimation unit 15A when calculating an adjustment coefficient used in the other frame type estimation process. For each temporary adjustment coefficient, the same process as the other frame type estimation process may be executed.

  Here, before executing the frame type estimation process, the video frame of the video stream divided and stored in the payload of each TS packet is extracted in advance at the start of processing, and the amount of data in each video frame is determined in advance. It is assumed that it has been calculated. The video frame extraction process and the intra-frame data amount calculation process may be executed by the frame type estimation unit 15A or a dedicated function unit and stored in the storage unit 14 at the time when the video stream is input, for example. .

  A video frame can be extracted using PUSI (payload_unit_start_indicator) or MB (MarkerBit) in an unencrypted TS header. The PUSI is a flag indicating whether or not a PES (Packetized Elementary Stream) header is present in the payload of the TS packet. The PES header indicates the data delimiter position regarding the stream in the payload. In the case of a video stream, video data, that is, the start point of each video frame is indicated. MB is a flag indicating the start point of video / audio data.

  The frame type estimation process in FIG. 2 is executed for each GOP structure. First, the frame type estimation unit 15A selects video frames for the maximum GOP length defined by the video stream configuration information 14A in the storage unit 14 from the video frames constituting the video stream (step 100). Of the selected frames, the frame with the largest amount of data corresponding to video data is estimated as an I frame (step 101). Normally, one I frame is included for each GOP, and the amount of data in the frame is larger than that of P and B frames.

Thereafter, the frame type estimation unit 15A selects, from among the GOPs starting from the I frame, frames subsequent to the I frame, that is, P and B frames for which the frame type is to be estimated (step 102).
Next, the frame type estimation unit 15A calculates the average value of the in-frame data amount of these target frames (step 103), and determines by multiplying the average value by the designated temporary adjustment coefficient a. A threshold value is calculated (step 104).

Subsequently, the frame type estimation unit 15A selects one unprocessed video frame from the target frames (step 105), and compares the in-frame data amount of this selected frame with a determination threshold value (step 106). .
If the intra-frame data amount is larger than the determination threshold (step 106: YES), the frame type estimation unit 15A estimates the selected frame as a P frame (step 107). When the amount of data in the frame is equal to or less than the determination threshold (step 106: NO), the frame type estimation unit 15A estimates the selected frame as a B frame (step 108).

After estimating the frame type of the selected frame in this way, the frame type estimation unit 15A checks whether there is an unprocessed target frame (step 109).
Here, when there is an unprocessed target frame (step 109: YES), the process returns to step 105 to execute processing relating to a new target frame, and when there is no unprocessed target frame (step 109: NO), A series of frame type estimation processing ends.

In FIG. 3, the horizontal axis indicates the frame number of the video frame, and the vertical axis indicates the amount of data in the frame of each video frame. Here, the maximum GOP length is 15 frames, and the GOP configuration is “IBBPBBPBBPBBPBB”.
First, of the 15 frames from the frame (1) to the frame (15), the frame (3) having the largest amount of data in the frame is estimated as the I frame. Next, out of 1 GOP from I frame (3) to frame (17), from frame (4) to frame (17) is selected as a target frame, and the average value of these intra-frame data amounts is calculated. A determination threshold value is calculated by multiplying the average value by the designated temporary adjustment coefficient a.

  Thereafter, the in-frame data amount of these target frames is compared with a determination threshold value, and the frame type of each target frame is estimated according to the comparison result. Here, since the amount of data in the frames (6), (9), (12), and (15) is larger than the determination threshold, it is estimated as a P frame. Also, the frame (4), the frame (5), the frame (7), the frame (8), the frame (10), the frame (11), the frame (13), the frame (14), the frame (16), and the frame (17 ) Is inferior to the determination threshold value, it is estimated as B frame.

[Optimum adjustment coefficient calculation process]
Next, with reference to FIG. 4, the optimum adjustment coefficient calculation process in the adjustment coefficient calculation apparatus 10 according to the present embodiment will be described. FIG. 4 is a flowchart showing an optimum adjustment coefficient calculation process according to the first embodiment.

  Before executing the optimal adjustment coefficient calculation process, as with the frame type estimation process, before executing the frame type estimation process, the video frame of the video stream divided and stored in the payload of each TS packet is extracted in advance at the start of the process. It is assumed that the amount of data in each video frame is calculated in advance.

  First, the feature value calculation unit 15B selects an unprocessed temporary adjustment coefficient a from among a plurality of temporary adjustment coefficients that are candidates for the optimum value of the adjustment coefficient (step 150). Note that the range of the temporary adjustment coefficient may be determined in advance based on, for example, past results.

  Next, the feature value calculation unit 15B configures the video stream by executing the frame type estimation process of FIG. 2 for an arbitrary encoder condition in the frame type estimation unit 15A using the selected temporary adjustment coefficient a. The frame type of all the video frames is estimated (step 151), and based on the obtained temporary estimation result, the temporary feature value Ra for evaluating the estimation result in the frame type estimation unit 15A using the temporary adjustment coefficient a. Is calculated (step 152).

  Subsequently, the optimum adjustment coefficient calculation unit 15C compares the temporary feature value Ra with the reference feature value Rs (14B) of the storage unit 14 (step 153), and based on these comparison results, optimizes the temporary feature value Ra from the temporary feature value Ra. The adjustment coefficient A (14C) is calculated and stored in the storage unit 14 (step 154).

As a specific example of this feature value, a frame ratio indicating a ratio of B frames included in a video stream, which will be described later, or a video quality characteristic of the video stream may be used.
For example, when the frame ratio is used as the feature value, the comparison result between the temporary frame ratio that is the temporary feature value Ra and the reference frame ratio that is the reference feature value Rs (14B), that is, the temporary adjustment coefficient a having the smallest difference. May be selected as the optimum adjustment coefficient A for any encoder condition. This temporary frame ratio matches the reference frame ratio calculated from the GOP structure if the temporary estimation result is valid.

  When the video quality characteristic is used as the feature value, the comparison result between the temporary video quality characteristic that is the temporary feature value Ra and the reference video quality characteristic that is the reference feature value Rs (14B), that is, the correlation coefficient is The optimal adjustment coefficient A may be calculated from the temporary adjustment coefficient a included in the predetermined correlation threshold range. The video quality characteristic is a model showing the relationship between the invalid frame rate and the video quality value. If the temporary estimation result is valid, the invalid frame rate of the video stream can be obtained with high accuracy, and as a result, the invalid frame rate is obtained. The video quality characteristic derived from the rate has a strong correlation with the video quality characteristic obtained by evaluating the actual video stream in advance.

  Thereafter, the optimal adjustment coefficient calculation unit 15C confirms whether or not the unprocessed temporary adjustment coefficient a exists (step 155). If the unprocessed temporary adjustment coefficient a exists (step 155: YES), the process proceeds to step 150. The process related to the new temporary adjustment coefficient a is returned and if there is no unprocessed temporary adjustment coefficient a (step 155: NO), the series of optimum adjustment coefficient calculation processes is terminated.

[Effect of the first embodiment]
Thus, in the present embodiment, the feature value calculation unit 15B uses the temporary adjustment coefficient a for each of the plurality of temporary adjustment coefficients a that are candidates for the optimal adjustment coefficient, and the frame type estimation unit 15A uses the temporary adjustment coefficient a. A frame type is estimated, a temporary feature value Ra indicating a characteristic unique to the video stream is calculated from the obtained temporary estimation result, and each of the temporary feature values Ra is converted into a reference feature value Rs ( 14B), and the optimum adjustment coefficient A is calculated based on these comparison results.

  Thereby, the estimation result of the frame type when each temporary adjustment coefficient a is used can be evaluated without using the frame type of the video stream used for calculating the optimum adjustment coefficient. Therefore, even in a video stream in which the payload portion of the TS packet is encrypted and the video data itself cannot be processed, the adjustment coefficient used for identifying the frame type can be automatically calculated.

[Second Embodiment]
Next, a video frame type estimation adjustment coefficient calculation apparatus 10 according to a second embodiment of the present invention will be described with reference to FIG. FIG. 5 is a block diagram illustrating a configuration of an adjustment coefficient calculation apparatus for video frame type estimation according to the second embodiment.
In the present embodiment, a case where a frame ratio is used as a feature value in the first embodiment will be described.

In FIG. 5, the feature value calculation unit 15B is provided with a reference frame ratio calculation unit 16A and a temporary frame ratio calculation unit 16B as main processing units.
The reference frame ratio calculation unit 16A has a function of calculating a reference frame ratio Rs (14B) indicating the ratio of the B frames included in the video stream based on the video stream configuration information 14A in the storage unit 14 and storing it in the storage unit 14. Have.

  The temporary frame ratio calculation unit 16B, for each of a plurality of temporary adjustment coefficients a that are candidates for the optimal adjustment coefficient A, based on the temporary estimation result obtained by the frame type estimation unit 15A, for the B frames included in the video stream. And a function of calculating a provisional frame ratio Ra indicating the ratio.

[Operation of Second Embodiment]
Next, with reference to FIGS. 6 to 8, the optimum adjustment coefficient calculation processing in the video frame type estimation adjustment coefficient calculation apparatus 10 according to the present embodiment will be described. FIG. 6 is a flowchart illustrating an optimum adjustment coefficient calculation process according to the second embodiment. FIG. 7 is a pseudo code example of the optimum adjustment coefficient calculation process according to the second embodiment. FIG. 8 is an explanatory diagram of an optimum adjustment coefficient calculation process according to the second embodiment.

  Before executing the optimal adjustment coefficient calculation process, as with the frame type estimation process, before executing the frame type estimation process, the video frame of the video stream divided and stored in the payload of each TS packet is extracted in advance at the start of the process. It is assumed that the amount of data in each video frame is calculated in advance.

  Further, the reference frame ratio calculation unit 16A refers to the video stream configuration information 14A in the storage unit 14, and based on the number of frames Nb in which B frames continue between I and P frames in the GOP to be estimated, I , The reference frame ratio Rs = 1 / Nb indicating the ratio of the P frame and the B frame is calculated and stored in the storage unit 14. At this time, the number of frames Nb is equal to a value obtained by subtracting 1 from the period M of P frames appearing at regular intervals after the I frame. Normally, in the GOP structure, it is defined as a period M of P frames.

  First, the feature value calculation unit 15B selects an unprocessed temporary adjustment coefficient a from among a plurality of temporary adjustment coefficients that are candidates for the optimum value of the adjustment coefficient (step 200). Note that the range of the temporary adjustment coefficient may be determined in advance based on, for example, past results.

  Next, the feature value calculation unit 15B configures the video stream by executing the frame type estimation process of FIG. 2 for an arbitrary encoder condition in the frame type estimation unit 15A using the selected temporary adjustment coefficient a. As the temporary feature value for estimating the frame type of all video frames (step 201) and evaluating the estimation result in the frame type estimation unit 15A using the temporary adjustment coefficient a based on the obtained temporary estimation result. A temporary frame ratio Ra is calculated (step 202). This step 202 includes the following three steps 210 to 212.

  First, the temporary frame ratio calculation unit 16B counts the number of B frames nb from the temporary estimation result obtained from the frame type estimation unit 15A (step 211), and counts the number of I and P frames n (step 211). A temporary frame ratio Ra = n / nb indicating the ratio of the I, P frame and the B frame is calculated from the B frame number nb and the I and P frame numbers n (step 212).

Subsequently, the optimal adjustment coefficient calculation unit 15C compares the temporary frame ratio Ra, which is a temporary feature value, with the reference frame ratio Rs (14B), which is a reference feature value of the storage unit 14 (step 203). This step 203 is composed of the following two steps 220 to 221.
First, the optimal adjustment coefficient calculation unit 15C calculates a new difference ΔR (= | Ra−Rs |) from the temporary frame ratio Ra and the reference frame ratio Rs (14B) (step 220), and the temporary adjustment selected so far. The difference obtained with the adjustment coefficient a is compared (step 221).

  Here, when the new difference ΔR is the smallest so far (step 221: YES), the optimum adjustment coefficient calculation unit 15C uses the temporary adjustment coefficient a as the optimum adjustment coefficient A (14C) for an arbitrary encoder condition. Is stored in the storage unit 14 (step 204).

Thereafter, the optimum adjustment coefficient calculation unit 15C confirms whether or not an unprocessed temporary adjustment coefficient a exists (step 205). If an unprocessed temporary adjustment coefficient a exists (step 205: YES), the process proceeds to step 200. The process related to the new temporary adjustment coefficient a is executed again, and if there is no unprocessed temporary adjustment coefficient a (step 205: NO), the series of optimum adjustment coefficient calculation processes is terminated.
In step 221, when the new difference ΔR has not been minimized so far (step 221: NO), the optimal adjustment coefficient calculation unit 15C proceeds to step 205.

  FIG. 8 shows a change in the difference ΔR when the temporary adjustment coefficient a is set in increments of 0.1 in the range from 0.5 to 2.0. When the temporary adjustment coefficient a is selected in order from 0.5, the difference ΔR monotonously decreases, and when the temporary adjustment coefficient a = 1.2, the difference ΔR becomes the minimum value, and thereafter the difference ΔR monotonously increases. . For this reason, 1.2 when the difference ΔR is the minimum value is selected as the optimum adjustment coefficient A.

[Effect of the second embodiment]
As described above, the present embodiment uses the temporary frame ratio indicating the ratio of the B frames included in the video stream calculated based on the temporary estimation result as the temporary feature value Ra, and the video as the reference feature value Rs. Using the reference frame ratio indicating the ratio of B frames included in the video stream calculated based on the video stream configuration information 14A of the stream, the optimal adjustment coefficient calculation unit 15C uses the temporary frame ratio Ra of the temporary adjustment coefficient a. And the reference frame ratio Rs is selected as the optimum adjustment coefficient A, so that the optimum adjustment coefficient A can be calculated very accurately with a feature value that can be calculated by simple arithmetic processing.

  In particular, the reference frame ratio calculation unit 16A and the temporary frame ratio calculation unit 16B used in the video frame type estimation adjustment coefficient calculation apparatus 10 according to the present embodiment can be realized by simple arithmetic processing as described above. It can be easily implemented in a quality estimation device, a video frame estimation device, and the like.

  In the present embodiment, the case of using the ratio of I, P frames and B frames as the frame ratio has been described as an example. However, the present invention is not limited to this. For example, it may be the ratio of the number of B frames to the total number of video frames constituting the video stream, or any other frame type configuration ratio for the video stream, such as the ratio of B frames to I frames or P frames. It may be a ratio.

[Third Embodiment]
Next, a video frame type estimation adjustment coefficient calculation apparatus 10 according to a third embodiment of the present invention will be described with reference to FIG. FIG. 9 is a block diagram illustrating a configuration of a video frame type estimation adjustment coefficient calculation apparatus according to the third embodiment.
In the present embodiment, a case will be described in which video quality estimation characteristics are used as feature values in the first embodiment.

In FIG. 9, the feature value calculation unit 15B is provided with an invalid frame rate calculation unit 17A and a provisional video quality characteristic specifying unit 17B as main processing units.
The invalid frame rate calculation unit 17A determines, for each of a plurality of temporary adjustment coefficients a that are candidates for the optimal adjustment coefficient A, the video frames constituting the video stream based on the temporary estimation result obtained by the frame type estimation unit 15A. And a function of calculating an invalid frame rate indicating a lost ratio.

The temporary video quality characteristic specifying unit 17B has a function of specifying the temporary video quality characteristic Ra indicating the correspondence between the invalid frame rate and the video communication quality based on the invalid frame rate calculated by the invalid frame rate calculation unit 17A. ing.
As a method of calculating the invalid frame rate, a known technique such as Patent Document 3 may be used. As a method for specifying the video quality characteristic, a known technique such as Patent Document 4 may be used.

  In the present embodiment, the reference feature value 14B of the storage unit 14 is composed of reference video quality characteristics. The video quality is a subjective quality when a user views a video that is received and decoded by a receiving terminal. For reference video quality characteristics, at least evaluation videos with different video frame loss conditions are prepared, subjective video quality evaluation experiments are actually performed on these evaluation videos, and the invalid frame rate and video communication are obtained from the obtained evaluation results. The video quality characteristic indicating the correspondence with the quality is derived.

At this time, in the subjective video quality evaluation experiment, for the evaluation video, a video condition indicating a difference in video content may be combined with a loss condition of the video frame. In addition, since the optimum adjustment coefficient A varies depending on the encoder conditions, it is actually necessary to prepare a combination of encoder conditions as an evaluation video.
In addition, regarding the evaluation method of video quality, ITU-T recommendation P.21. The ACR (5-level scale quality scale) defined in 910 may be used, and MOS may be used as the subjective quality feature value.

[Operation of Third Embodiment]
Next, with reference to FIGS. 10 to 13, the optimum adjustment coefficient calculation processing in the video frame type estimation adjustment coefficient calculation apparatus 10 according to the present embodiment will be described. FIG. 10 is a flowchart illustrating an optimum adjustment coefficient calculation process according to the third embodiment. FIG. 11 is a pseudo code example of the optimum adjustment coefficient calculation process according to the third embodiment. FIG. 12 is a calculation example of provisional video quality characteristics according to the third embodiment. FIG. 13 is an explanatory diagram of an optimal adjustment coefficient calculation process according to the third embodiment.

Before executing the optimal adjustment coefficient calculation process, as with the frame type estimation process, before executing the frame type estimation process, the video frame of the video stream divided and stored in the payload of each TS packet is extracted in advance at the start of the process. It is assumed that the amount of data in each video frame is calculated in advance.
Further, it is assumed that the reference video quality characteristic Rs (14B) obtained in the subjective video quality evaluation experiment described above is stored in the storage unit 14 in advance.

  First, the feature value calculation unit 15B selects an unprocessed temporary adjustment coefficient a from among a plurality of temporary adjustment coefficients that are candidates for the optimum value of the adjustment coefficient (step 300). Note that the range of the temporary adjustment coefficient may be determined in advance based on, for example, past results.

  Next, the feature value calculation unit 15B uses the selected temporary adjustment coefficient a for the frame type estimation unit 15A for each video stream having different loss conditions for at least video frames prepared in advance for an arbitrary encoder condition. Frame type estimation processing is executed to estimate the frame types of all video frames constituting the video stream (step 301), and based on the obtained temporary estimation results, the frame types using the temporary adjustment coefficient a A temporary video quality characteristic Ra is calculated as a temporary feature value for evaluating the estimation result in the estimation unit 15A (step 302). This step 302 includes the following two steps 310 to 311.

  First, the invalid frame rate calculation unit 17A indicates the ratio of loss of the video frames constituting the video stream based on the temporary estimation result for each loss condition obtained by the frame type estimation unit 15A for the adjustment coefficient a. A frame rate is calculated (step 310).

Next, the provisional video quality characteristic specifying unit 17B, based on the invalid frame rate for each loss condition calculated by the invalid frame rate calculation unit 17A, indicates the provisional video quality characteristic indicating the correspondence between the invalid frame rate and the video communication quality. Ra is specified (step 311).
FIG. 12 shows an example in which video quality values (MOS) obtained by evaluating the quality of an evaluation video stream at different invalid frame rates are plotted, and a regression equation of these plots is derived as a temporary video quality characteristic Ra. Has been.

  Subsequently, the optimum adjustment coefficient calculation unit 15C compares the temporary video quality characteristic Ra, which is a temporary feature value, with the reference video quality characteristic Rs (14B), which is a reference feature value of the storage unit 14, to thereby calculate the temporary video quality characteristic Ra. Is calculated and stored in the storage unit 14 (step 303).

  Thereafter, the optimal adjustment coefficient calculation unit 15C checks whether there is an unprocessed temporary adjustment coefficient a (step 304). If there is an unprocessed temporary adjustment coefficient a (step 304: YES), the process proceeds to step 300. It returns and performs the process regarding the new temporary adjustment coefficient a.

On the other hand, when there is no unprocessed temporary adjustment coefficient a (step 304: NO), the optimal adjustment coefficient calculation unit 15C selects the maximum correlation coefficient Rmax from the correlation coefficients of each temporary adjustment coefficient a, and this The correlation threshold value Rth is calculated by multiplying the maximum correlation coefficient by the threshold value calculation coefficient (step 305). Among the temporary adjustment coefficients a, the correlation coefficient is from this correlation threshold value Rth to the maximum correlation coefficient Rmax. The temporary adjustment coefficient a included in the correlation threshold range Rw is selected (step 306).
Then, the optimum adjustment coefficient calculation unit 15C calculates the representative value of the selected temporary adjustment coefficient a as the optimum adjustment coefficient A and stores it in the storage unit 14 (step 307), and ends the series of optimum adjustment coefficient calculation processes.

  FIG. 13 shows the change of the correlation coefficient R when the temporary adjustment coefficient a is set in increments of 0.2 in the range from 0.1 to 1.9. When the temporary adjustment coefficient a is selected in order from 0.1, the correlation coefficient R increases and changes within a certain range, and thereafter the correlation coefficient R decreases. Here, when the threshold value calculation coefficient K = 0.005, since the maximum correlation coefficient Rmax = 0.931, the correlation threshold value Rth = Rmax × (1−K) = 0.926, and the correlation The threshold range Rw is 0.926 to 0.931.

  Therefore, the temporary adjustment coefficients ax to ay are selected as the temporary adjustment coefficient a in which the correlation coefficient R falls within the correlation threshold range Rw, and the median value, average value, maximum value, etc. of these temporary adjustment coefficients ax to ay are selected. Is calculated as the optimum adjustment coefficient A by statistical processing. In this example, an average value 0.9 of the temporary adjustment coefficients ax and ay is calculated as the optimum adjustment coefficient A.

[Effect of the third embodiment]
As described above, according to the present embodiment, as the temporary feature value Ra, at least for each of various video communication conditions, the temporary estimation result obtained from the video stream distributed under the video communication conditions and the frame loss status of each video frame The temporary video quality characteristic indicating the relationship between the invalid frame rate and the video quality value specified by the invalid frame rate calculated from the above is used, and the video communication condition is set as the reference feature value Rs at least for each video communication condition. The reference video quality characteristic indicating the relationship between the invalid frame rate and the video quality value related to the video stream, obtained by evaluating the video stream distributed below in advance, is used.

Then, in the optimum adjustment coefficient calculation unit 15C, the correlation coefficient between the temporary video quality characteristic and the reference video quality characteristic among the temporary adjustment coefficient a is included in the range from the maximum value to a predetermined correlation threshold value. Since these are selected and the representative values of these temporary adjustment coefficients are calculated as the optimum adjustment coefficients, the optimum adjustment coefficient A can be calculated very accurately.
In particular, the invalid frame rate calculation unit 17A and the provisional video quality characteristic specifying unit 17B used in the video frame type estimation adjustment coefficient calculation device 10 according to the present embodiment are those provided in advance in the video quality estimation device. It can also be used, and can be implemented in a video quality estimation apparatus with the addition of a few functional units.

[Extended embodiment]
The present invention has been described above with reference to the embodiments, but the present invention is not limited to the above embodiments. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.
For example, network quality values such as a packet loss rate and the number of bucket losses may be used instead of the invalid frame rate. The third embodiment can also be applied to the adjustment coefficient B in Non-Patent Document 2. Also, in the description of each embodiment, a compression encoding method compliant with MPEG standards such as MPEG-2 standard (ISO / IEC 13818) and MPEG-4 standard (ISO / IEC 14496, ITU-T H.264) is used. The video frame used has been described as an example. However, the present invention is not limited to this, and each embodiment is also applied to a video frame derived from the MPEG standard or using another compression coding method similar to the MPEG standard. Can be applied.

  DESCRIPTION OF SYMBOLS 10 ... Adjustment coefficient calculation apparatus for frame classification estimation, 11 ... Communication I / F part, 12 ... Operation input part, 13 ... Screen display part, 14 ... Memory | storage part, 14A ... Video stream structure information, 14B ... Reference | standard feature value, 14C ... optimal adjustment coefficient, 14P ... program, 15 ... arithmetic processing section, 15A ... frame type estimation section, 15B ... feature value calculation section, 15C ... optimal adjustment coefficient calculation section, 16A ... reference frame ratio calculation section, 16B ... temporary frame ratio Calculation unit, 17A... Invalid frame rate calculation unit, 17B... Temporary video quality characteristic specifying unit.

Claims (5)

Of each video frame constituting the video stream, the in-frame data amount of each target frame that is the target of estimation of the frame type was calculated by adjusting the average value of the in-frame data amount of these target frames with a predetermined adjustment coefficient. Video frame type estimation adjustment coefficient calculation for calculating an optimum adjustment coefficient that is an optimum value of the adjustment coefficient for a video frame type estimation device that estimates the frame type of these target frames by comparing with a determination threshold value An adjustment coefficient calculation method for video frame type estimation used in an apparatus,
The frame type estimation unit has designated the intra-frame data amount of each target frame that is the target of frame type out of the respective video frames constituting the video stream, and the average value of the intra-frame data amount of these target frames. A frame type estimation step for estimating the frame type of these target frames by comparing with a determination threshold value calculated by adjusting with a temporary adjustment coefficient;
For each of the plurality of temporary adjustment coefficients that are candidates for the optimum adjustment coefficient, the feature value calculation unit estimates the frame type of the target frame by the frame type estimation unit using the temporary adjustment coefficient, and obtained temporary estimation A feature value calculating step for calculating a temporary feature value indicating a characteristic unique to the video stream from the result;
A storage step of storing in advance, as a reference feature value, a characteristic value indicating a characteristic unique to the video stream, obtained when the frame type of the video stream is legitimately determined;
An optimum adjustment coefficient calculating unit comprising: an optimum adjustment coefficient calculating step of comparing each of the temporary feature values with the reference feature value and calculating the optimum adjustment coefficient based on the comparison result ;
The temporary feature value includes a temporary frame ratio that is calculated based on the temporary estimation result and indicates a composition ratio of an arbitrary frame type related to the video stream.
The reference feature value includes a reference frame ratio indicating a configuration ratio of an arbitrary frame type related to the video stream, calculated based on video stream configuration information of the video stream.
The optimal adjustment coefficient calculating step includes a step of selecting, as the optimal adjustment coefficient, the smallest difference between the temporary frame ratio and the reference frame ratio among the temporary adjustment coefficients. Adjustment coefficient calculation method for estimation. Of each video frame constituting the video stream, the in-frame data amount of each target frame that is the target of estimation of the frame type was calculated by adjusting the average value of the in-frame data amount of these target frames with a predetermined adjustment coefficient. Video frame type estimation adjustment coefficient calculation for calculating an optimum adjustment coefficient that is an optimum value of the adjustment coefficient for a video frame type estimation device that estimates the frame type of these target frames by comparing with a determination threshold value An adjustment coefficient calculation method for video frame type estimation used in an apparatus,
The frame type estimation unit has designated the intra-frame data amount of each target frame that is the target of frame type out of the respective video frames constituting the video stream, and the average value of the intra-frame data amount of these target frames. A frame type estimation step for estimating the frame type of these target frames by comparing with a determination threshold value calculated by adjusting with a temporary adjustment coefficient;
For each of the plurality of temporary adjustment coefficients that are candidates for the optimum adjustment coefficient, the feature value calculation unit estimates the frame type of the target frame by the frame type estimation unit using the temporary adjustment coefficient, and obtained temporary estimation A feature value calculating step for calculating a temporary feature value indicating a characteristic unique to the video stream from the result;
A storage step of storing in advance, as a reference feature value, a characteristic value indicating a characteristic unique to the video stream, obtained when the frame type of the video stream is legitimately determined;
An optimum adjustment coefficient calculating unit comprising: an optimum adjustment coefficient calculating step of comparing each of the temporary feature values with the reference feature value and calculating the optimum adjustment coefficient based on the comparison result;
The tentative feature value is an invalid frame rate calculated from the tentative estimation result and the frame loss status of each video frame obtained from the video stream distributed under the video communication condition at least for each of various video communication conditions. Consists of temporary video quality characteristics that indicate the relationship between invalid frame rate and video quality value,
The reference feature value indicates a relationship between an invalid frame rate and a video quality value relating to the video stream obtained by evaluating the video stream distributed under the video communication condition in advance for at least each video communication condition. It consists of standard video quality characteristics,
In the optimum adjustment coefficient calculating step, a correlation coefficient between the temporary video quality characteristic and the reference video quality characteristic is included in a range from the maximum value to a predetermined correlation threshold among the temporary adjustment coefficients. And calculating a representative value of these temporary adjustment coefficients as the optimum adjustment coefficient. A method of calculating an adjustment coefficient for video frame type estimation. Of each video frame constituting the video stream, the in-frame data amount of each target frame that is the target of estimation of the frame type was calculated by adjusting the average value of the in-frame data amount of these target frames with a predetermined adjustment coefficient. Video frame type estimation adjustment coefficient calculation for calculating an optimum adjustment coefficient that is an optimum value of the adjustment coefficient for a video frame type estimation device that estimates the frame type of these target frames by comparing with a determination threshold value A device,
Among the video frames that make up the video stream, the amount of data in each frame that is subject to estimation of the frame type is adjusted, and the average value of the data in the frame of these target frames is adjusted using the specified temporary adjustment coefficient. A frame type estimation unit for estimating the frame type of these target frames by comparing with the determination threshold value calculated by
For each of the plurality of temporary adjustment coefficients that are candidates for the optimum adjustment coefficient, the frame type estimation unit estimates the frame type of the target frame using the temporary adjustment coefficient, and the obtained temporary estimation result is used for the video stream. A feature value calculation unit for calculating a temporary feature value indicating a unique feature;
A storage unit that preliminarily stores, as a reference feature value, a feature value indicating a characteristic unique to the video stream, which is obtained when the frame type of the video stream is legitimately determined;
An optimum adjustment coefficient calculating unit that compares each of the temporary feature values with the reference feature value and calculates the optimum adjustment coefficient based on the comparison results ;
The temporary feature value includes a temporary frame ratio that is calculated based on the temporary estimation result and indicates a composition ratio of an arbitrary frame type related to the video stream.
The reference feature value includes a reference frame ratio indicating a configuration ratio of an arbitrary frame type related to the video stream, calculated based on video stream configuration information of the video stream.
The optimum adjustment coefficient calculating unit selects, as the optimum adjustment coefficient, an adjustment for video frame type estimation that has the smallest difference between the temporary frame ratio and the reference frame ratio among the temporary adjustment coefficients. Coefficient calculation device. Of each video frame constituting the video stream, the in-frame data amount of each target frame that is the target of estimation of the frame type was calculated by adjusting the average value of the in-frame data amount of these target frames with a predetermined adjustment coefficient. Video frame type estimation adjustment coefficient calculation for calculating an optimum adjustment coefficient that is an optimum value of the adjustment coefficient for a video frame type estimation device that estimates the frame type of these target frames by comparing with a determination threshold value A device,
Among the video frames that make up the video stream, the amount of data in each frame that is subject to estimation of the frame type is adjusted, and the average value of the data in the frame of these target frames is adjusted using the specified temporary adjustment coefficient. A frame type estimation unit for estimating the frame type of these target frames by comparing with the determination threshold value calculated by
For each of the plurality of temporary adjustment coefficients that are candidates for the optimum adjustment coefficient, the frame type estimation unit estimates the frame type of the target frame using the temporary adjustment coefficient, and the obtained temporary estimation result is used for the video stream. A feature value calculation unit for calculating a temporary feature value indicating a unique feature;
A storage unit that preliminarily stores, as a reference feature value, a feature value indicating a characteristic unique to the video stream, which is obtained when the frame type of the video stream is legitimately determined;
An optimum adjustment coefficient calculating unit that compares each of the temporary feature values with the reference feature value and calculates the optimum adjustment coefficient based on the comparison results;
The tentative feature value is an invalid frame rate calculated from the tentative estimation result and the frame loss status of each video frame obtained from the video stream distributed under the video communication condition at least for each of various video communication conditions. Consists of temporary video quality characteristics that indicate the relationship between invalid frame rate and video quality value,
The reference feature value indicates a relationship between an invalid frame rate and a video quality value relating to the video stream obtained by evaluating the video stream distributed under the video communication condition in advance for at least each video communication condition. It consists of standard video quality characteristics,
The optimum adjustment coefficient calculating unit includes a correlation coefficient between the temporary video quality characteristic and the reference video quality characteristic in the range from the maximum value to a predetermined correlation threshold value among the temporary adjustment coefficients. And a representative value of these temporary adjustment coefficients is calculated as the optimum adjustment coefficient. An adjustment coefficient calculation apparatus for estimating video frame type. The program for functioning a computer as each part which comprises the adjustment coefficient calculation apparatus for video frame classification estimation of Claim 3 or Claim 4 .

Images