CN100591145C - Video quality evaluation device and video quality evaluation method - Google Patents

Abstract

A subjective quality estimation unit (11) receives a reference video signal (RI) as a video signal not deteriorated and a deteriorated video signal (PI) as the reference video signal deteriorated, calculates a video signal characteristic amount of the both signals, and estimates the subjective quality of the deteriorated video signal according to the difference between the video signal characteristic amounts of the both signals calculated. A characteristic amount calculation unit (12) calculates a video signal characteristic amount of the reference video signal. A correction information storage unit (13) contains correction information for correcting the subjective quality in accordance with the video signal characteristic amount. A correction calculation unit (14) acquires correction information corresponding to the video signal characteristic amount of the reference video signal received from the characteristic amount calculation unit (12), from a correction information storage unit(13) and passes it to a correction unit (15). According to the correction information received, the correction unit (15) corrects the subjective quality of the deteriorated video signal estimated by the subjective quality estimation unit (11).

Description

Video quality evaluation device and video quality evaluation method

Technical field

The present invention relates to not to be after people see actual image, carry out the subjective quality evaluation test of quality evaluation, but infer video quality evaluation device, video quality evaluation method and the video quality evaluation program of subjective quality by the mensuration of the characteristic quantity of the physics of signal of video signal.

In addition, the invention still further relates in video quality evaluation device, video quality evaluation method and video quality evaluation program, carry out the reference images signal and by through coding or network transmission etc. and image coalignment, image matching method and the video matching program of the coupling of space between the deterioration signal of video signal of its deterioration and time location.

Background technology

Image information, by certain processing, for example coding or network transmission etc., generally its quality can worsen.About the image of such deterioration, the degree of the actual deterioration of seeing the back and feeling of people is called subjective quality.

Always, as the method for estimating this subjective quality, carry out actual person in charge's quality evaluation test of seeing image according to people.But,, exist and to take a lot of work and the problem of spended time in the quality evaluation of image for the actual quality of seeing this image of image post-evaluation of people.In addition, be proficient in the people and the beginner of video quality evaluation, estimate different mostly.

Therefore, as the method for measuring the subjective quality of inferring image by physics, the physical features amount of the deterioration signal of video signal of with good grounds reference images signal and deterioration poor, perhaps according to the deterioration amount of only trying to achieve, the method for the deterioration degree of evaluation quality objectively by the physical features amount that worsens signal of video signal.

But, the subjective quality of image, if qualification image, though also can infer accurately sometimes the subjective assessment value (for example with reference to ANSI T1.801.03.1996, " Digital Transport ofOne-Way Video Signal Parameters for Objective Performance Assessment "; The ridge this, high bridge " about the discussion of the law of application of quality of image objective evaluation technology ", it is big that letter is learned ソ, Sep.2002; Ridge basis, chestnut field, high bridge " about improving the properties Research of quality of image objective evaluation ", it is big that letter is learned ソ, Mar.2003), but for image arbitrarily, because greatly rely on the character of image, even so in the identical occasion of the degree of deterioration amount, as a result of person in charge's quality of Pan Duaning also has difference more.

From such reason, in the occasion of asking the image subjective quality, still be that actual this image of seeing of people was estimated the such subjective assessment test of its quality afterwards mostly, such technology status is continuing always.

In such situation, having proposed to make it can estimate subjective quality as the occasion of seeing near the people is purpose, be conceived to the fringe region of image, signal of video signal is used the Sobel filter, calculate the objective evaluation (with reference to PCT Pub.No.:WO99/45715) of its deterioration amount.

But, only use such characteristic quantity, can't obtain replacing the sufficient evaluation precision of actual subjective assessment.

On the other hand, in the method for the subjective quality of inferring image, the prerequisite of the physical features amount of the deterioration signal of video signal of reference images signal and deterioration needs coupling reference images signal and worsens the space of signal of video signal and the position of time as a comparison.That is, between reference images signal and deterioration signal of video signal, need departing from of direction match time or departing from of locus.

Originally, such matching treatment is with manually carrying out.But, soon, proposed to carry out automatically the scheme of matching treatment.For example, at United States Patent (USP) 5,446, in 492,, also can tackle it in the occasion that the time has taken place to postpone during matching treatment as the pre-treatment of objective evaluation.

According to the technology of the document, as television broadcasting, if to as if with size with the signal of frame rate, then in case obtain synchronously, just can obtain later good synchronous.

But, in the image distribution or communication service that begins in recent years to popularize, because in personal computer (PC) terminal, receive image, so the size or the length-width ratio of the image that receives are various via the IP network of internet etc.In addition, also lose the information of as packet, concentrating to a certain degree sometimes.Therefore, the position correspondence that is difficult to obtain the reference images signal and worsens the space of signal of video signal.On the other hand, because the generation of the fluctuation of IP bag arrival interval or bag loss, taken place to depart from or the fluctuation of image Displaying timer or freeze these non-existent phenomenons in the past such as the timing that the image on time shaft shows.Therefore, do not obtain the method for the coupling that departs from the time such between reference images signal and the deterioration signal of video signal.

In the method for the subjective quality of inferring image, if can not obtain the reference images signal fully and worsen the space of signal of video signal and the location matches of time, can not carry out the inferring of subjective quality of image the most at last, it is also low promptly to enable its precision.

In addition, among the PCT Pub.No.:WO99/45715 in front, document as an example of having recorded and narrated last matching treatment of time, can enumerate ITU-T Contribution COM-12-29 " Draft newrecommendation on multimedia communication delay; synchronization; and framerate measurement ", December 1997.

Summary of the invention

The present invention proposes in view of above-mentioned prior art, its objective is, provide can to any image high-precision infer video quality evaluation device, video quality evaluation method and the video quality evaluation program of unified subjective quality.

In addition, the objective of the invention is, be provided at when inferring benchmark signal of video signal in the subjective quality and worsening the physical features amount of signal of video signal, can obtain their space and image coalignment, image matching method and the video matching program of the location matches of time really.

For achieving the above object, the main idea of the invention of first form is a kind of video quality evaluation device, have: input is as the reference images signal of the signal of video signal that does not worsen and the deterioration signal of video signal after this reference images signal degradation, calculate the signal of video signal characteristic quantity of two signals, poor according to the signal of video signal characteristic quantity of two signals that calculate inferred the subjective quality of the subjective quality of described deterioration signal of video signal and infers portion; Be used to revise the update information storage part of the update information of described subjective quality with the storage that is mapped of signal of video signal characteristic quantity; Import the described signal of video signal characteristic quantity of described reference images signal, obtain and the corresponding update information of having imported of signal of video signal characteristic quantity from described update information storage part, according to the subjective quality correction portion of the described subjective quality of obtaining of inferring of update information correction.

The main idea of the invention of second form is, in the invention of first form, described subjective quality is inferred portion, according to the spatial information of the image state in each frame that is illustrated in the image that comprises in the signal of video signal and be illustrated in any one at least deterioration signal of video signal of the temporal information that the image of interframe changes and the difference between the reference images signal is inferred described subjective quality.

The main idea of the invention of the 3rd form is, in the invention of second form, described subjective quality is inferred portion, at least one side of mobile electric power amount (M) that poor as the described signal of video signal characteristic quantity between described deterioration signal of video signal and described reference images signal, computational chart are shown in the edge electric power amount (E) of the deterioration amount in each frame of the image that comprises in the signal of video signal and are illustrated in the deterioration amount of interframe.

The main idea of the invention of the 4th form is, in the invention of second form, described subjective quality is inferred portion, according to ITU-R Recommendation P.910 among the spatial information of regulation and the temporalinformation difference any one at least, that worsen between signal of video signal and reference images signal infer described subjective quality.

The main idea of the invention of the 5th form is, in the invention of first form, the update information storage part, as the update information that is used to revise described subjective quality, with the spatial information that is illustrated in the image state in each frame of image that comprises in the signal of video signal and be illustrated in temporal information that the image of interframe the changes storage correction factor that is mapped.

The main idea of the invention of the 6th form is, in the invention of first form, described subjective quality is inferred portion, have: import described reference images signal and described deterioration signal of video signal, generate about from the reference images frame of described reference images signal and departing from and the calibration information generating unit of the calibration information that departs from the space of described reference images frame and described deterioration image frame from time of the deterioration image frame of described deterioration signal of video signal; According to described calibration information, eliminating after described spatial offset and described time departs from, calculate the space characteristics amount calculating part of space characteristics amount separately according to the spatial information of the image state of described reference images frame of expression and described deterioration image frame; According to described calibration information, eliminating after described spatial offset and described time departs from, for described reference images frame and described deterioration image frame, calculate the temporal characteristics amount calculating part of temporal characteristics amount separately according to the temporal information of the image variation of representing interframe separately; Infer the portion of inferring of the subjective quality of described deterioration signal of video signal according to described space characteristics amount and described temporal characteristics amount.

In addition, for achieving the above object, the invention of the 7th form is a kind of video quality evaluation method, this method, be have with the signal of video signal characteristic quantity be mapped storage be used for revising as the deterioration signal of video signal after the reference images signal degradation of the signal of video signal that does not worsen be pushed the update information of the subjective quality of making the update information storage part, infer the video quality evaluation method of video quality evaluation device of the subjective quality of described deterioration signal of video signal; It is characterized in that, import described reference images signal and described deterioration signal of video signal, calculate the signal of video signal characteristic quantity of two signals, poor, the described subjective quality of inferring described deterioration signal of video signal according to the signal of video signal characteristic quantity of two signals of calculating, obtain the corresponding update information of signal of video signal characteristic quantity with the described reference images signal of calculating from described update information storage part, according to the described subjective quality of inferring of the update information correction that obtains.

In addition, for achieving the above object, the invention of the 8th form is a kind of video quality evaluation program, it is characterized in that, make computer as following cell operation: input is as the reference images signal of the signal of video signal that does not worsen and the deterioration signal of video signal after this reference images signal degradation, calculate the signal of video signal characteristic quantity of two signals, poor according to the signal of video signal characteristic quantity of two signals that calculate inferred the subjective quality of the subjective quality of described deterioration signal of video signal and infers the unit; Be used to revise the update information memory cell of the update information of described subjective quality with the storage that is mapped of signal of video signal characteristic quantity; Obtain the update information of the signal of video signal characteristic quantity of the described reference images signal that correspondence calculates from described update information memory cell, according to the subjective quality amending unit of the described subjective quality of obtaining of inferring out of update information correction.

In addition, for achieving the above object, the invention of the 9th form is a kind of video quality evaluation device, it is characterized in that, have: input is as the reference images signal of the signal of video signal that does not worsen and the deterioration signal of video signal after this reference images signal degradation, generates about from the reference images frame of described reference images signal and departing from and the calibration information generating unit of the calibration information that departs from the space of described reference images frame and described deterioration image frame from time of the deterioration image frame of described deterioration signal of video signal; According to described calibration information, eliminating after described spatial offset and described time departs from, calculate the space characteristics amount calculating part of space characteristics amount separately according to the spatial information of the image state of described reference images frame of expression and described deterioration image frame; According to described calibration information, eliminating after described spatial offset and described time departs from, for described reference images frame and described deterioration image frame, calculate the temporal characteristics amount calculating part of temporal characteristics amount separately according to the temporal information of the image variation of representing interframe separately; Infer portion according to the subjective quality that described space characteristics amount and described temporal characteristics amount are inferred the subjective quality of described deterioration signal of video signal.

The main idea of the invention of the tenth form is, in the invention of the 9th form, further have: in the document form of the reference images that comprises in the described reference images signal that the document form of the deterioration image that comprises in described deterioration signal of video signal is transformed in correspondence, export the formal argument portion of the information related, with the corresponding information related, store the correction factor storage part of the coefficient of the subjective quality that is used to infer described deterioration signal of video signal with described document form with its document form after reunification; Described subjective quality is inferred portion, obtain the described coefficient of the corresponding information related with the described unified document form of importing from described formal argument portion from described correction factor storage part, according to described space characteristics amount, described temporal characteristics amount and the described coefficient of obtaining, infer the subjective quality of described deterioration signal of video signal.

The main idea of the invention of the 11 form is, in the invention of the tenth form, described formal argument portion, as with the described unified related information of document form, export the signal form of described deterioration signal of video signal, the amount of information of the deterioration image that sends by described deterioration signal of video signal, and in the coded system of described deterioration signal of video signal at least any one, described correction factor storage part, the signal form of corresponding described deterioration signal of video signal, the amount of information of the deterioration image that sends by described deterioration signal of video signal, and in the coded system of described deterioration signal of video signal at least any one, the coefficient that storage is best.

The main idea of the invention of the 12 form is, in the invention of the 9th or the tenth form, described space characteristics amount calculating part, according to described reference images signal and described deterioration signal of video signal, the index of the deterioration that quantification brightness value in frame border jumpy is taken place is calculated as described space characteristics amount.

The main idea of the invention of the 13 form is, in the invention of the 12 form, described space characteristics amount calculating part, according to described reference images signal and described deterioration signal of video signal, the edge electric power amount that will stipulate in ANSIT1.801.03.1995 is calculated as described space characteristics amount.

The main idea of the invention of the 14 form is, the the 9th, the tenth or the invention of the 12 form in, described space characteristics amount calculating part, more described reference images frame calculates in the index of the situation that described deterioration image frame brightness value border jumpy that should the reference images frame is taken place on the horizontal vertical direction quantification as described space characteristics amount.

The main idea of the invention of the 15 form is, the the 9th, the tenth, the 12 or the invention of the 14 form in, described space characteristics amount calculating part, according to each as in the frame more than or equal to the piece of the set more than 1 pixel in reference images ITU-R Recommendation P.910 in regulation Temporal Information value and worsen the poor of Temporal Information value that image stipulates in P.910 at ITU-R Recommendation, the variable quantity of the interframe of image is calculated as described temporal characteristics amount.

The main idea of the invention of the 16 form is, in the invention of the 9th or the tenth form, and then has with the storage that is mapped of space characteristics amount and temporal characteristics amount and is used to revise the update information storage part of the update information of described subjective quality; Import the described space characteristics amount of described reference images signal and described temporal characteristics amount, from described update information storage part obtain the update information of the space characteristics amount of corresponding input and temporal characteristics amount, according to the subjective quality correction portion of the described subjective quality of obtaining of inferring out of update information correction.

In addition, for achieving the above object, the main idea of the invention of the 17 form is a kind of video quality evaluation method, it is characterized in that, input is as the reference images signal of the signal of video signal that does not worsen and the deterioration signal of video signal after this reference images signal degradation; Generation is about from the reference images frame of described reference images signal and departing from and the calibration information that departs from the space of described reference images frame and described deterioration image frame from time of the deterioration image frame of described deterioration signal of video signal; According to described calibration information, eliminating after described spatial offset and described time departs from, calculate separately space characteristics amount according to the spatial information of the image state of described reference images frame of expression and described deterioration image frame; According to described calibration information, having eliminated after described spatial offset and described time departs from, for described reference images frame and described deterioration image frame, the temporal information that changes according to the image of expression interframe is separately calculated temporal characteristics amount separately; Infer the subjective quality of described deterioration signal of video signal according to described space characteristics amount and described temporal characteristics amount.

In addition, for achieving the above object, the main idea of the invention of the 18 form is a video quality evaluation method, it is characterized in that, input is as the reference images signal of the signal of video signal that does not worsen and the deterioration signal of video signal of this reference images signal degradation; The document form of the deterioration image that comprises is transformed to the document form of the reference images that in the described reference images signal of correspondence, comprises in described deterioration signal of video signal; Generate the information related with this document form of unifying; Generation about the time between the deterioration image frame of the deterioration image that was transformed from the reference images frame of described reference images signal with from the document form of described deterioration signal of video signal depart from and described reference images frame and described deterioration image frame between the calibration information that departs from space; According to described calibration information, eliminating after described spatial offset and described time departs from, calculate separately space characteristics amount according to the spatial information of the image state of described reference images frame of expression and described deterioration image frame; According to described calibration information, having eliminated after described spatial offset and described time departs from, for described reference images frame and described deterioration image frame, the temporal information that changes according to the image of expression interframe is separately calculated temporal characteristics amount separately; According to described space characteristics amount, described temporal characteristics amount and corresponding related information with the described document form of unifying, be used for the coefficient that subjective quality is inferred, infer the subjective quality of described deterioration signal of video signal.

In addition, for achieving the above object, the invention of the 19 form is a kind of video quality evaluation program, and this program makes computer as following cell operation: input as the deterioration signal of video signal of the reference images signal of the signal of video signal that does not worsen and its reference images signal degradation, generate about from the reference images frame of described reference images signal and from the time between the deterioration image frame of described deterioration signal of video signal depart from and described reference images frame and described deterioration image frame between the calibration information generation unit of the calibration information that departs from space; According to described calibration information, eliminating after described spatial offset and described time departs from, calculate the space characteristics amount computing unit of space characteristics amount separately according to the spatial information of the image state of described reference images frame of expression and described deterioration image frame; According to described calibration information, eliminating after described spatial offset and described time departs from, for described reference images frame and described deterioration image frame, calculate the temporal characteristics amount computing unit of temporal characteristics amount separately according to the temporal information of the image variation of representing interframe separately; According to described space characteristics amount and described temporal characteristics amount, infer the subjective quality of the subjective quality of described deterioration signal of video signal and infer the unit.

The main idea of the invention of the 20 form is, in the invention of the 19 form, and then make computer as following cell operation: the document form of the deterioration image that in described deterioration signal of video signal, comprises, be transformed in the document form of the reference images that in the described reference images signal of correspondence, comprises, export the formal argument unit of the information related with this document form of unifying; With the correction factor memory cell that makes it be mapped with the information related, store the coefficient of the subjective quality that is used to infer described deterioration signal of video signal with described document form; Described subjective quality is inferred the unit, from described correction factor memory cell, obtain described coefficient corresponding and from the related information of the described document form of unifying of described formal argument unit input, according to described space characteristics amount, described temporal characteristics amount and the described coefficient of obtaining, infer the subjective quality of described deterioration signal of video signal.

In addition, for achieving the above object, the invention of the 21 form is a kind of image coalignment, it is characterized in that, have: input is as the reference images signal of the signal of video signal that does not worsen and the deterioration signal of video signal after this reference images signal degradation, the document form of the deterioration image that in described deterioration signal of video signal, comprises, be transformed to the formal argument portion of the document form of the reference images that in the described reference images signal of correspondence, comprises; The number of the deterioration image frame that comprises in reference images frame that coupling comprises in described reference images signal and the described deterioration signal of video signal and the Displaying timer matching part of Displaying timer; With for described reference images frame and described deterioration image frame to several frames before and after picture frame and its, departing from or the frozen state of described deterioration image of the frame of supervision between described reference images frame and described deterioration image frame, and obtain the corresponding relation of interframe and the corresponding relation between pixel coupling synchronously/location matches portion.

The main idea of the invention of the 22 form is, in the invention of the 21 form, and described formal argument portion, at least one makes with described reference images and coincide in the data mode of the described deterioration image of conversion, size, the length-width ratio.

The main idea of the invention of the 23 form is, in the invention of the 21 form, described Displaying timer matching part, in the described reference images frame occasion different with the frame rate of described deterioration image frame, by interpolation or delete described deterioration image frame, the frame rate of described reference images frame and described deterioration image frame is coincide.

The main idea of the invention of the 24 form is, in the invention of the 21 form, and described Displaying timer matching part, for the Displaying timer of described reference images frame, the Displaying timer of the described deterioration image frame of harmonizing.

The main idea of the invention of the 25 form is, in the invention of the 21 form, described Displaying timer matching part in the dissimilar occasion of the Displaying timer of described reference images frame, is taken as preset time to described reference images frame and described deterioration image frame both sides' Displaying timer at interval.

The main idea of the invention of the 26 form is, in the invention of the 21 form, described synchronously/location matches portion, for described reference images frame and described deterioration image frame to several frames before and after picture frame and its, the passing of the characteristic quantity separately of whole frames or its specific region, the grand Synchronous Processing of the corresponding relation that departs from the time that becomes minimum described reference images frame and described deterioration image frame of the described characteristic quantity of execution decision.

The main idea of the invention of the 27 form is, in the invention of the 26 form, described synchronously/location matches portion, for described reference images frame and described deterioration image frame to several frames before and after picture frame and its, move corresponding relation and the pixel corresponding relation of the time of described reference images frame and described deterioration image frame on one side, relatively whole characteristic quantity separately of frames or its specific region on one side, the difference of carrying out the described characteristic quantity of decision become the corresponding relation of time of minimum described reference images frame and described deterioration image frame and pixel corresponding relation little synchronously/location matches handles.

The main idea of the invention of the 28 form is, in the invention of the 27 form, described synchronously/location matches portion, the initial stage carry out described grand Synchronous Processing and described little synchronously/location matches handles.

The main idea of the invention of the 29 form is, in the invention of the 28 form, described synchronously/location matches portion, be absorbed in the occasion of frozen state at described deterioration image frame, by frame number being counted the continuation time of measuring described frozen state.

The main idea of the invention of the 30 form is, in the invention of the 29 form, described synchronously/location matches portion, for described reference images frame and described deterioration image frame to several frames before and after picture frame and its, derive the characteristic quantity separately of whole frames, change in time and the characteristic quantity of described deterioration image frame when not changing in time at characteristic quantity, judge that described deterioration image frame is absorbed in frozen state for described reference images frame.

The main idea of the invention of the 31 form is, in the invention of the 28 form, described synchronously/location matches portion, be absorbed in the occasion of frozen state at described deterioration image frame, perhaps can not with the occasion of described reference images frame synchronization, carry out described grand Synchronous Processing once more.

The main idea of the invention of the 32 form is, in the invention of the 28 form, described synchronously/location matches portion, be absorbed in the occasion of the state that frame departs from described deterioration image frame, output frame depart from number.

The main idea of the invention of the 33 form is, in the invention of the 21 form, and then have from described synchronously/location matches portion imports described reference images signal and described deterioration signal of video signal, the brightness of described deterioration image and colouring information and described reference images are coincide, and the deterioration image that described brightness and colouring information have coincide turn back to described synchronous/the brightness/color correction portion of location matches portion.

In addition, for achieving the above object, the invention of the 34 form is a kind of image matching method, it is characterized in that, input is as the reference images signal of the signal of video signal that does not worsen and the deterioration signal of video signal after this reference images signal degradation; The document form of the deterioration image that in described deterioration signal of video signal, comprises, be transformed to the document form of the reference images that in the described reference images signal of correspondence, comprises; The number and the Displaying timer of the deterioration image frame that comprises in reference images frame that coupling comprises in described reference images signal and the described deterioration signal of video signal; With for described reference images frame and described deterioration image frame to several frames before and after picture frame and its, departing from or the frozen state of described deterioration image of the frame of supervision between described reference images frame and described deterioration image frame, and obtain the coupling of the corresponding relation and the corresponding relation between pixel of interframe.

In addition, for achieving the above object, the invention of the 35 form is a kind of video matching program, it is characterized in that, make computer as following cell operation: input is transformed to the variation unit of the document form of the reference images that comprises in the described reference images signal of correspondence as the reference images signal of the signal of video signal that does not worsen and the deterioration signal of video signal after this reference images signal degradation, the document form of the deterioration image that comprises in described deterioration signal of video signal; The number of reference images frame that coupling comprises in described reference images signal and the deterioration image frame that in described deterioration signal of video signal, comprises and the Displaying timer matching unit of Displaying timer; And for described reference images frame and described deterioration image frame to several frames before and after picture frame and its, monitor frame between described reference images frame and described deterioration image frame depart from or the frozen state of described deterioration image, simultaneously obtain the corresponding relation of interframe and the corresponding relation between pixel coupling synchronously/the location matches unit.

In addition, for achieving the above object, the invention of the 36 form is a kind of video quality evaluation device, it is characterized in that having: input is transformed to the formal argument portion (41) of the document form of the reference images that comprises in the described reference images signal of correspondence as the reference images signal (RI) of the signal of video signal that does not worsen and the deterioration signal of video signal (PI) after this reference images signal degradation, the document form of the deterioration image that comprises in described deterioration signal of video signal; The number of reference images frame that coupling comprises in described reference images signal and the deterioration image frame that in described deterioration signal of video signal, comprises and the Displaying timer matching part (42) of Displaying timer; For described reference images frame and described deterioration image frame to several frames before and after picture frame and its, monitor frame between described reference images frame and described deterioration image frame depart from or the frozen state of described deterioration image, simultaneously obtain the corresponding relation of interframe and the corresponding relation between pixel coupling synchronously/location matches portion (43); With input by described synchronously/location matches portion (43) obtained coupling described reference images signal (RI) and described deterioration signal of video signal (PI), calculate two signals the signal of video signal characteristic quantity, infer portion (11) according to the subjective quality poor, that infer the subjective quality of described deterioration signal of video signal of the signal of video signal characteristic quantity of two signals of calculating.

Description of drawings

Fig. 1 is the block diagram of structure of first example of expression video quality evaluation device of the present invention;

Fig. 2 is the figure that is illustrated in the update information of storing in the update information database of Fig. 1;

Fig. 3 is the figure of expression based on the relation of the subjective quality of finally inferring subjective quality Q and actual measurement at the correction formula of once inferring subjective quality SQ revised;

Fig. 4 is the block diagram of structure of second example of expression video quality evaluation device of the present invention;

Fig. 5 is the figure that is used for calculating the horizontal vertical edge amount that the space characteristics amount calculating part at Fig. 4 uses;

Fig. 6 is the block diagram of structure of the 3rd example of expression video quality evaluation device of the present invention;

Fig. 7 is the figure that is illustrated in a plurality of conditions of storing in the weight coefficient database of Fig. 6 and the weight coefficient corresponding with condition;

Fig. 8 is the figure that the standard video data qualification used in the checking of example of the present invention is enumerated as learning data and verification msg;

Fig. 9 is SI (spatial information) value that calculates according to the learning data of Fig. 8 and verification msg and the distribution map of TI (temporal information) value;

Figure 10 is the figure that infers the result that the learning data that existing peak value SN obtains than (PSNR) is used in expression;

Figure 11 is the figure that infers the result that the learning data that existing edge electric power amount (Ave_EE) obtains is used in expression;

Figure 12 is that expression uses of the present invention second and the figure that infers the result of the learning data that obtains of the video quality evaluation device of the 3rd example;

Figure 13 is that expression uses of the present invention second and the figure that infers the result of the verification msg that obtains of the video quality evaluation device of the 3rd example;

Figure 14 is the figure that infers the result that represents only to use the learning data that existing edge electric power amount (Ave_EE) obtains;

Figure 15 is the figure of the relation of the minimum value (Min_HV) of expression horizontal vertical edge amount and subjective assessment value;

Figure 16 is the figure of the relation of expression piece mean motion electric power amount (Ave_MEB) and subjective assessment value;

Figure 17 is the block diagram of structure of the example of expression image coalignment of the present invention;

Figure 18 is the flow chart of action step of the example of expression image coalignment of the present invention;

Figure 19 is the figure that is used for illustrating the processing of Displaying timer matching part;

Figure 20 is the figure that is used for illustrating synchronous/grand moment Synchronous Processing of location matches portion;

Figure 21 be used for illustrating synchronously/location matches portion is little synchronously/figure that handles is derived in the position.

Embodiment

Describe the example of video quality evaluation device of the present invention, video quality evaluation method and video quality evaluation program and image coalignment, image matching method and video matching program with reference to the accompanying drawings in detail.First to the 3rd example is the working of an invention form of video quality evaluation device, video quality evaluation method and video quality evaluation program, and the 4th example is the working of an invention form of image coalignment, image matching method and video matching program.

＜the first example 〉

Illustrate with reference to Fig. 1 video quality evaluation device of the present invention first example respectively install part.Fig. 1 is the block diagram of structure of first example of expression video quality evaluation device of the present invention.

The video quality evaluation device of first example has subjective quality at least and infers portion 11, feature value calculation unit 12, update information database 13, corrected Calculation portion 14 and correction portion 15.

Subjective quality is inferred portion 11, input reference signal of video signal RI and deterioration signal of video signal PI.Reference images signal RI is the signal of video signal before worsening, and on the other hand, worsens signal of video signal PI, is the signal of video signal that reference images signal RI for example is encoded or has worsened via network.

Then, subjective quality is inferred portion 11, about reference images signal RI with worsen signal of video signal PI and calculate poor as the signal of video signal characteristic quantity separately of physical features amount.In the signal of video signal characteristic quantity, for example have be illustrated in comprise in the signal of video signal at the spatial information (SI) of the image state of the frame that image is arranged and be illustrated in the temporal information (TI) that the image in the interframe that image is arranged that comprises in the signal of video signal changes.As these spatial informations SI and temporal information TI, for example have at ITU-R Recommendation P.910 " Subjective Video Quality Assessment Methods for Multimedia Applications " Appendix A in the Spatial Information and the Temporal Information that stipulate.

Then, subjective quality is inferred portion 11, from the deterioration of the poor quantification deterioration signal of video signal PI of the signal of video signal characteristic quantity that calculates, according to this quantification deterioration infer subjective quality.That is subjective quality is inferred portion 11, by reference images signal RI with worsen the deterioration that signal of video signal PI quantification worsens signal of video signal PI, according to this by quantification deterioration infer subjective quality.The subjective quality of inferring out is inferred portion 11 as once inferring subjective quality SQ output from subjective quality.Once infer subjective quality SQ, for example if by edge electric power amount (E) and motion electric power amount (M) decision, the function representation shown in the then general available following formula (1):

SQ＝F(M，E) ...(1)。

This function F is tried to achieve by the subjective assessment experiment in advance.In addition, edge electric power amount (E) and motion electric power amount (M), in the quality of image objective evaluation yardstick (ANSIT1.801.03.1996, " Digital Transport of One-Way Video Signal Parameters forObjective Performance Assessment ") of U.S. ANSI regulation, be prescribed.

Feature value calculation unit 12, input reference signal of video signal RI calculates signal of video signal characteristic quantity FI according to this reference images signal RI.As this signal of video signal characteristic quantity FI for example also have living space as mentioned above information SI or temporal information TI.Feature value calculation unit 12 for example determines that any one quantitatively calculates its characteristic quantity at least among spatial information SI and the temporal information TI.

Update information database 13 is with the signal of video signal characteristic quantity storage update information that is mapped.The signal of video signal characteristic quantity for example is spatial information SI or temporal information TI as mentioned above.Update information is to be used to revise the correction formula of once inferring subjective quality SQ or the correction factor of inferring portion's 11 outputs as subjective quality.The back will describe these update informations in detail with reference to Fig. 2.These correction formulas or correction factor, decision by experiment is stored in the update information database 13 in advance.In image, added the occasion that worsens, seen subjective deterioration such characteristic of feeling which kind of degree behind this image by subjective assessment experiment unified calculation people in advance with signal of video signal characteristic quantity FI.According to correction formula and the correction factor of this property calculation for signal of video signal characteristic quantity FI.

That is, subjective assessment characteristic when unifiedly calculating the physical features amount of reference images signal RI and on having the image of its feature, adding deterioration by the subjective assessment test in advance, according to this physical features amount that calculates and the corresponding relation between the subjective assessment characteristic, correction is once inferred subjective quality SQ from the signal of video signal characteristic quantity of reference images signal RI by what subjective quality inferred that portion 11 derives, makes it to carry out accurately the conversion to unified objective evaluation value.

Corrected Calculation portion 14 from feature value calculation unit 12 input signal of video signal characteristic quantity FI, extracts out and corresponding correction formula and the correction factor of this signal of video signal characteristic quantity FI from update information database 13.That is, corrected Calculation portion 14, retrieval correction formula and the correction factor corresponding with signal of video signal characteristic quantity FI in update information database 13 extracted corresponding correction formula and correction factor out from update information database 13.Then, corrected Calculation portion 14 exports these correction formulas and correction factor as update information CI.

Correction portion 15 is inferred portion's 11 inputs from subjective quality and is once inferred subjective quality SQ, from the 14 input update information CI of corrected Calculation portion.Then, correction portion 15 once inferring in the correction formula of subjective quality SQ substitution with the correction factor that comprises, is once inferred subjective quality SQ as what infer finally that subjective quality Q output revises in update information CI.Finally infer subjective quality Q, revised quantificational expression worsen signal of video signal PI subjective quality SQ once infer subjective quality SQ.

Below with reference to Fig. 2 update information database 13 update informations storage, corresponding signal of video signal characteristic quantity are described.Fig. 2 is the figure that is illustrated in the update information of storing in the update information database 13 of Fig. 1.

As shown in Figure 2, in update information database 13, be mapped with each value of signal of video signal characteristic quantity, a plurality of correction factor data base systems.According to Fig. 2, by feature value calculation unit 12 the signal of video signal characteristic quantity (spatial information SI and temporal information TI) of reference images signal RI is calculated as (SI, TI)=(75,30) occasion, to correction factor that should the signal of video signal characteristic quantity be (α, β)=(α 2, and β 2).In fact, corrected Calculation portion 14 extracts the corresponding correction factor with this signal of video signal characteristic quantity FI from feature value calculation unit 12 input signal of video signal characteristic quantity FI out together from the database and the correction formula of update information database 13.And then, the occasion that also changes in corresponding signal of video signal characteristic quantity correction formula, in update information database 13, correction factor not only, the correction formula of corresponding signal of video signal characteristic quantity is also by data base system.

At last, with reference to Fig. 3 illustrate correction portion 15 correction once infer export behind the subjective quality SQ finally infer subjective quality Q.Fig. 3 is the figure that the relation between the revised subjective quality of finally inferring subjective quality Q and actual measurement that the correction formula of once inferring subjective quality SQ is obtained is used in expression.

In first example, correction formula is following formula (2).

Finally infer subjective quality Q=α * once infer subjective quality SQ+ β ... (2)

Here, α and β are correction factors.Fig. 3 represents by importing 3 figure lines that correction formula obtains that signal of video signal calculates for 3 kinds that are input in the signal of video signal evaluating apparatus.Transverse axis represent as subjective quality infer portion 11 output once infer subjective quality SQ, the longitudinal axis is represented for the subjective quality of once inferring subjective quality SQ, calculating by the actual subjective quality evaluation test of seeing its quality of image post-evaluation that obtains by the input signal of video signal of people.Representing respectively with circular mark, square mark and triangle mark among the figure, is that expression is estimated point for each input signal of video signal at the subjective quality of once inferring subjective quality SQ.On the other hand, 3 lines are correction formulas based on this example of corresponding respectively 3 kinds of input signal of video signal, according to this example, once infer subjective quality SQ and revise with the correction formula of such each image of correspondence.

Like this, according to first example, ask human vision property from the physical features amount of reference images for this image, as update information data base system for the characteristic quantity of reference images, for from reference images with worsen the subjective quality of inferring that the difference of characteristic quantity of the physics of image derives, pay with weight with this update information, like this, for any image just can with infer subjective quality as the identical precision unification of subjective estimate method of existing method.

In addition, in above-mentioned example, the structure of the signal of video signal characteristic quantity that calculates reference images signal RI as feature value calculation unit 12 logically as the monomer setting, but also such feature value calculation unit 12 can be set especially, the signal of video signal characteristic quantity of the reference images signal RI of portion's 11 derivation is inferred in the constant utilization of former state by subjective quality.

In addition, corrected Calculation portion 14 and correction portion 15 can logically be one not only physically, also.That is, also can be in the signal of video signal characteristic quantity of correction portion 15 direct input reference signal of video signal RI, correction portion 15 from update information database input to update information that should the signal of video signal characteristic quantity.

＜the second example 〉

Illustrate with reference to Fig. 4 video quality evaluation device of the present invention second example respectively install part.Fig. 4 is the block diagram of structure of second example of expression video quality evaluation device of the present invention.

The video quality evaluation device of second example as shown in Figure 4, has calibration information generating unit 21, space characteristics amount calculating part 22, temporal characteristics amount calculating part 23 and subjective quality and infers portion 24.

Calibration information generating unit 21, input reference signal of video signal RI and deterioration signal of video signal PI, receive the reference images frame from reference images signal RI respectively, receive the deterioration image frame from worsening signal of video signal PI, departing from of the frame time and the space of detection reference image frame and deterioration image frame, generate the calibration information that departs from about frame these times and the space.

So-called reference images frame and worsen the temporal of image frame and depart from is meant image that calibration information generating unit 21 carves the reference images frame that receives and deterioration image frame at a time departing from time.For example, calibration information generating unit 21 be engraved at a time in the reference images frame frame that receives certain image A, this time be engraved in and worsen the occasion that receives in the image frame than the frame before frame 3 frames of this image A, calibration information generating unit 21 detects for the reference images frame and has worsened image frame evening the fact of 3 frames, and this information is generated as calibration information.

In addition, so-called reference images frame and the departing from of space of worsening image frame, be meant that calibration information generating unit 21 carves the locus of the image that receives at a time and departed from.For example, the unswerving in time occasion of frame at reference images frame and deterioration image frame, calibration information generating unit 21 be engraved at a time the center mapping that is received in image center ball (ボ-Le) in the reference images frame frame, this time be engraved in worsen in the image frame from right 1 pixel in the center of image, on the occasion of frame of center mapping of the position ball that departs from of 2 pixels, calibration information generating unit 21 detects for the reference images frame and worsens the image frame fact that departed from of 1 pixel, last 2 pixels to the right, and this information is generated as calibration information.

Space characteristics amount calculating part 22, input reference signal of video signal RI and deterioration signal of video signal PI and calibration information, make the reference images frame and worsen image frame the space depart from and the time depart from disappearance, thereafter, according to reference images signal RI with worsen signal of video signal PI computer memory characteristic quantity.In this second example, as this space characteristics amount, the index of the edge electric power amount (Ave_EE) that will describe in detail below using and the minimum value (Min_HV) of horizontal vertical edge amount.

Edge electric power amount (Ave_EE):

This index is based on reference images signal RI and worsens the index that the deterioration (for example ambiguity) that takes place is gone up on signal of video signal PI quantification brightness value in frame border jumpy (being called the edge).In second example, as an example, emphasize the edge from the brightness value of pixel by using the Sobel filter, thus, come the deterioration on the quantification edge.In addition, here the edge electric power amount (Ave_EE) of quantification, at ANSI TI.801.03.1996, be prescribed in " Digital Transport of One-Way Video Signal Parametersfor Objective Performance Assessment ".

Edge electric power amount (Ave_EE) is with (3) formula quantification.

[mathematical expression 1]

$\mathrm{Ave}\_\mathrm{EE}=\sqrt{\frac{1}{M}\underset{m=0}{\overset{M-1}{\mathrm{\Σ}}}{\left(\frac{{\mathrm{SI}}_{\mathrm{in}}\left(m\right)-{\mathrm{SI}}_{\mathrm{out}}\left(m\right)}{{\mathrm{SI}}_{\mathrm{in}}\left(m\right)}\right)}^2}\·\·\·\·\·\·\left(3\right)$

In the formula, SI _μ(m) (μ=in or out) provided by (4) formula.

[mathematical expression 2]

$\mathrm{SI\μ}\left(m\right)=\sqrt{\frac{1}{N}\underset{i,j}{\mathrm{\Σ}}\{{\mathrm{SI}}^{2}h\_\mathrm{\μ}(i,j,m)+{\mathrm{SI}}^{2}v\_\mathrm{\μ}(i,j,m)\}-{\left(\frac{1}{N}\sqrt{{\mathrm{SI}}^{2}h\_\mathrm{\μ}(i,j,m)+{\mathrm{SI}}^{2}v\_\mathrm{\μ}(i,j,m)}\right)}^2}$

……(4)

μ=in or out

In addition, here, SI _H-μ(i, j, m), SI _V-μ(m), expression is positioned at the position of m frame (i, the Sobel filter of j) locating is provided by (5) formula and (6) formula respectively respectively for i, j.

SI _h-μ(i，j，m)＝{-Y _μ(i-1，j-1，m)-2Y _μ(i，j-1，m)-Y _μ(i+1，j-1，m)

+Y _μ(i-1，j+1，m)+2Y _μ(i，j+1，m)+Y _μ(i+1，j+1，m)}

……(5)

SI _v-μ(i，j，m)＝{-Y _μ(i-1，j-1，m)+Y _μ(i+1，j-1，m)-2Y _μ(i-1，j，m)

+2Y _μ(i+1，j，m)-Y _μ(i-1，j+1，m)+Y _μ(i+1，j+1，m)}

……(6)

In the formula, Y _In(m) expression is positioned at position (i, the brightness value of the pixel of j) locating (from 0 to 255), the Y of the m frame of reference images frame for i, j _Out(m) expression is positioned at position (i, the brightness value of the pixel of j) locating (from 0 to 255) of the m frame that worsens image frame for i, j.

In addition, back and figure together explanation only use that edge electric power amount (Ave_EE) infers the time the result, although subjective assessment value variation image very greatly only uses this edge electric power amount (Ave_EE) can not catch the deterioration that takes place fully in these images but existence objective evaluation value is almost constant.

The minimum value (Min_HV) of horizontal vertical edge amount:

This index, be and the reference images frame relatively, come the index of quantification in the degree that brightness value border jumpy (edge) in should the described deterioration image frame of reference images frame is taken place on the horizontal vertical direction.In this second example, as an example, the ratio of the edge amount that the place by the horizontal vertical edge amount that taken place by the place in the horizontal/of frame and the direction beyond it takes place, catch the characteristic quantity (Min_HV) of amount distortion, the degree of coming quantification to take place at the horizontal vertical top edge.This degree, as shown in Figure 5, ask for the edge amount (HV) of the some part that enters the horizontal/zone that determines by constant r0=20 and Δ θ=0.05236 and enter the ratio of the edge amount (HV strigula) in this zone in addition by importing image and worsening image.Adopt the reason of such index to be, mostly with the block unit image information of encoding, so be reflected in so-called distortion in such coding really.

The minimum value (Min_HV) of horizontal vertical edge amount is with formula (7) quantification.

[mathematical expression 3]

$\mathrm{Min}\_\mathrm{HV}=\underset{m}{\min }\left\{\frac{{\mathrm{HVR}}_{\mathrm{in}}\left(M\right)-{\mathrm{HVR}}_{\mathrm{out}}\left(m\right)}{{\mathrm{HVR}}_{\mathrm{in}}\left(m\right)}\right\}\·\·\·\·\·\·\left(7\right)$

In the formula, HVR _μ(m) (μ=in or out) provided by (8) formula.

[mathematical expression 4]

${\mathrm{HVR}}_{\mathrm{\μ}}\left(m\right)=\frac{{\mathrm{HV}}_{\mathrm{\μ}}(r_{\min },\mathrm{\Δ\θ},m)+0.5}{\stackrel{\‾}{{\mathrm{HV}}_{\mathrm{\μ}}}(r_{\min },\mathrm{\Δ\θ},m)+0.5}\·\·\·\·\·\·\left(8\right)$

μ=in or out

In addition, the definition represented of useful here (9) formula.

[mathematical expression 5]

${\mathrm{HV}}_{\mathrm{\μ}}(r_{\min },\mathrm{\Δ\θ},m)=\frac{1}{P}\underset{i,j}{\mathrm{\Σ}}{\mathrm{SI}}_{r\_\mathrm{\μ}(i,j,m)}\·\·\·\·\·\·\left(9\right)$

μ=in or out

Wherein, as condition, satisfy (10) formula and (11) formula.

SI _{r_μ}(i，j，m)≥r _min＞0 ……(10)

k _n/2-Δθ＜SI _{θ_μ}(i，j，m)＜k _n/2+Δθ(k＝0，1，2，3)

……(11)

Here, SI _{θ _ μ}(i, j, m)=tan ^-1[SI _{V_ μ}(i, j, m)/SI _{H_ μ}(i, j, m)].

In addition, the P of (9) formula is the pixel count of the scope that occupies of the shade of Fig. 5.

On the other hand, the definition that has (12) formula to represent.

[mathematical expression 6]

$\stackrel{\‾}{{\mathrm{HV}}_{\mathrm{\μ}}}(r_{\min },\mathrm{\Δ\θ},m)=\frac{1}{P}\underset{i,j}{\mathrm{\Σ}}{\mathrm{SI}}_{r\_\mathrm{\μ}}(i,j,m)\·\·\·\·\·\·\left(12\right)$

μ=in or out

Wherein, as condition, satisfy (13) formula and (14) formula.

SI _{r_μ}(i，j，m)≥r _min＞0 ……(13)

k _∏/2+Δθ＜SI _{θ_μ}(i，j，m)＜(k+1) _∏/2-Δθ(k＝0，1，2，3)

……(14)

In addition, the P of (12) formula is the pixel count of the scope of satisfied (13) formula and (14) formula.

The minimum value (Min_HV) of horizontal vertical edge amount, be in edge electric power amount (Ave_EE), catch less than deterioration, specifically, be to catch because the characteristic quantity at the newly-generated edge of piece distortion.Can illustrate with reference to the accompanying drawing back, can catch its deterioration sensitively by this index.This index based on set quota among the ANSI, but, shown in (7) formula, is modified into the generation quantity that can only catch new edge in high sensitivity alone by using minimum value.

Yet, temporal characteristics amount calculating part 23, input reference signal of video signal RI and deterioration signal of video signal PI and calibration information are eliminated the spatial offset and the time of reference images frame and deterioration image frame and are departed from, thereafter, according to reference images signal RI and deterioration signal of video signal PI characteristic quantity computing time.As this temporal characteristics amount, the conduct that will describe in detail below using is based on the piece mean motion electric power amount (Ave_MEB) of the index of the difference of TI (temporal information) value in this second example.The TI value is brightness value poor of the pixel between the image frame, has been prescribed in P.910 at ITU-R Recommendation.

This piece mean motion electric power amount (Ave_MEB):

This index, be for derive reference images frame as each piece of several pixels set in the frame and worsen image frame the TI value poor, according to the TI value of every of reference images frame index to this difference operative normization.

That is piece mean motion electric power amount (Ave_MEB) is represented with (15) formula.

[mathematical expression 7]

$\mathrm{Ave}\_\mathrm{MEB}=\frac{1}{M}\underset{M=0}{\overset{M-1}{\mathrm{\Σ}}}\frac{1}{N_b}\sqrt{\underset{(k,l)}{\mathrm{\Σ}}{\left(\frac{{\mathrm{TI}}_{b\_\mathrm{in}}(k,l,m)-{\mathrm{TI}}_{b\_\mathrm{out}}(k,l,m)}{{\mathrm{TI}}_{b\_\mathrm{in}}(k,l,m)}\right)}^2}\·\·\·\·\·\·\left(15\right)$

In the formula, TI _B-μ(k, l m) represent with (16) formula.

[mathematical expression 8]

${\mathrm{TI}}_{b\_\mathrm{\μ}}(k,l,m)=\frac{1}{64}\underset{i,j}{\mathrm{\Σ}}{(Y_{\mathrm{\μ}}(8k+i,8l+j,m)-Y_{\mathrm{\μ}}(8k+i,8l+j,m-1))}^2$

……(16)

μ=in or out

Piece mean motion electric power amount (Ave_MEB) is the deterioration of catching in edge electric power amount (Ave_EE) not to the utmost, specifically, is the characteristic quantity of seizure based on the generation of the deterioration of the motion in each zone.Can illustrate with reference to the accompanying drawing back, can catch its deterioration sensitively by this index.This index is to derive every TI value for the motion of catching each zone, and then carries out standardization by the value with reference images, and the yardstick alone that sensitivity is increased.

Subjective quality is inferred portion 24, the temporal characteristics amount of calculating the space characteristics amount of being calculated by space characteristics amount calculating part 22 with by temporal characteristics amount calculating part 23 is as input, infers as the subjective assessment value for the subjective quality of the deterioration signal of video signal PI of reference images signal RI.This subjective assessment value (Y) is calculated with (17) formula.

Y＝αX ₁+βX ₂+γX ₃+δ ……(17)

In the formula, X ₁=Ave_EE, X ₂=Min_HV, X ₃=Ave_MEB.

This subjective assessment value is the value that finally obtains.

Here, α, β, γ, δ be for from relation for the subjective assessment value of the deterioration image of trying to achieve by subjective assessment test, decide the weight coefficient of trying to achieve in advance for temporal characteristics amount and space characteristics amount with the corresponding relation of subjective assessment value.In addition, α, β, γ, δ for example, are according to the signal form that worsens signal of video signal PI, are decided by amount of information (size) that worsens the deterioration image that signal of video signal PI sends and the coded system that worsens signal of video signal.Therefore, from the signal form that worsens signal of video signal PI, by amount of information (size) that worsens the deterioration image that signal of video signal PI sends and the coded system that worsens signal of video signal, can calculate α, β, γ, δ in advance.

According to the video quality evaluation device of second example of above explanation, can infer subjective quality for image unification arbitrarily.How many back will describe this second example in detail with reference to accompanying drawing video quality evaluation devices can improve with respect to existing video quality evaluation device, precision.

In addition, also can infer the subjective quality that the video quality evaluation device shown in Figure 4 that illustrated is assembled into video quality evaluation device shown in Figure 1 in the portion 11 in second example, use video quality evaluation device shown in Figure 1.That is, video quality evaluation device input reference signal of video signal RI shown in Figure 4 and deterioration signal of video signal PI, the value of subjective quality being inferred portion's 24 outputs is as once inferring subjective quality SQ to correction portion 15 outputs.Correction portion 15 receives update information CI from corrected Calculation portion 14, and subjective quality SQ is once inferred in input, calculates and finally infers subjective quality Q.

＜the three example 〉

Illustrate with reference to Fig. 6 video quality evaluation device of the present invention the 3rd example respectively install part.Fig. 6 is the block diagram of structure of the 3rd example of expression video quality evaluation device of the present invention.

The video quality evaluation device of the 3rd example, with the video quality evaluation device of second example relatively, difference is: to the reference images signal RI of device input with worsen signal of video signal PI and import with different mutually document forms respectively; Moreover, worsen the signal form of signal of video signal PI, be unknown by amount of information (size) that worsens the deterioration image that signal of video signal PI sends and the coded system that worsens signal of video signal.Therefore, to giving same symbol, omit explanation with the same part of the video quality evaluation device of second example.

The video quality evaluation device of the 3rd example as shown in Figure 6, has format conversion portion 35, calibration information generating unit 21, space characteristics amount calculating part 22, temporal characteristics amount calculating part 23, subjective quality and infers portion 34 and weight coefficient database 36.

Format conversion portion 35, at the input reference image file with after worsening image file, and the document form occasion different with the document form of reference images file of deterioration image file is transformed to the form identical with the document form of reference images file to the document form that worsens image file.More detailed theory, for example, in the signal form, distribution of color, size, length-width ratio and the coded system that worsen image file different occasion, the document form that worsens image file is transformed to the form identical with the document form of reference images file with the reference images file.Then, format conversion portion 35 worsens signal of video signal PI from conversion for the deterioration image file output with reference images file same file form.From form the deterioration signal of video signal PI of transformation component 35 is output to calibration information generating unit 21, space characteristics amount calculating part 22, temporal characteristics amount calculating part 23.

More specifically say, format conversion portion 35, for example, in the imagination reference images is YUV form reference images, the reference images occasion different with the color format of deterioration image of VGA size, if worsening image is the RGB form, then use the transform of in Rec.IUT-R BT.601 " STUDIO ENCODINGPARAMETERS OF DIGITAL TELEVISION FOR STANDARD 4:3 ANDWIDE-SCREEN 16:9 ASPECT RATIOS " waits, putting down in writing to change this image.In addition, reference images is transformed to identical occasion with the size or the length-width ratio that worsen image, can calculate as integral multiple simply sometimes, but putting invalid occasion at this need be transformed to arbitrary dimension in format conversion portion 35.In this occasion, use known method to carry out conversion (for example reference to arbitrary dimension, Muramatsu S.and Kiya H.: " Scale Factor of Resolution Conversion Based onOrthogonal Transforms; " IEICE Trans.Fundamentals., E76-A, 7, pp.1150-1153 (July 1993), perhaps loose village just I, your family's benevolence will, " for the resolution conversion method of the multiplying power of rational arbitrarily that changes image encoded " collection of thesis A of electronic information communication association, Vol.77-A, No.3, pp.369-378, March 1994).In addition, the occasion that signal of video signal has been compressed is transformed to non-compressed format in advance.

In addition, since image coding or the characteristic of transfer system or the difference of country variant specification, feasible color or the lightness distribution difference that worsens image, therefore, for avoiding causing a deviation, carry out standardization to reference images and the distribution that worsens the brightness value etc. of image to the objective evaluation value.That is, for reference images and deterioration image, for each of brightness, color difference signal or rgb value, distribution from the pixel value of particular frame, derive its maximum, minimum value, mean value, the such statistic of variance, for making its distribution identical, worsen whole pixel values of image by the quantity displacement that for example departs from mean value, perhaps carry out conversion, come standardization is carried out in the distribution of reference images and deterioration image by using at the transform shown in the minimum value that is being used to make reference images and worsen image (18) formula consistent with peaked scope.

[mathematical expression 9]

In the formula, Y: the pixel value of the deterioration image after the conversion

X: the pixel value of the deterioration image before the conversion

Portion 34 output and the document form of the deterioration signal of video signal PI of conversion, i.e. the document form related information same with reference images signal RI are inferred to subjective quality by format conversion portion 35.The information related with this document form, for example be signal form, distribution of color, size, length-width ratio and the coded system that worsens image file, but in the 3rd example, infer signal form, size and the coded system that portion's 34 outputs worsen image file to subjective quality.

Weight coefficient database 36 as a set condition, stores a plurality of these conditions to the signal form, size and the coded system that worsen image file, and corresponding each condition preestablishes weight coefficient α, β, γ, δ.These weight coefficients, be from and by subjective assessment experiment try to achieve at the relation between the subjective assessment value that worsens image, for temporal characteristics amount and space characteristics amount, in advance try to achieve and the subjective assessment value between the coefficient of corresponding relation.Fig. 7 represents the weight coefficient of a plurality of conditions and corresponding these conditions of 36 storages of weight coefficient database.

Subjective quality is inferred portion 34, correspondence is from signal form, size and the coded system of the deterioration image file of format conversion portion 35 inputs, obtain weight coefficient α, β, γ, δ from weight coefficient database 36, it is different that this point and the subjective quality of second example are inferred portion 24.To infer portion 24 identical with subjective quality for other.

In addition, in Fig. 6, format conversion portion 35 and calibration information generating unit 21 are disposed respectively, are assembled in the video quality evaluation device of Fig. 6 but format conversion portion 35 and calibration information generating unit 21 also can be used as a formation portion.

According to basis the 3rd example of above explanation,, also can infer the subjective quality that worsens image even in the document form occasion different that worsens image file with the reference images file.Have again, about worsening signal form, size and the coded system of image file, because can corresponding a plurality of patterns, so can infer the subjective quality of various deterioration image files.

More existing below video quality evaluation device to Figure 16, illustrates the actual result of inferring precision of the quality of image value of inferring about the video quality evaluation device of the second and the 3rd example with reference to Fig. 8.

At first, illustrate and be used in the checking of inferring precision and derive the subjective assessment data that the weight coefficient in 36 storages of weight coefficient database uses.The subjective assessment data, shown in Fig. 8 with the selected standard video of ITU-R (with reference to ITU-R BT.802-1, " Test Pictures and Sequences forsubjective Assessments of Digital Codecs ' Coveying Signals produced Accordingto Recommendation ITU-RBT.601; " 1994, and ITU-R BT.1201-2, " Testmaterials to be used in subjective assessment; " 2001) 36 kinds as image, as shown in Figure 8, standard video is divided into the verification msg of inferring precision that is used to verify the video quality evaluation value, with learning data as the data of in deriving coefficient, using in advance.

Be used to verify the subjective assessment data of inferring precision, be selected as reducing because the influence that the imbalance of the characteristic of reference images causes.That is, consideration at ITU-TP.910 (with reference to ITU-TP.910, " SubjectiveVideo Quality assessment Methods for multimedia applicatons " Aug.1996.) spatial information (SI) of expression, the distribution of temporal information (TI) in, select the image of similar number to region D from the regional A of Fig. 9, like this, can be image as reference images with various SI values and TI value.In addition, for these reference images, use based on the coding deterioration of MPEG4 and in the scope of 256kbps～8Mbps, make to become the deterioration image of distributing to for 4 stages.

The subjective quality evaluation assessment, the DSCQS method that use is often used in the benchmark test of the codecs such as checking of MPEG is (with reference to ITU-R BT.500-10, " Methodology firthe subjective assessment of the quality of television pictures " March 2000.), the people of reception test is from 24 of 18 years old to 40 years old common men and women.

The target of video quality evaluation device is inferred precision, and it is suitable making the degree that equals the dispersion of evaluation and test point in evaluation.When the degree of the dispersion of subjective assessment value is derived as one-sided 95% credibility interval, be 7.24 in total data.Therefore, infer precision, be no more than this value as target with average variance (RMSE) as the target of video quality evaluation device.

In Figure 10 and Figure 11, be illustrated in respectively in the existing video quality evaluation device, general that use, use the result who has adopted PSNR (Peak Signal Noise Ratio: peak value SN than) and inferred out as the data of the known edge electric power amount (Ave_EE) of the parameter of ANSI, for learning data.Figure 10 is the result that infers in the occasion of having used PSNR, and Figure 11 is the result that infers in the occasion of having used edge electric power amount (Ave_EE).According to Figure 10 and Figure 11, the RMSE of PSNR is 9.57, and the RMSE of edge electric power amount (Ave_EE) is 7.47, than PSNR, edge electric power amount (Ave_EE) side expresses good characteristic, but RMSE surpasses desired value, it is inadequate inferring precision.

Figure 12 and Figure 13 have represented to use the video quality evaluation device of the second and the 3rd all example of edge electric power amount (Ave_EE), the minimum value (Min_HV) of horizontal vertical edge amount, Ave_MEB, for the result that infers of learning data, verification msg.According to Figure 12 and Figure 13, because the RMSE of learning data is 6.43, the RMSE of verification msg is 6.49, so, not only for learning data, can realize also that for verification msg enough targets infer precision as non-learning data.Therefore, obvious, the video quality evaluation device of the second and the 3rd example has and can replace people to see the precision of inferring that the subjective assessment quality test of actual its quality of image post-evaluation uses, and has reached realistic scale.

Figure 14 is the figure that infers the result that represents only to use the part of the learning data that edge electric power amount (Ave_EE) obtains.As shown in figure 14, obvious, the gradient of locating to evaluate and test a little in input image [3], [6], [7], [9] is big, although the objective evaluation value changes hardly, the subjective assessment value but alters a great deal, and only uses edge electric power amount (Ave_EE) to catch and is not all in the deterioration that takes place in these images.

Figure 15 is the figure of the relation of the minimum value (Min_HV) of expression horizontal vertical edge amount and subjective assessment value.As shown in figure 15, as can be known,, in input image [3], [7], catch this deterioration sensitively according to the minimum value (Min_HV) of horizontal vertical edge amount.

Figure 16 is the figure of the relation of expression piece mean motion electric power amount (Ave_MEB) and subjective assessment value.As shown in figure 16, as can be known,, in input image [6], [9], catch this deterioration sensitively according to piece mean motion electric power amount (Ave_MEB).

As mentioned above, according to the of the present invention second and the 3rd example, physical features amount as image, except the parameter of catching the deterioration that edge region takes place, by appending the parameter alone of two additional precision deficiencies, can be in specific image, keeping inferring the technological expansion of precision to any image.

＜the four example 〉

The following describes the working of an invention form that relates to image coalignment, image matching method and video matching program.

Figure 17 is the block diagram of structure of the example of expression image coalignment of the present invention.

The image coalignment of the 4th example, have at least format conversion portion 41, Displaying timer matching part 42, synchronously/location matches portion 43, brightness/color correction portion 44 and deterioration amount leading-out portion 45.

Format conversion portion 41 makes the document form that comprises in owing to the deterioration signal of video signal after coding or the deterioration of the loss in network, coincide with the form of the reference images that comprises in the reference images signal.Displaying timer matching part 42 coincide the image Displaying timer of reference images signal and deterioration signal of video signal.Synchronously/location matches portion 43, from deterioration amount leading-out portion 45 obtain deterioration amount or synchronous runout information on one side, obtain the coupling of the time and space direction of reference images signal and deterioration signal of video signal on one side.In addition, as required, brightness/color correction portion 44 is revised with the brightness or the different of distribution of color of worsening signal of video signal for the reference images signal.

In addition, in this example, in reference images signal and deterioration signal of video signal, comprise frame rate information or frame and show constantly/be taken into time information.In addition as required, also comprise signal format information.Thus, this image coalignment, for reference images and worsen image, on one side in memory storage object frame and front and back thereof count frame, carry out non real-time on one side and handle.

Figure 18 is the flow chart of action step of the example of expression image coalignment of the present invention.

At first, format conversion portion 41 in reference images and the different occasion of signal form, size, length-width ratio that worsens image, worsens the conversion (step S1) of the signal form of image.For example, if the data mode of reference images is the YUV form of non-compression, the data mode that worsens image is the RGB form of non-compression, then uses Rec.ITU-R BT.601 " STUDIO ENCODINGPARAMETERS OG DIGITAL TELEVISION FOR SDANDART 4:3 ANDWIDE-SCREEN 16:9 ASPECT RATIOS " the transform conversion worsen image.In addition, originally format conversion portion 41 is that the occasion of compressed format will be transformed to non-compressed format in advance worsening image.And then format conversion portion 41 in the size occasion different with length-width ratio, carries out conversion so that they become identical.For example, sometimes can calculate as integral multiple simply, but in the occasion of only doing so not enough, need be transformed to arbitrary dimension, in this occasion, use known method to carry out conversion (for example reference to arbitrary dimension, Muramatsu S.and Kiya H.: " Scale Factor of Resolution ConversionBased on Orthogonal Transforms; " IEICE Trans.Fundamentals., E76-A, 7, pp.1150-1153 (July 1993), perhaps loose village just I, your family's benevolence will, " for the resolution conversion method of the multiplying power of rational arbitrarily that changes image encoded " collection of thesis A of electronic information communication association, Vol.77-A, No.3, pp.369-378, March 1994.).Thereafter, format conversion portion 41 hands to Displaying timer matching part 42 to the deterioration signal of video signal after reference images signal and the conversion.

Then, Displaying timer matching part 42 coincide for making the Displaying timer by the deterioration signal of video signal of format conversion portion 41 format conversions and reference images signal, carries out the processing (step S2) of the polishing etc. of frame.

Specifically, shown in Figure 19 (a), at interval than the demonstration of the reference images occasion of constant length at interval, Displaying timer matching part 42 shown in the hypomere of Figure 19 (b), replenishes image by the polishing full frame for worsening image in the demonstration that worsens image.

In addition, shown in Figure 19 (b), in the occasion that the demonstration that worsens image is fluctuateed at interval, the additional image that worsens of image image before shown in the hypomere of Figure 19 (b), by the Displaying timer synchronous with reference images, is worsened with being right after in Displaying timer matching part 42.In addition, at this moment, also can make the image that is not with before the deterioration image and replenish, replenish but go up approaching image with the time in the deterioration image.That is, for example in the occasion of Figure 19 (b), second frame that worsens image as conversion has replenished frame I, if but supposition replenishes with upward approaching image of time in the deterioration image, then to replenish frame II.

And then, shown in Figure 19 (c), not the occasion of the frame rate of fixing at reference images signal self, then want polishing to become with the correct time interval different and show the reference images signal and worsen signal of video signal with them.

Thereafter, Displaying timer matching part 42 is handed to reference images signal and deterioration signal of video signal synchronously/location matches portion 43.

Synchronously/and location matches portion 43, imagination (1) is estimated initial state, and (2) synchronous regime behind the three state of (3) frozen state, is stipulated its action.

(1) at first, in estimating initial state, synchronously/location matches portion 43, for obtaining the grand coupling of time orientation, relatively for the reference images signal of certain hour with worsen the passing of the such characteristic quantity of the mean value of brightness/aberration/rgb value of signal of all or its specific region of 1 frame of signal of video signal, obtaining matching becomes the highest frame information, thus, derives depart from (the step S3) of grand time orientation.Specifically, as shown in figure 20, characteristic quantity being departed from time orientation, is minimum from the difference of each time series value, and perhaps cross-correlation coefficient is derived depart from (frame is poor) of time orientation for maximum condition.

In addition, synchronously/location matches portion 43, for obtaining the coupling of little time and space direction, in reference images, count frame by deterioration amount leading-out portion 45 for the front and back of the image of having obtained coupling, as shown in figure 21, make and worsen image in the pixel that moves up and down, the position that is minimum to difference is as having obtained the position of mating most, as location of pixels message pick-up (step S4).

And then, synchronously/and location matches portion 43, deliver the reference images that has obtained coupling and worsen image to brightness/color correction portion 44 for carrying out the brightness/color correction.Brightness/color correction portion 44 makes the brightness and mean value, minimum value and the maximum of colouring information and their distribution that worsen image, obtains coupling with these of reference images.For example, in reference images and the different occasion of Luminance Distribution that worsens image, brightness/color correction portion 44, according to separately mean value and the variance yields of reference images with the Luminance Distribution that worsens image, implement linear standardization for reference images and the Luminance Distribution that worsens image, about being that the information of carrying out the standardized transform of this linearity is handed to synchronously/location matches portion 43 as update information.Synchronously/and location matches portion 43, receive this update information, carry out brightness/color correcting process (step S5) according to the update information that receives.

Below, as the processing in (2) synchronous regime (3) frozen state, synchronously/location matches portion 43, carry out the processing of step S6 shown in Figure 180 to step S22.

Littler at reference images object frame number i than the final frame number N of reference images, perhaps worsen imaged object frame number j than worsening image final frame number M hour, carry out the processing (step S6) of following step S7 to step S22.

So, at first, judge whether F1 is 1, promptly whether reference images and deterioration image are in asynchronous regime (step S7).In reference images with worsen the occasion (F1=0) that image is in synchronous regime, transfer to step S8, in reference images with worsen the occasion (F1=1) that image is in asynchronous regime, transfer to step S16.

At first, below explanation reference images and deterioration image are in the occasion (F1=0) of synchronous regime.

Synchronously/and location matches portion 43, deliver the deterioration image and the reference images that depart from or revised brightness/color information that obtain at above-mentioned evaluation initial state, the time and space direction to deterioration amount leading-out portion 45.Deterioration amount leading-out portion 45, for worsen image to picture frame, frame counted in the frame of the reference images of correspondence and front and back thereof carry out little Synchronous Processing (with reference to Figure 21) as object, when deriving the deterioration amount that obtains thus, the frame difference between reference images and the deterioration image previous frame is separately derived (step S8) as deterioration amount.Synchronously/and location matches portion 43, receive those deterioration amount that derive by deterioration amount leading-out portion 45, in those deterioration amount, worsen according to frame difference, judgement whether image is frozen state (step S9).That is, with respect to difference table certain value is shown about reference images, and if the difference that worsens image is expressed and is close in 0, then synchronously/location matches portion 43 judges that worsening images is frozen states.

When worsening image and be frozen state, synchronously/location matches portion 43 several freezing when being set at asynchronous regime (F1=1) as " 1 " (count=1) (step S 10), transfer to step S22.

On the other hand, when worsening image and being not frozen state, then, synchronously/location matches portion 43, judge obtain in little Synchronous Processing of step S8 by deterioration amount leading-out portion 45, worsen image to the deterioration amount of picture frame whether minimum (step S11).In the deterioration amount to picture frame that worsens image is hour, is set at synchronous regime (F2=0) (step S12), transfers to step S14.On the other hand, in the deterioration amount to picture frame that worsens image is not hour, be judged to be frame and depart from state (frame-skipping state, return frame state), transfer to step S13, the frame number of deterioration amount minimum is set at the j value, if the words of frame-skipping are set at F2=1, otherwise (returning frame state etc.) is set at F2=2, and the number that departs from of frame at that time is set at the Count value.

At step S14, export F2 value and Count value at the output reference image with when worsening image., Count value be reset to 0, transfer to step S22 thereafter.

Then, below explanation is judged the occasion that is in asynchronous regime (F1=1) for reference images and deterioration image at step S7.

Synchronously/location matches portion 43, at step S16, S8 is same with step, for worsen image to picture frame, receive by the frame of corresponding reference images and front and back thereof being counted the deterioration amount that little Synchronous Processing that frame carries out as object obtains from deterioration amount leading-out portion 45, simultaneously, will and reference images and worsen image separately the frame difference between the former frame receive as deterioration amount.Synchronously/and location matches portion 43, S9 is identical with step, in these deterioration amount, according to frame difference, judges and worsens whether image is frozen state (step S17).

When the deterioration image is frozen state, synchronous/location matches portion 43 will freeze number (count value) and add 1 (step S18), transfer to step S22.

On the other hand, when the deterioration image is not frozen state, synchronously/and the 43 judgement frozen states end of location matches portion, carry out the grand moment Synchronous Processing (step S19) same, freezeout number (count value) (step S20) with step S3., F1 value and count value be reset to 0, transfer to step S22 thereafter.

At step S22, with reference images object frame number i and worsen imaged object frame number j and add 1, reach the occasion that the final frame number N of reference images and this j value reach the final frame number M of deterioration image, processing end in this i value.

In addition, suppose always at certain hour, for reference images signal and deterioration signal of video signal, savings has 1 frame all or the such characteristic quantity of mean value of the brightness/aberration/rgb value of specific region, and fail for a certain reason to get when synchronous, when for example more than the characteristic quantity in becoming the frame of object is leaving the mean value threshold value (constant times of the standard deviation as 3 σ) of regular period, disconnection having taken place, think the evaluation A-stage that enters (1), restart to handle from step S3.

More than be the details of the image matching treatment of example of the present invention.

According to this example, in the format conversion of carrying out image, by grand matching treatment and little matching treatment are carried out Synchronous Processing often as the basis, even the size or the different occasion of length-width ratio of the image that receives when in personal computer terminal, receiving image, perhaps can not obtain the occasion of the reference images and the correspondence of the locus of worsening image in the information loss of concentrating owing to certain degree as loss of packets, perhaps and then because the fluctuation of IP packet arrival interval or the generation of loss of packets, the timing that shows at the image of time shaft departs from or the fluctuation of image Displaying timer or freeze the occasion that so always non-existent phenomenon takes place, and can both suitably obtain reference images and worsen the time orientation of image and the coupling of direction in space.

But, in image matching method of the present invention, be envisioned for from synchronously/reference images of location matches portion 43 output and worsen image and satellite information (state of time orientation) input video quality evaluation device (subjective quality that for example is transfused to the video quality evaluation device of first example is inferred portion 11), but, also can be in deterioration amount leading-out portion 45, not to derive deterioration amount but derivation objective evaluation value, if by synchronously/and 43 its results of output of location matches portion, also can be used as video quality evaluation device and use.

In addition, in each example of above explanation, the indication of representing in its treatment step can be carried out based on the program as software.General-purpose computing system, good this program of storage in advance by reading in this program, can allow this general-purpose computing system come work as video quality evaluation device and image coalignment.The indication of recording and narrating in above-mentioned each example, the program as allowing computer carry out is recorded in disk (floppy disk, hard disk etc.), CD (CD-ROM, CD-R, CD-RW, DVD-ROM, DVD ± R, DVD ± RW etc.), on semiconductor memory reservoir or the recording medium similar with it.Computer is from this recording medium read-in programme, if carry out the indication of recording and narrating according to this program by CPU in program, then computer just moves as the video quality evaluation device and the image coalignment of above-mentioned example.

In addition, the present invention is not limited to indiscriminately imitating of above-mentioned example, and can the implementation phase in the scope that does not break away from its main idea, change structural element and specialize.In addition, the suitable combination according to disclosed a plurality of structural elements in above-mentioned example can form various inventions.For example, also can from all structural elements shown in the example, delete several structural elements.And then, also can appropriate combination relate to the structural element of different examples.

According to video quality evaluation device of the present invention, video quality evaluation method and video quality evaluation system, even unknown reference images, also can for any image high-precision infer unified subjective quality.

Specifically, according to video quality evaluation device of the present invention, video quality evaluation method and video quality evaluation system, ask for visual characteristic from the physical features amount of reference images at the people of this image, carry out data base system as update information for the reference images characteristic quantity, infer subjective quality with this update information additional weight by what derive for difference from reference images and the physical features amount that worsens image, just can be for any image to derive subjective quality uniformly with the equal precision of the subjective estimate method of existing method.

In addition, as the physical features amount of image, except the parameter of catching the deterioration that edge region takes place,, just can infer the technology of precision to the maintenance in specific image of image expansion arbitrarily by appending remedy the precision deficiency two parameters alone.

According to image coalignment of the present invention, image matching method and video matching program, in the inferring of subjective quality, at the benchmark signal of video signal with when worsening physical features amount between signal of video signal, can positively obtain the coupling of their the position space and the time.

Specifically, in the format conversion of carrying out image, by being that mate synchronously on the basis often with grand matching treatment and little matching treatment, even the size or the different occasion of length-width ratio of the image that receives when in personal computer terminal, receiving image, perhaps can not obtain the occasion of the reference images and the correspondence of the locus of worsening image in the information loss of concentrating owing to certain degree as loss of packets, perhaps and then because the fluctuation of IP packet arrival interval or the generation of loss of packets, the timing that shows at the image of time shaft departs from or the fluctuation of image Displaying timer or freeze the occasion that so always non-existent phenomenon takes place, and can both suitably obtain reference images and worsen the time orientation of image and the coupling of direction in space.

Therefore, according to being the video quality evaluation device and the method for prerequisite,, can improve its precision even also can carry out the objective evaluation of the quality of image in above-mentioned each occasion with such image coalignment and method.

Claims (7)

1. a video quality evaluation device is characterized in that,

Have:

Input is as the reference images signal (RI) of the signal of video signal that does not worsen and the deterioration signal of video signal (PI) after this reference images signal degradation, calculate the signal of video signal characteristic quantity of two signals, poor according to the signal of video signal characteristic quantity of two signals that calculate inferred the subjective quality of the subjective quality of described deterioration signal of video signal and infers portion (11);

Be used to revise the update information storage part (13) of the update information of described subjective quality with the storage that is mapped of signal of video signal characteristic quantity; With

Import the described signal of video signal characteristic quantity of described reference images signal, obtain and the corresponding update information of having imported of signal of video signal characteristic quantity from described update information storage part (13), subjective quality correction portion (12 according to the described subjective quality of obtaining of inferring out of update information correction, 14,15).

2. video quality evaluation device according to claim 1 is characterized in that,

Described subjective quality is inferred portion (11), according to the spatial information (SI) of the image state in each frame that is illustrated in the image that comprises in the signal of video signal and be illustrated in poor between any one at least deterioration signal of video signal in the temporal information (TI) that the image of interframe changes and reference images signal, infer described subjective quality.

3. video quality evaluation device according to claim 2 is characterized in that,

Described subjective quality is inferred portion (11), poor as the described signal of video signal characteristic quantity between described deterioration signal of video signal and described reference images signal, computational chart are shown in the edge electric power amount (E) of the deterioration amount in each frame of the image that comprises in the signal of video signal and are illustrated at least one side in the mobile electric power amount (M) of deterioration amount of interframe.

4. video quality evaluation device according to claim 2 is characterized in that,

Described subjective quality is inferred portion (11) and is inferred described subjective quality according to the difference between, deterioration signal of video signal and reference images signal any one at least among the spatial information of regulation and the temporal information in P.910 at ITU-R Recommendation.

5. the described video quality evaluation device of claim 1 is characterized in that,

Update information storage part (13), as the update information that is used to revise described subjective quality, with the spatial information (SI) that is illustrated in the image state in each frame of image that comprises in the signal of video signal and be illustrated in temporal information (TI) that the image of interframe changes be mapped the storage correction factor (α, β).

6. video quality evaluation device according to claim 1 is characterized in that,

Described subjective quality is inferred portion (11) and is had:

Import described reference images signal (RI) and described deterioration signal of video signal (PI), generate about from the reference images frame of described reference images signal with from the calibration information generating unit (21) of the calibration information that departs from the space of the departing from of time of the deterioration image frame of described deterioration signal of video signal, described reference images frame and described deterioration image frame;

According to described calibration information, eliminating after described spatial offset and described time departs from, calculate the space characteristics amount calculating part (22) of space characteristics amount separately according to the spatial information of the image state of described reference images frame of expression and described deterioration image frame;

According to described calibration information, eliminating after described spatial offset and described time departs from, for described reference images frame and described deterioration image frame, calculate the temporal characteristics amount calculating part (23) of temporal characteristics amount separately according to the temporal information of the image variation of representing interframe separately; With

Infer the portion of inferring (24) of the subjective quality of described deterioration signal of video signal according to described space characteristics amount and described temporal characteristics amount.

7. video quality evaluation method, this method is to be pushed the update information (α of the subjective quality of making having with the signal of video signal characteristic quantity deterioration signal of video signal (PI) of storage after being used to revise reference images signal (RI) as the signal of video signal that worsens and worsening that be mapped, update information storage part (13) β), infer the video quality evaluation method in the video quality evaluation device of subjective quality of described deterioration signal of video signal, it is characterized in that

Import described reference images signal and described deterioration signal of video signal;

Calculate the signal of video signal characteristic quantity of two signals;

Poor according to the signal of video signal characteristic quantity of two signals of calculating inferred the described subjective quality of described deterioration signal of video signal;

From described update information storage part (13) obtain the update information corresponding with the signal of video signal characteristic quantity of the described reference images signal of calculating (α, β);

(α β) revises described subjective quality of inferring out according to the update information that obtains.

Images