CN1893676A - Video evaluation apparatus, spatio-temporal frequency analyzing apparatus, video evaluation method, spatio-temporal frequency analyzing method, video evaluation program, and spatio-temporal frequency - Google Patents

Abstract

An object is to provide a video evaluation apparatus and video evaluation method capable of appropriately evaluating a video consisting of a plurality of frame images, using an objective standard. The object is achieved as follows. A reference characteristic, which is a predetermined spatio-temporal frequency characteristic about a video, is stored, and a video characteristic analyzer 101 calculates a spatio-temporal frequency characteristic of a target video consisting of a plurality of frame images. Then a video evaluation determiner 102 calculates an evaluated value of the target video, based on a relative relation between the calculated spatio-temporal frequency characteristic and the stored reference characteristic. This enables appropriate evaluation of the video.

Description

Image evaluation device, method, the timely spatial-frequency analysis device of program, method, program

Technical field

The present invention relates to estimate image evaluation device, time space frequency analysis device, image evaluation method, time space frequency analysis method, the timely spatial frequency analysis program of image evaluation program of image.

Background technology

" image " is to constitute continuously by " frame " that makes single static portrait.In the image processing that obtains, stores, transmits, shows, encodes and decipher that comprises image, generally the spatial resolution of each frame (the following resolution of singly being called) is as object with image certain or that slowly change.In addition, the difference of the resolution that human eye can clear and definite recognition image.So, resolution different directly related with evaluation mostly to the quality of image, general high-resolution image is preferred.

Wherein, even each frame low image that is resolution, when each frame shows as continuous image, the people of this image can take place to observe, perhaps can be observed the image of the distinctness of image resolution ratio raising, the perhaps more level and smooth such optical illusion phenomenon of sensorimotor, this point is at document (Takeuchi T, Karen K.De Valois, " Sharpening imagemotionbased onspatio-temporal characteristics of human vision ", Human Vision and Electronic Imaging X, March 18,2005; Volume5666, No.5666, p.83-94, on the books in ISBN/ISSN:0-8194-5639-X).This optical illusion phenomenon is called motion sharpens (motion sharpening) phenomenon.

Put down in writing in above-mentioned document, the cause of this motion sharpens phenomenon may be to be included in the size of motion of the subject in the image and the high fdrequency component of spatial frequency.In addition, also record, in the motion sharpens phenomenon, by inserting the high frame of resolution between the image that constitutes at the frame low by resolution, the resolution of the frame that perceived resolution is low when observing as moving image is than original sensation height.Particularly, be high-resolution situation at 50% of the frame that constitutes image, seem that sometimes the resolution of this image is as broad as long by the moving image that high-resolution constitutes with whole frames, in other words, the resolution illusion about the same that occurs Yu feel.

On the other hand, known in human vision property, in the sensitivity of eyes, there is characteristic at the time space frequency of image (below be called the time space frequency characteristic).In other words, in human eye, exist at this highly sensitive character that is included in the specific time space frequency in the image.In addition, the sensitivity of known eyes as human vision property reduces along with the increasing of time space frequency of observed image.Therefore, can think, when the sensitivity of eyes reduces, even the time space frequency of image changes, the also situation of imperceptible its variation sometimes.

Summary of the invention

Yet, in above-mentioned document, about the resolution of when each frame is observed as continuous image, feeling, under the situation of the spatial frequency that reduces each frame (situation that resolution is reduced), in human eye, seem to exist the frequency of which kind of degree that laboratory report is a bit arranged, but about the not record of quality how to evaluate this point of image with resolution of in this human eye, feeling (below be called perceived resolution).In other words, record does not become the result of the evaluation of the different qualities that human eye feels, is what kind of degree in the variation of the time space frequency of image.Therefore, in the technology of putting down in writing in above-mentioned document, the problem of existence is to use the objective benchmark based on human vision property to carry out suitable evaluation to the quality of the image that is made of multiframe.

In addition, method as the quality of estimating image, adducible just like obtaining the method for impairment grade from power spectrum shown in the technology of putting down in writing in the Japanese patent laid-open 11-266469 communique, in the method, can not carry out with above-mentioned motion sharpens phenomenon is the appropriate evaluation of prerequisite.

Therefore, for solving above-mentioned problem, the purpose of this invention is to provide image evaluation device, time space frequency analysis device, image evaluation method, time space frequency analysis method, the timely spatial frequency analysis program of image evaluation program that to use objective benchmark the image that is made of multiframe to be carried out appropriate evaluation.

For solving above-mentioned problem, image evaluation device of the present invention has memory cell, and storage is about the predetermined datum characteristic as the time space frequency characteristic of image; The image characteristics analytic unit is calculated the time space frequency characteristic by the object image of a plurality of image constructions; And the image evaluation determining unit of calculating the evaluation of estimate of above-mentioned object image based on time space frequency characteristic of calculating by above-mentioned image characteristics analytic unit and the relativeness that is stored in the datum characteristic in the said memory cells.

In addition, image evaluation method of the present invention has spatial frequency and calculates the unit, calculates the spatial frequency of a plurality of images; Spatial frequency component arrangement unit, the component transformation that will calculate the spatial frequency of calculating the unit by above-mentioned spatial frequency for and the corresponding data of predetermined benchmark; The time space frequency is calculated the unit, the time by at least one or more space frequency component in the spatial frequency component of above-mentioned spatial frequency component arrangement unit conversion is changed carry out frequency translation and calculate the time space frequency; And frequency characteristic calculates the unit, utilizes by above-mentioned time space frequency and calculates at least one or more high fdrequency component and low frequency component among the time space frequency of calculating the unit, calculates the data of the time space frequency characteristic that can show image.

Utilize this invention, can store predetermined datum characteristic as the time space frequency characteristic about image, output is by the object image of a plurality of image constructions, calculate the time space frequency of object image, according to the time space frequency characteristic of calculating and the relativeness of datum characteristic, determine the evaluation of estimate of the object image of input.Therefore, can carry out, and the image evaluation in the image be compared, can carry out appropriate evaluation image based on evaluation as the characteristic of the image of consecutive image.Therefore,, obtain high praise sometimes, can estimate rightly as image as image even exist more than of images in the image owing to reasons such as resolution is low are estimated the situations of step-down.

In addition, image evaluation device of the present invention, preferably also has the frequency component adjustment unit, based on the evaluation of estimate of determining by above-mentioned image evaluation determining unit,, adjust the frequency component of regulation to the image that the N between any two images in above-mentioned object image opens, wherein, N is the integer more than or equal to 1, and above-mentioned image evaluation determining unit is calculated the evaluation of estimate of utilizing the image that said frequencies component adjustment unit adjusts signal frequency component.

Utilize this invention, can be according to the evaluation of estimate of determining, N between any two images in the object image is opened the image of (N is the integer more than or equal to 1), adjust the frequency component of regulation, calculate the evaluation of estimate that signal frequency component is opened image through the above-mentioned N that adjusts.Therefore, can revalue, can estimate the image of all frequency components from all angles to the image of frequency component through the part of adjustment.

In addition, in image evaluation device of the present invention, above-mentioned image characteristics analytic unit preferably includes: spatial frequency is calculated the unit, calculates the spatial frequency of a plurality of images; Spatial frequency component arrangement unit, the component transformation that will calculate the spatial frequency of calculating the unit by above-mentioned spatial frequency for and the corresponding data of predetermined benchmark; The time space frequency is calculated the unit, the time by at least one or more space frequency component in the spatial frequency component of above-mentioned spatial frequency component arrangement unit conversion is changed carry out frequency translation and calculate the time space frequency; And characteristic calculates the unit, utilizes by above-mentioned time space frequency and calculates high fdrequency component and low frequency component among the time space frequency of calculating the unit, calculates the data of the time space frequency characteristic that can show image.

Utilize this invention, calculate the spatial frequency of a plurality of images, with the component transformation of the spatial frequency calculated for and the corresponding data of predetermined datum characteristic.So, can change time and carry out frequency translation through at least one or more space frequency component in the spatial frequency component of conversion, calculate the time space frequency, among the time space frequency of calculating, use high fdrequency component and low frequency component to calculate the data of the time space frequency characteristic that can show image.Therefore, can show the time space frequency characteristic of image rightly, can carry out, and the image evaluation in the image be compared, can carry out appropriate evaluation image based on evaluation as the characteristic of the image of consecutive image.

In addition, image evaluation device of the present invention has memory cell, and storage is about the predetermined datum characteristic as the time space frequency characteristic of image; The 1st image characteristics analytic unit is calculated based on the 1st image characteristics by the time space frequency characteristic of the object image of a plurality of image constructions; The 2nd image characteristics analytic unit is calculated based on the 2nd image characteristics that above-mentioned object images is limited the time space frequency characteristic in the resulting frequency component restriction of a part of frequency component image; And image evaluation determining unit, based on the variable quantity of the 2nd image characteristics of calculating by above-mentioned the 2nd image characteristics analytic unit of above-mentioned the 1st image characteristics of relatively calculating by above-mentioned the 1st image characteristics analytic unit be stored in the relativeness of the datum characteristic in the said memory cells, calculate the evaluation of estimate of the above-mentioned object image of being imported.

In addition, the image evaluation method has the 1st image characteristics analytical procedure, calculates based on the 1st image characteristics by the time space frequency characteristic of the object image of a plurality of image constructions; The 2nd image characteristics analytical procedure, the 2nd image characteristics of calculating the time space frequency characteristic in the frequency component restriction image that obtains based on above-mentioned object images is limited a part of frequency component; And the image evaluation determining step, based on the variable quantity of the 2nd image characteristics of above-mentioned relatively the 1st image characteristics with about the relativeness of the predetermined datum characteristic as the time space frequency characteristic of image, calculate the evaluation of estimate of the above-mentioned object image of input.

Utilize this invention, can store predetermined datum characteristic as the time space frequency characteristic about image, output is by the object image of a plurality of image constructions, calculate the 1st image characteristics, calculate the 2nd image characteristics of the time space frequency characteristic in the frequency component restriction image that obtains based on object images is limited a part of frequency component and calculate the evaluation of estimate of the object image of input according to the variable quantity of the 2nd image characteristics of relative the 1st image characteristics and the relativeness of datum characteristic based on the time space frequency characteristic of object image.

Therefore, a part of frequency component of restriction can be used and the frequency component that obtains restriction image carries out the evaluation to the object image, the difference on the frequency component in the image that the frequency component restriction of time space frequency characteristic causes can be distinguished.Therefore, can limit image according to frequency component estimates rightly to the characteristic of the frequency component of the image that will estimate.Such as, frequency component restriction image is become under restriction people's the situation of image of the appreciable frequency component of vision, by estimating, can carry out the evaluation of image of characteristic of relevant people's the imperceptible frequency component of vision according to the variable quantity between this frequency component restriction image and the object image and the relativeness of datum characteristic.

In addition, above-mentioned the 2nd image characteristics analytic unit of image evaluation device of the present invention preferably limits a part of frequency component by the image of the N between any two images in above-mentioned object image being opened (N is the integer more than or equal to 1) and obtains frequency component restriction image for preferred.

Utilize the present invention, can open (N is the integer more than or equal to 1) image, obtain limiting the frequency component restriction image of frequency component that N opens the part of image the N between any two images in the object image.Therefore, can carry out the evaluation of object image, can consider that a part of N in the object image opens the evaluation of the frequency component of image according to this frequency component that is restricted restriction image.

In addition, image evaluation device of the present invention preferably also has the frequency component adjustment unit, based on above-mentioned evaluation of estimate the frequency component that is limited by above-mentioned the 2nd image characteristics analytic unit is adjusted; Above-mentioned image evaluation determining unit is calculated the evaluation of estimate of the image of signal frequency component being adjusted by said frequencies component adjustment unit.Utilize this invention, can be according to evaluation of estimate, the characteristic of the frequency component that is restricted is adjusted, can the characteristic through the frequency component adjusted be revalued.

In addition, image evaluation device of the present invention preferably also has frequency component limit amount adjustment unit, based on above-mentioned evaluation of estimate, changes any two images of being determined the restriction object of frequency component by above-mentioned the 2nd image characteristics analytic unit.Utilize this invention, can be according to evaluation of estimate, any two images of the restriction object that becomes frequency component are changed, can revalue according to image through the restriction object that changes.

In addition, in image evaluation device of the present invention, the formation of best above-mentioned the 1st image characteristics analytic unit has spatial frequency and calculates the unit, calculates the spatial frequency of a plurality of images; Spatial frequency component arrangement unit, the component transformation that will calculate the spatial frequency of calculating the unit by above-mentioned spatial frequency for and the corresponding data of predetermined benchmark; The time space frequency is calculated the unit, the time by at least one or more space frequency component in the spatial frequency component of above-mentioned spatial frequency component arrangement unit conversion is changed carry out frequency translation and calculate the time space frequency; And characteristic calculates the unit, utilizes by above-mentioned time space frequency and calculates high fdrequency component and low frequency component among the time space frequency of calculating the unit, calculates the data of the time space frequency characteristic that can show image.

Utilize this invention, calculate the spatial frequency of a plurality of images, with the component transformation of the spatial frequency calculated for and the corresponding data of predetermined datum characteristic.So, can change time and carry out frequency translation through at least one or more space frequency component in the spatial frequency component of conversion, calculate the time space frequency, among the time space frequency of calculating, use high fdrequency component and low frequency component to calculate the data of the time space frequency characteristic that can show image.Therefore, can show the time space frequency characteristic of image rightly, can carry out, and the image evaluation in the image be compared, can carry out appropriate evaluation image based on evaluation as the characteristic of the image of consecutive image.

In addition, in image evaluation device of the present invention, preferably the formation of above-mentioned the 2nd image characteristics analytic unit comprises that spatial frequency calculates the unit, from the frequency component restriction image that the frequency component of an object images restriction part is obtained, calculates spatial frequency; To calculate the component transformation of the spatial frequency of calculating the unit for putting the unit in order by above-mentioned spatial frequency with the spatial frequency component of the corresponding data of predetermined benchmark, the time space frequency is calculated the unit, time of at least one or more space frequency component in the spatial frequency component that utilizes above-mentioned spatial frequency component arrangement unit conversion is changed carry out frequency translation and calculate the time space frequency; And characteristic calculates the unit, utilizes by above-mentioned time space frequency and calculates high fdrequency component and low frequency component among the time space frequency of calculating the unit, calculates the data of the time space frequency characteristic that can show image.

Utilize this invention, the frequency component that obtains from object images is limited a part of frequency component restriction image is calculated spatial frequency, with the component transformation of the spatial frequency calculated for and the corresponding data of predetermined datum characteristic.So, can change time and carry out frequency translation through at least one or more space frequency component in the spatial frequency component of conversion, calculate the time space frequency, among the time space frequency of calculating, use high fdrequency component and low frequency component to calculate the data of the time space frequency characteristic that can show image.Therefore, can show the time space frequency characteristic of image rightly, can carry out, and the image evaluation in the image be compared, can carry out appropriate evaluation image based on evaluation as the characteristic of the image of consecutive image.

In addition, the said reference characteristic of image evaluation device of the present invention, the preferably frequency characteristic of determining based on human vision property.Utilize this invention, by with datum characteristic as the frequency characteristic definite according to human vision property, can carry out evaluation based on human vision property.

In addition, time space frequency-response analysis device of the present invention has spatial frequency and calculates the unit, calculates the spatial frequency of a plurality of images; Spatial frequency component arrangement unit, the component transformation that will calculate the spatial frequency of calculating the unit by above-mentioned spatial frequency for and the corresponding data of predetermined benchmark; The time space frequency is calculated the unit, the time by at least one or more space frequency component in the spatial frequency component of above-mentioned spatial frequency component arrangement unit conversion is changed carry out frequency translation and calculate the time space frequency; And frequency characteristic calculates the unit, utilizes by above-mentioned time space frequency and calculates at least one or more high fdrequency component and low frequency component among the time space frequency of calculating the unit, calculates the data of the time space frequency characteristic that can show image.

Utilize this invention, calculate the spatial frequency of a plurality of images, with the component transformation of the spatial frequency calculated for and the corresponding data of predetermined datum characteristic.So, can change time and carry out frequency translation through at least one or more space frequency component in the spatial frequency component of conversion, calculate the time space frequency, the data that high fdrequency component among the time space frequency that use is calculated and low frequency component are calculated the time space frequency characteristic that can show image.Therefore, can show the time space frequency characteristic of image rightly, can carry out, and the image evaluation in the image be compared, can carry out appropriate evaluation image based on evaluation as the characteristic of the image of consecutive image.

In addition, as described below, can also record and narrate as the invention of the timely spatial frequency characteristic routine analyzer of image evaluation program.This is the category difference, comes down to same invention, can reach same effect and effect.

Image evaluation program of the present invention has the memory module of storage about the predetermined datum characteristic as the time space frequency characteristic of image, calculates by the image characteristics analysis module of the time space frequency characteristic of the object image of a plurality of image constructions and based on the time space frequency characteristic of being calculated by above-mentioned image characteristics analysis module be stored in the image evaluation determination module that the relativeness of the datum characteristic in the above-mentioned memory module is calculated the evaluation of estimate of above-mentioned object image.

In addition, image evaluation program of the present invention has memory module, and storage is about the predetermined datum characteristic as the time space frequency characteristic of image; The 1st image characteristics analysis module is calculated based on the 1st image characteristics by the time space frequency characteristic of the object image of a plurality of image constructions; The 2nd image characteristics analysis module, the 2nd image characteristics of calculating the time space frequency characteristic in the frequency component restriction image that obtains based on above-mentioned object images is limited a part of frequency component; And image evaluation determination module, based on the variable quantity of the 2nd image characteristics of calculating by above-mentioned the 2nd image characteristics analysis module of the 1st image characteristics of relatively calculating by above-mentioned the 1st image characteristics analysis module be stored in the relativeness of the datum characteristic in the above-mentioned memory module, calculate the evaluation of estimate of the above-mentioned object image of input.

In addition, time space frequency-response analysis program of the present invention, its formation has spatial frequency and calculates module, calculates the spatial frequency of a plurality of images; The spatial frequency component sorting module, the component transformation that will calculate the spatial frequency that module calculates by above-mentioned spatial frequency for and the corresponding data of predetermined benchmark; The time space frequency is calculated module, the time by at least one or more space frequency component in the spatial frequency component of above-mentioned spatial frequency component sorting module conversion is changed carry out frequency translation and calculate the time space frequency; And frequency characteristic calculates module, utilizes by above-mentioned time space frequency and calculates high fdrequency component and low frequency component among the time space frequency that module calculates, calculates the data of the time space frequency characteristic that can show image.

The present invention can carry out based on the evaluation as the characteristic of the image of consecutive image, and the image evaluation in the image is compared, and can carry out appropriate evaluation to image.Therefore,, obtain high praise sometimes, can estimate rightly as image as image even exist more than of images in the image owing to reasons such as resolution is low are estimated the occasions of step-down.

In addition, the present invention can use a part of frequency component of restriction and the frequency component that obtains restriction image carries out the evaluation to the object image, can distinguish the difference on the frequency component in the image that the frequency component restriction of time space frequency characteristic causes.Therefore, can limit image according to frequency component estimates rightly to the characteristic of the frequency component of the image that will estimate.Such as, make frequency component restriction image become the occasion of the image of the appreciable frequency component of vision that limits the people, by estimating, can carry out the evaluation of image of characteristic of relevant people's the imperceptible frequency component of vision according to the variable quantity between this frequency component restriction image and the object image and the relativeness of datum characteristic.

Description of drawings

Fig. 1 carries out illustrative block diagram to the formation of the function formation of the image evaluation device of execution mode 1.

Fig. 2 is the concept map that the two field picture that reads in is shown.

Fig. 3 is for to carry out illustrative diagrammatic sketch to the relation of power spectrum in the frequency field and spatial frequency.

Fig. 4 is the concept map that the notion when two field picture P0～P3 extracts two dimensional image in the horizontal direction is shown.

Fig. 5 is the concept map that the notion when two field picture P0～P3 extracts two dimensional image in vertical direction is shown.

Fig. 6 is the flow chart that the flow process of the image evaluation processing in the execution mode 1 is shown.

The formation block diagram that Fig. 7 constitutes for the function of image evaluation device that execution mode 2 is shown for example.

Fig. 8 is the diagrammatic sketch of example that the processing of spatial frequency is shown.

Fig. 9 is the flow chart that the flow process of the image processing in the execution mode 2 is shown.

The formation block diagram that Figure 10 constitutes for the function of image processor that execution mode 3 is shown for example.

Figure 11 is the flow chart that the flow process of the image processing in the execution mode 3 is shown.

The formation block diagram that Figure 12 constitutes for the function of time space frequency-response analysis device that execution mode 4 is shown for example.

Figure 13 illustrates the flow chart that the image evaluation of the time space frequency-response analysis device of execution mode 4 is handled.

Figure 14 is for realizing the module structure drafting with the program P10 of image evaluation device 10 identical functions.

Figure 15 is for realizing the module structure drafting with the program P50 of image evaluation device 50 identical functions.

Figure 16 is for realizing the module structure drafting with the program P80 of image processor 80 identical functions.

Figure 17 is for realizing the module structure drafting with the program P1000 of time space frequency-response analysis device 100 identical functions.

Embodiment

With reference to the accompanying drawings each execution mode of image evaluation device of the present invention is described below.In addition, in each figure, give prosign for same parts, the repetitive description thereof will be omitted.

[execution mode 1]

At first, embodiments of the present invention 1 are described.The formation block diagram that Fig. 1 constitutes for the function of image evaluation device 10 that execution mode 1 is shown for example.This image evaluation device 10 is to have storage devices such as CPU (central processing unit), memory physically and be used for the computer of input/output unit (communicator etc.) of input and output image.Such as, image evaluation device 10 are devices of the information processing carried out of mobile communication terminals such as the fixed communication terminal of personal computer etc. or portable telephone etc.

With reference to Fig. 1 the function formation of image evaluation device 10 is described.As shown in Figure 1, the formation of image evaluation device 10 comprises image characteristics analysis portion 101 (image characteristics analytic unit), image evaluation determination portion 102 (image evaluation determining unit) and datum characteristic storage part 103 (memory cell).

Image characteristics analysis portion 101 is that the input signal of video signal 104 that will import as motion image signal from the outside is decomposed into two field picture, based on a plurality of two field pictures that decompose, calculates the part about the characteristic quantity of the time space frequency characteristic of input signal of video signal.

Herein, with reference to Fig. 2,, the relevant method of calculating the characteristic quantity of time space frequency characteristic is specifically described as the analysis of time space frequency characteristic.Fig. 2 is the concept map of the image that is made of a plurality of frames.In addition, for convenience of description, the order according to from the outside input describes continuous 4 two field pictures as two field picture P0～P3.4 continuous two field picture P0～P3 from input signal of video signal 104 decomposes read in preservation by image characteristics analysis portion 101 orders.So the motion feature amount in 4 of two field picture P0～P3 is calculated by image characteristics analysis portion 101.Below the detailed process till calculating the motion feature amount is described.

At first, image characteristics analysis portion 101, order is read in continuous 2 two field pictures.Image characteristics analysis portion 101, based between 2 two field pictures that read in mutually between consistent degree more than or equal to the so-called similar picture signal pattern mutually of setting, calculate motion vector (MVbx, MVby) as the space addendum modification between two two field pictures.

In the present embodiment, image characteristics analysis portion 101, two field picture is divided into a plurality of segments of prescribed level, the exploration of use figure BMA etc. is handled from two field picture P0, each segment at two field picture P1, explore with the picture signal pattern of this each segment the most similar (such as, consistent degree is more than or equal to setting) the picture signal pattern.So image characteristics analysis portion 101 is illustrated in the segment of picture signal pattern unanimity between two field picture P0 and the P1 by judgement, in two field picture, departs from the bias of much positions and calculate motion vector.

Thereafter, image characteristics analysis portion 101, according to the motion vector of calculating in the full figure piece (MVbx, MVby), the mean value of the motion vector by calculating each segment is calculated all motion vector Vi of frame (MVbx, MVby).Image characteristics analysis portion 101, also average the and frame rate f of the size by utilizing 3 Vi between 4 frames calculates, and calculates motion feature amount V according to formula (1).

[formula 1]

In following formula 1, i is the numbering of expression frame.And α is a constant.

In addition, in the present embodiment, be to calculate motion vector from continuous 2 frames, calculate motion vector but also can use more than or equal to 2 frames.In addition, also can use mean value, maximum, the minimum value of the motion vector that obtains from many frames.

In addition, in the present embodiment, being to use of illustrating utilized the motion feature amount of the value that all motion vector Vi of frame calculate, but also can use the motion feature amount of the motion vector of obtaining in the full figure piece (MVbx, MVby).In addition, also can use,, replace the mean value of the size in the motion vector such as, any one in maximum, median or the minimum value of the size (MVbx, MVby) of the motion vector that each segment that is included in the two field picture is obtained.

In addition, in the present embodiment, be to utilize the size of the motion vector of entire image to calculate the motion feature amount, but in addition also can utilize the motion characteristics amount of watching image attentively, such as, utilize the direction poor, each motion vector of distribution, each motion vector size of the motion vector that segment had in expression travel direction, the two field picture poor, calculate the motion feature amount more than or equal to the amount of the variation of the time orientation of the ratio of the motion vector of the size of setting.

In addition, the phase place in the time of also can using the frequency translation of image for the motion feature amount is calculated.In this case, also can carry out discrete Fourier transform or the like respectively by whole pixels to 2 two field pictures, carry out the conversion from the pixel space to the frequency space, same frequency component to each frame is implemented division, use has utilized the amount of movement of the size of phase place, calculates the size of the phase place of each frequency component.So, after calculating the phase place of each frequency component, the value of whole phase places is added up to, as the motion feature amount.In addition, the utilization of the phase place after the frequency translation is not limited thereto, so long as represent the momental amount of 2 interframe, how to calculate can.

Afterwards, image characteristics analysis portion 101 is read in the two field picture P0 that is preserved, and calculates the spatial frequency characteristic quantity R of conduct about the characteristic quantity of the spatial frequency characteristic of two field picture P0.In specific words, at first, image characteristics analysis portion 101 is carried out discrete Fourier transform by the whole pixels to two field picture, is transformed to the frequency space from pixel space.In the present embodiment, image characteristics analysis portion 101 is carried out two dimensional discrete Fourier transform to two field picture, two field picture is transformed to the data mode with spatial frequency zone.So image characteristics analysis portion 101 is calculated spatial frequency characteristic quantity about the frequency component of the spatial frequency of the image in the frequency space according to formula (2).

Herein, spatial frequency characteristic quantity R is the characteristic quantity of the characteristic of the frequency component of expression when being transformed to the frequency space of image.Use is calculated this spatial frequency characteristic quantity R by the weighting function that big or small Ps showed (formula (2)) of the energy of the spatial frequency S of configuration frame image and each spatial frequency S by image characteristics analysis portion 101.

[formula 2]

In following formula 2, i represents frame number, j representation space frequency.In addition, β _jExpression is to the weight coefficient of each spatial frequency S.

Spatial frequency characteristic to image describes below.Fig. 3 is the key diagram of the spatial frequency characteristic of the relation of the big or small P of energy that the spatial frequency in the one dimension is shown for example and spatial frequency S, Fig. 3 (a) is the key diagram of the spatial frequency characteristic of an image of expression, and Fig. 3 (b) is the key diagram of the appreciable spatial frequency characteristic of expression human eye.As shown in Figure 3, the spatial frequency characteristic of image, at varying in size of each its energy of spatial frequency, and the spatial frequency characteristic that human eye can be felt also is varying in size at each its energy of spatial frequency.

As mentioned above, image characteristics analysis portion 101 is calculated the spatial frequency characteristic quantity R of two field picture P0, and is identical therewith, also calculates spatial frequency characteristic quantity R for the two field picture P1～P3 that preserves.

In addition, in the present embodiment, spatial frequency characteristic quantity R is to calculate in each frame, but also can use the spatial frequency of calculating in many frame to calculate spatial frequency characteristic quantity R.Such as, also can use the mean value of many frame spatial frequencys and ranking operation or the like.

In addition, in the present embodiment, be to calculate the spatial frequency characteristic quantity according to the weighting function of the big or small Ps of spatial frequency S and this energy, but so long as the function of the characteristic of the frequency component of the spatial frequency of the frequency field of presentation video also can be calculated the spatial frequency characteristic quantity according to any type of function.

In addition, in the present embodiment, in calculating the spatial frequency characteristic quantity, be that example illustrates in the entire frame image, to carry out discrete Fourier transform, but also can use one-dimensional discrete Fourier transform, be divided into and repeatedly be transformed to frequency field all directions of image.In addition, also can not be transformed to frequency field, from pixel space, such as, use computing differential value etc. is obtained the characteristic of the frequency component of spatial frequency.

Afterwards, image characteristics analysis portion 101 is calculated the time orientation frequecy characteristic amount of two field picture P0～P3.In specific words, image characteristics analysis portion 101 is extracted the two dimensional image that is made of horizontal direction and time-axis direction by horizontal direction 1 row * time-axis direction 4 row formations.Notion during herein, to the extraction two dimensional image describes.Fig. 4 is the concept map that the notion when two field picture P0～P3 extracts two dimensional image in the horizontal direction is shown.As shown in Figure 4, extract the image P01 of the horizontal direction 1 row size among the two field picture P0.Equally, extract image P11, P21, the P31 of the horizontal direction 1 row size among two field picture P1, P2, the P3, generate two dimensional image P100.Afterwards, extract image P02, P12, P22, P32, generate two dimensional image P200.Below, same, order generates two dimensional image.Image characteristics analysis portion 101 is carried out two dimensional discrete Fourier transform to the two dimensional image that extracts, and two dimensional image is transformed to frequency field.

Like this, image characteristics analysis portion 101, the same two dimensional image that is made of horizontal direction and time-axis direction that extracts is transformed to frequency field in the position under 1 pixel on the vertical direction of direction in space.Image characteristics analysis portion 101 repeats this processing on the entire frame image.

Image characteristics analysis portion 101 after the processing to the vertical direction of two field picture finishes, is calculated average level frequency component TFBh as the full extraction two dimensional image frequency component of the frequency component in the whole two dimensional images that extract according to formula (3).

[formula 3]

P _hThe size of the energy of the frequency component that expression comprises.In addition, Sh represents frequency.φ _mExpression is defined as calculating the value of mean value to the weight coefficient of each frequency component.

Afterwards, image characteristics analysis portion 101 is extracted the two dimensional image that is made of vertical direction and time-axis direction by vertical direction 1 row * time-axis direction 4 row formations from the left end of image.Notion during herein, to the extraction two dimensional image describes.Fig. 5 is the concept map that the notion when two field picture P0～P3 extracts two dimensional image in vertical direction is shown.As shown in Figure 5, extract the image P01a of the vertical direction 1 row size among the two field picture P0.Equally, extract image P11a, P21a, the P31a of the vertical direction 1 row size among two field picture P1, P2, the P3, generate two dimensional image P100a.Afterwards, extract image P02a, P12a, P22a, P32a, generate two dimensional image P200a.Below, order generates two dimensional image.

Image characteristics analysis portion 101 is carried out two dimensional discrete Fourier transform to the two dimensional image that extracts, and two dimensional image is transformed to frequency field.Like this, image characteristics analysis portion 101 is transformed to frequency field from 1 the same extraction of locations of pixels of advancing by the two dimensional image that vertical direction and time-axis direction constitute to the right on the horizontal direction of direction in space.Image characteristics analysis portion 101 repeats this processing on the entire frame image.Image characteristics analysis portion 101 after the processing to the horizontal direction of two field picture finishes, is calculated average vertical frequency component TFBv as the full extraction two dimensional image frequency component of the frequency component in the whole two dimensional images that extract according to formula (4).

[formula 4]

P _vThe size of the energy of the frequency component that expression comprises.In addition, Sv represents frequency.μ _mExpression is defined as calculating the value of mean value to the weight coefficient of each frequency component.

Afterwards, image characteristics analysis portion 101, the average level frequency component TFB that calculates by calculating _hWith average vertical frequency component TFB _vWeighting function, carry out calculating of time orientation frequecy characteristic amount C.In specific words, image characteristics analysis portion 101 is calculated time orientation frequecy characteristic amount C according to following formula (5).

[formula 5]

Time orientation frequecy characteristic amount C=M (TFB _h, TFB _v)=δ * TFB _h+ ε * TFB _v(5)

In above-mentioned formula, in formula (5), δ and ε are weight coefficients.

In addition, in the present embodiment, be by calculating average level frequency component TFB _hWith average vertical frequency component TFB _vWeight coefficient, carry out calculating of time orientation frequecy characteristic amount, but so long as use average level frequency component TFB _hWith average vertical frequency component TFB _vFunction, be that the function of which kind of form can.

Like this, motion feature amount V, the spatial frequency characteristic quantity R and the time orientation frequecy characteristic amount C that utilize image characteristics analysis portion 101 to calculate output to image evaluation determination portion 102 as the characteristic quantity 105 about the time space frequency characteristic of image.This moment spatial frequency characteristic, in its notion, so long as can show the parameter of the variation in the time orientation of the feature of representing by the spatial frequency of image, be not limited to characteristic quantity 105, not some also passable among motion feature amount V, spatial frequency characteristic quantity R and time direction frequecy characteristic amount C.

In addition, in the present embodiment, the mean value of calculating full extraction two dimensional image frequency component is as average level frequency component TFB _hWith average vertical frequency component TFB _v, but also can only calculate mean value to the characteristic image among the frequency component of each two dimensional image.Such as, the frequency that also can calculate among the full extraction two dimensional image frequency component constitutes more than or equal to the mean value in the image of threshold value.In addition, such as, also can calculate the high more weighted average of center weight of the approaching more image of being gazed at easily.

In addition, in the present embodiment, what illustrate is the two dimensional images of 4 row that extract horizontal direction one row (or vertical direction delegation) * time orientation of direction in space carry out frequency conversion process to entire image method, but in addition, such as, also two field picture can be divided into the figure block unit, it is carried out frequency conversion process to the time orientation extraction and to the result who generates as two dimensional image.

In addition, in the present embodiment, what illustrate is the method for repetition towards the conversion of the two-dimentional frequency field of time orientation, but also can carry out conversion to three-dimensional frequency field.In addition, also can use one-dimensional Fourier transform to time orientation.In addition, also can not be transformed to frequency field, and from pixel space, such as, use computing differential value or the like is obtained the characteristic to the frequency component of time orientation.

In addition, in the present embodiment, be to use the function in the frequency component of time orientation of image, but also can use based on the variation of expression to time orientation, such as, wait characterization with the variation of the correlation of the pixel in time orientation and the function that is changed to the basis of spatial frequency characteristic quantity.

In addition, in the present embodiment, be to carry out calculating and in time orientation frequecy characteristic amount, carrying out discrete Fourier transform and carry out of spatial frequency characteristic quantity, but the method that is transformed to frequency component also can be used additive methods such as discrete cosine transform, wavelet transform about the calculating of the characteristic quantity of the frequency component of frequency field.

In addition, the number of frame of calculating motion feature amount and time direction frequecy characteristic amount is so long as continuous 2 or more, and what use open can.

In addition, in the present embodiment, as the time space frequency characteristic of image, use the characteristic quantity of the characteristic of the characteristic be directed to motion that image has and spatial frequency and time space frequency, but in addition also can utilize the characteristic quantity that is associated with the time space frequency characteristic of image.

Then, carry out the explanation of figure.Image evaluation determination portion 102, according to the characteristic quantity 105 that receives from image characteristics analysis portion 101 with deposit in datum characteristic storage part 103 the characteristic quantity 106 of time space frequency characteristic of human vision property of the relevant benchmark that becomes relative evaluation, calculate the evaluation of estimate 107 of image of relativeness of the time space frequency characteristic of the time space frequency characteristic that has based on image and human eye.

Below, be specifically described to the time space frequency characteristic of calculating the object image with at the method for the evaluation of estimate of the relativeness of the visual characteristic of image with reference to Fig. 3.At first, image evaluation determination portion 102 is calculated poor about the characteristic quantity 105 of the time space frequency characteristic of image and the characteristic quantity 106 about spatial frequency characteristic (being stored in the datum characteristic in the datum characteristic storage part 103) that human eye can be felt of being calculated by image characteristics analysis portion 101.So, image evaluation determination portion 102, Δ R, time orientation frequecy characteristic amount C and motion feature amount V according to the difference of frequency component that has based on image and the appreciable frequency component of human eye, utilize predetermined function, calculate evaluation of estimate 107 based on the relativeness of the time space frequency characteristic of time space frequency characteristic and human eye.In the present embodiment, according to formula (6), image evaluation determination portion 102, the Δ R of difference that calculates the characteristic of the frequency component that has based on image and the appreciable frequency component of human eye.So according to formula (7), image evaluation determination portion 102 is calculated evaluation of estimate D.

[formula 6]

$\mathrm{\ΔR}=\underset{i=0}{\overset{3}{\mathrm{\Σ}}}(F(P_s,S)-H(P_s,S))---\left(6\right)$

In formula (6), H (Ps, S) characteristic quantity 106 of the time space frequency characteristic that can feel of the relevant human eye of expression.In addition, in formula (6), frame number i is from 0 to 3, and this frame number i is the amount that is changed by the specific mode of two field picture.

[formula 7]

In formula (7), ρ, ω represent weight coefficient.

In addition, (Ps S) might not be certain certain characteristic quantity, also can be the characteristic quantity 106 of the relevant time space frequency characteristic determined by characteristic quantity 105 as the H of the characteristic of the appreciable frequency component of human eye.

In addition, (Ps S), uses expression to the characteristic about the sensitivity of the frequency of the time space of the quality of image, but is not limited thereto the H of the characteristic of the appreciable frequency component of human eye.Such as, also can use the characteristic of a certain specific quality of images such as characteristic of the sensitivity of spot to image, torsion resistance, flicker.In addition, in this case, evaluation of estimate D can calculate the value of expression to a certain specific quality evaluation of image.

In addition, in the present embodiment, image evaluation determination portion 102, be to use difference delta R to calculate evaluation of estimate D based on characteristic quantity 105 and characteristic quantity 106, but so long as the function of calculating by comparative feature amount 105 and characteristic quantity 106, not necessarily must be poor, also can be any function (such as, the expression ratio function).

In addition, in the present embodiment, the characteristic quantity 105 that uses in calculating evaluation of estimate D uses the poor Δ R and the time direction frequecy characteristic amount C of motion feature amount V, spatial frequency characteristic quantity, but some also passable less than wherein.

Below the image evaluation in the image evaluation device 10 of execution mode 1 is handled and described.Fig. 6 is the flow chart that the image evaluation processing of image evaluation device 10 is shown.

At first, continuous 4 two field pictures (two field picture P0 is to two field picture P3) (step S401) that decompose from input signal of video signal 104 are read in and preserved to image characteristics analysis portion 101, order.

Below, image characteristics analysis portion 101, the motion vector of calculating each interframe according to two field picture P0～two field picture P3 is calculated the mean value of the motion vector of two field picture P0 to P3 as the motion feature amount.Image characteristics analysis portion 101 is applied to the mean value and the frame rate f of motion vector above-mentioned formula (1) and calculates motion feature amount V (step time space frequency characteristic 402).

Below, image characteristics analysis portion 101 is carried out frequency translation to the pixel data of the two field picture P0 that reads in, and two field picture is transformed to frequency space (step S403) from pixel space.According to the spatial frequency that obtains through conversion, image characteristics analysis portion 101 is calculated spatial frequency characteristic quantity R (step S404) about the characteristic of the frequency component of the spatial frequency of two field picture P0 according to above-mentioned formula (2).Image characteristics analysis portion 101 repeats above processing (S403, S404) (S405) to two field picture P1～two field picture P3.

When the processing to whole two field picture P0～P3 finished, image characteristics analysis portion 101 was according to the average level frequency component TFB of two field picture P0～two field picture P3 _hAnd average vertical frequency component TFB _vCalculate time orientation frequecy characteristic amount C (step S406).

Motion feature amount V, spatial frequency characteristic quantity R that these are calculated and time direction frequecy characteristic amount C are input to the characteristic quantity 105 of image evaluation determination portion 102 conducts about the time space frequency characteristic of image.

Then, image evaluation determination portion 102 is read in the characteristic quantity 106 (step S407) of the relevant time space frequency characteristic as human vision property of depositing in datum characteristic storage part 103.So, image evaluation determination portion 102, characteristic quantity 106 that relatively reads in and characteristic quantity 105 are calculated the evaluation of estimate 107 (evaluation of estimate D) (step S408) of the relativeness that causes based on the time space frequency characteristic by the time space frequency characteristic that has of object image of input and human eye.

As mentioned above, utilize the image evaluation device 10 in the execution mode 1, can calculate the characteristic quantity of the time space frequency characteristic of relevant image, compare with the relevant predetermined characteristic quantity that becomes the time space frequency characteristic of benchmark by the characteristic quantity of will calculate, can calculate evaluation of estimate based on the relativeness of the time space frequency characteristic that has of object image of input and the predetermined time space frequency characteristic that becomes benchmark.So, according to the time space frequency characteristic of input signal of video signal, with the comparison of the predetermined time space frequency characteristic that becomes benchmark in appropriate evaluation of estimate can be shown.Such as, under the situation of the time space frequency characteristic that has as human vision property with the predetermined time space frequency characteristic that becomes benchmark, can illustrate rightly based on the evaluation of estimate of considering the relativeness of the influence of human eye.

[execution mode 2]

Secondly, embodiments of the present invention 2 are described.The formation block diagram that Fig. 7 constitutes for the function of image evaluation device 50 that execution mode 2 is shown for example.This image evaluation device 50 is to have storage devices such as CPU (central processing unit), memory physically and be used for the computer of input/output unit (communicator etc.) of input and output image.Such as, image evaluation device 50 are devices of the information processing carried out of mobile communication terminals such as the fixed communication terminal of PC terminal etc. or portable telephone etc.

Secondly, with reference to Fig. 7 the function formation of image evaluation device 50 is described.As shown in Figure 7, the formation of image evaluation device 50 comprises that image characteristics analysis portion 501 (the 1st image characteristics analytic unit), frequency component restriction image adds the Ministry of worker 502 (the 2nd image characteristics analytic unit), time space characteristic variations amount is calculated portion 505, image evaluation determination portion 506 (image evaluation determining unit) and datum characteristic storage part 507.In addition, frequency component restriction image adds the Ministry of worker 502 and comprises the frequency component restriction and add the Ministry of worker 503 and image characteristics analysis portion 504.Below each formation is described.

Image characteristics analysis portion 501 is the parts that have with image characteristics analysis portion 101 said functions described in the above-mentioned execution mode 1.In other words, image characteristics analysis portion 501 is calculated the characteristic quantity 510 of time space frequency characteristic to the image of input signal of video signal 508, the characteristic quantity 510 of calculating is outputed to time space characteristic variations amount calculate portion 505.

Frequency component restriction image adds the Ministry of worker 502, be that the input signal of video signal 508 that is made of a plurality of two field pictures is applied the frequency component restriction, will output to time space characteristic variations amount by the characteristic quantity 511 that the frequency component limited images that has applied the frequency component restriction generates and calculate the part of portion 505.This frequency component restriction image adds the Ministry of worker 502 and comprises the frequency component restriction and add the Ministry of worker 503 and image characteristics analysis portion 504, below these component parts is described.

The frequency component restriction adds the Ministry of worker 503, be that the N between any two two field pictures of the input signal of video signal 508 that is made of a plurality of two field pictures is opened (N is the integer more than or equal to 1) two field picture, follow the processing of the frequency component restriction of a part of signal frequency component of restriction, the part of generated frequency component restriction processing image 509.This frequency component restriction adds the Ministry of worker 503, and the frequency component restriction processing image 509 that generates is outputed to image characteristics analysis portion 504.

More detailed speech it, frequency component restriction adds the Ministry of worker 503, will be decomposed into two field picture from the outside as the input signal of video signal 508 of motion image signal input.So, the frequency component restriction adds the Ministry of worker 503, determine two two field pictures that N opens two field picture that accompany in a plurality of two field pictures that decompose, the N between the two field picture that is clipped in definite regulation is opened two field picture generate the frequency component restriction processing image 509 that limits a part of frequency component at least.The frequency component restriction adds the Ministry of worker 503, and the frequency component restriction processing image 509 that generates is outputed to image characteristics analysis portion 504.In addition, frequency component restriction adds the Ministry of worker 503, also can uncertain two field picture, and to whole two field picture limit frequency components.

With 4 two field pictures be below example describe (such as, with reference to Fig. 2).The frequency component restriction adds the Ministry of worker 503, will read in preservation in proper order from 4 continuous two field pictures that input signal of video signal 508 decomposes.Frequency component restriction adds the Ministry of worker 503, determines two field picture P0 and P3, carries out the restriction of signal frequency component of the two field picture of two field picture P1 and P2.In specific words, the frequency component restriction adds the Ministry of worker 503, reads in two field picture P1 and P2, and whole pixels of two field picture are carried out discrete Fourier transform, carries out the conversion from the pixel space to the frequency space.In addition, the frequency component restriction adds the Ministry of worker 503, stores the multiple frequency component limit amount B that is made of the restriction and the cut-off frequency of energy level (power spectrum) accordingly with the frequency component of each spatial frequency.This frequency component limit amount B, comprise to each frequency component determine frequency energy level the upper limit the energy level restricted information and be used for making the cut-off frequency of high-frequency cut-off.So the frequency component restriction adds the Ministry of worker 503, according to the frequency component of the spatial frequency of passing through conversion, determines a frequency component limit amount B.The frequency component restriction adds the Ministry of worker 503, by according to frequency component limit amount B two field picture P1 and the P2 that is transformed to the frequency space being processed generated frequency component restriction processing image 509.

Below frequency component limit amount B is described.Fig. 8 is the key diagram that the restriction of energy level is shown and limits the appearance of cut-off frequency.As shown in Figure 8, the upper limit of energy level (power spectrum) determined already, has frequency above the energy of this upper limit and adds the Ministry of worker 503 by the frequency component restriction and remove.In addition, as shown in Figure 8,, determine the upper limit (cut-off frequency) of frequency, surpass the frequency of this upper limit, add the Ministry of worker 503 by the frequency component restriction and remove in order to remove high frequency.Like this, can obtain adding the Ministry of worker 503 by frequency component restriction removes above the frequency of the upper limit of energy level and surpasses the frequency space of the frequency of cut-off frequency, generated frequency component restriction processing image 509.

In addition, in the present embodiment, be frequency component, among pre-prepd multiple frequency component limit amount B, determine a frequency component limit amount B, but the mode that also can have nothing to do with the spatial frequency with image is determined frequency component limit amount B according to the spatial frequency of image.In addition, also can automatically select and calculate a frequency component limit amount B according to the characteristic quantity of the spatial frequency of image.

In addition, the frequency component limit amount B of afore mentioned rules reaches about the information of the frame of determining (being used for determining the information of any two two field pictures), for it is imported from the outside, be set by the user, the frequency component restriction adds the Ministry of worker 503 and also can have input/output function.

In addition, in the present embodiment, processing in the frequency field is shown, but also can determines to use the processing of filter etc. as the unit of the frequency component restriction of carrying out spatial frequency.In addition, though what illustrate for example is by determining that cut-off frequency removes the processing of high fdrequency component from the frequency space, also can processing a certain specific frequency.Such as, if filter also can use high pass filter, band pass filter, band stop filter etc. that specific frequency is processed.

In addition, in the present embodiment, be two field picture number N and the frequency component limit amount B that determines respectively between specific two, but also can determine and two field picture number N correspondent frequency component limit amount B.In addition, in the present embodiment, be that two field picture P1 and P2 are determined same frequency component limit amount B, but also can determine because of the different frequency component limit amount B of frame.

Turning back to Fig. 7 then describes.Image characteristics analysis portion 504 is the parts that have with above-mentioned execution mode 1 described image characteristics analysis portion 101 said functions.In other words, image characteristics analysis portion 504 is calculated about frequency component being limited the characteristic quantity 511 of the time space frequency characteristic of the image of processing image 509, outputs to time space characteristic variations amount and calculates the part of portion 505.

Time space characteristic variations amount is calculated portion 505, is the part according to the variable quantity 512 of calculating the time space frequency characteristic about the characteristic quantity 510 and the characteristic quantity 511 of time space frequency characteristic from two of 504 outputs of image characteristics analysis portion 501 and image characteristics analysis portion.In specific words, time space characteristic variations amount is calculated portion 505, calculates based on the Δ R about the difference of the spatial frequency characteristic quantity R (with reference to formula (2)) of the spatial frequency characteristic quantity R (with reference to formula (2)) of the characteristic quantity 510 of time space frequency characteristic and characteristic quantity 511.In addition, time space characteristic variations amount is calculated portion 505, calculates the Δ C based on the difference of the time orientation frequecy characteristic amount C (with reference to formula (5)) of the time orientation frequecy characteristic amount C (with reference to formula (5)) of characteristic quantity 510 and characteristic quantity 511.Such as, time space characteristic variations amount is calculated portion 505, calculates variable quantity 512 according to following formula (8) and formula (9).

[formula 8]

${\mathrm{\&Delta;R}}_m=\underset{i=1}{\overset{2}{\mathrm{\&Sigma;}}}\left(\right|{\mathrm{<figure-callout\; id="510"\; label="F"\; filenames="A20061010009000541.png"\; state="\{\{state\}\}">F</figure-callout>}}_{510},(P_s,S)-F_{511i}(P_s,S)|/F_{510i}(P_s,S))---\left(8\right)$

[formula 9]

$\mathrm{\&Delta;C}=|C_{510}-{\mathrm{<figure-callout\; id="511"\; label="C"\; filenames="A20061010009000541.png"\; state="\{\{state\}\}">C</figure-callout>}}_{511}|/C_{510}---\left(9\right)$

In formula (8) and formula (9), the spatial frequency characteristic quantity R and the time orientation frequecy characteristic amount C of the numeral of interpolation difference representation feature amount 510 and characteristic quantity 511.In addition, in formula (8), frame number i is 1 to 2, calculates the variable quantity of two field picture P1 and P2, and frame number i is changed by the specific mode of any two two field pictures.

In addition, in the present embodiment, time space characteristic variations amount is calculated portion 505, use is calculated variable quantity 512 by the function that the difference of the difference of the spatial frequency characteristic quantity R in characteristic quantity 510 and the characteristic quantity 511 and time direction frequecy characteristic amount C constitutes, but also can be according to (the Δ R of any variable quantity at least among spatial frequency characteristic quantity R and the time orientation frequecy characteristic amount C _mOr Δ C) calculates variable quantity 512.In addition, based on the function of above-mentioned formula (8) and formula (9), can not also from difference, but the function of calculating from the ratio of characteristic quantity 510 and characteristic quantity 511.

Like this, time space characteristic variations amount is calculated portion 505, and the variable quantity 512 of calculating with motion feature amount 510, is outputed to image evaluation determination portion 506.

Image evaluation determination portion 506 is calculated based on the motion feature amount 510 of calculating portion 505 output from time space characteristic variations amount and variable quantity 512 and is deposited in time space frequency characteristic that the image about the characteristic quantity 513 of the time space frequency characteristic of the human vision property of the benchmark that becomes relative evaluation the datum characteristic storage part 507 has and the evaluation of estimate 514 of the relativeness of the time space frequency characteristic of human eye.In addition, evaluation of estimate 514 both can be the evaluation to whole image, also can be the evaluation that the N that is clipped between any two two field pictures is opened two field picture and these any two two field pictures.

In specific words, image evaluation determination portion 506, according to as the characteristic quantity 513 of datum characteristic, variable quantity 512 (based on the Δ R of the difference of spatial frequency characteristic quantity _m, based on the difference delta C of time orientation frequecy characteristic amount) and motion feature amount 510, utilize following formula (10), calculate evaluation of estimate 514 based on the relativeness of the time space frequency characteristic of time space frequency differential and human eye.This evaluation of estimate 514 is amounts of relativeness of the characteristic of the variable quantity of characteristic of the frequency component that has of expression image and the appreciable frequency component of human eye.

[formula 10]

In following formula, ρ, ω represent weight coefficient, and τ represents constant.In addition, (Ps, S) characteristic quantity 113 of the appreciable time space frequency characteristic of the relevant human eye of expression is the amount that leaves in the datum characteristic storage part 507 to H.In addition, V ₅₁₀Expression is to the motion feature amount of the characteristic quantity 510 of input signal of video signal.

In addition, (Ps S), needs not be some certain characteristic quantities to the H of the characteristic of the frequency component that can feel as human eye, also can be the characteristic quantity of being determined by characteristic quantity 510 513 about the time space frequency characteristic.

In addition, the characteristic of the frequency component that human eye can be felt is used the characteristic of expression to the sensitivity of the frequency of time space, but is not limited thereto.Such as, also can use the characteristic of a certain specific quality of images such as characteristic of the time space frequency of spot to image, torsion resistance, flicker.In addition, under situation, evaluation of estimate E can calculate the value of expression to a certain specific quality evaluation of image.

In addition, in the present embodiment, illustrate by as the characteristic quantity 513 of datum characteristic, based on the Δ R of the difference of spatial frequency characteristic quantity _m, based on the Δ C of the difference of time orientation frequecy characteristic amount and the evaluation of estimate E that motion feature amount 510 is calculated, but so long as calculate by variable quantity 512 and characteristic quantity 513 are compared, according to the function that the relativeness between variable quantity 512 and the characteristic quantity 513 is calculated, be which kind of function can.Utilize this function, can represent that the frequency component that restriction is applied to a part of signal frequency component on the input image limits the relative evaluation value for visual characteristic that human eye is given great influence.

In addition, in the present embodiment, the motion feature amount of using in calculating evaluation of estimate E 510 is used motion feature amount V, based on the difference delta R of spatial frequency characteristic quantity _mWith Δ C based on the difference of time orientation frequecy characteristic amount, but wherein some also passable.

Then, processing describes to the image evaluation in the image evaluation device of execution mode 2.Fig. 9 is the flow chart that the image evaluation processing of image evaluation device 50 is shown.

Image characteristics analysis portion 501 is calculated the characteristic quantity 510 about the time space frequency characteristic that the image of importing signal of video signal 508 is had.

On the other hand, the frequency component restriction adds the Ministry of worker 503, determines two two field pictures (step S701) among a plurality of two field pictures.So, frequency component restriction adds the Ministry of worker 503, is identified for opening two field picture and carrying out the frequency component limit amount (step S702) that the frequency component restriction is handled being present in N between two frames in the two field picture according to the frequency component through the spatial frequency of the two field picture of conversion.

Then, the frequency component restriction adds the Ministry of worker 503, removes the frequency of being determined by frequency component limit amount B in order to open two field picture from N, and this two field picture is carried out frequency component restriction processed, generated frequency component restriction processing image 509 (step S703).So image characteristics analysis portion 504 is calculated the characteristic quantity 511 (step S704) about the time space frequency characteristic that the image to the frequency component restriction processing image 509 that generates has.In other words, image characteristics analysis portion 504, to each two field picture,, calculate characteristic quantity 511 about the time space frequency characteristic of frequency component restriction processing image 509 by the image evaluation till the step S401～S405 of explanation in above-mentioned execution mode 1 is handled (with reference to Fig. 6).On the other hand, image characteristics analysis portion 501 is equally also calculated the motion feature amount 510 about the time space frequency characteristic of input signal of video signal 508.

Then, time space characteristic variations amount is calculated portion 505, calculates from two the motion feature amounts 510 of image characteristics analysis portion 501 and 504 outputs of image characteristics analysis portion and the difference of characteristic quantity 511, calculates variable quantity 512 (step S705).So, image evaluation determination portion 506, read in be stored in the datum characteristic storage part 507 about characteristic quantity 513 (step S706) as the time space frequency characteristic of human vision property.Image evaluation determination portion 506, the characteristic quantity 513 and the variable quantity 512 that read in are compared, calculate based on the relativeness between the time space frequency characteristic of the input time space frequency characteristic that has of image and human eye and the evaluation of estimate 514 (step S707) of motion feature amount 510.

As mentioned above, utilize the image evaluation device 50 in the execution mode 2, can generate by the two field picture in the input image being limited the frequency component that processing the generated restriction processing image 509 of a part of signal frequency component.So, by calculating about image that limits a part of frequency component and the characteristic quantity of importing the time space frequency characteristic of image, the characteristic quantity of the time space frequency characteristic that has with the variable quantity of the characteristic quantity of calculating with about human vision property compares, and can calculate the evaluation of estimate of the relativeness that the time space frequency characteristic of the variable quantity of the time space frequency characteristic that causes based on the frequency component restriction that is applied to the input image and human eye causes.So,, the relative value of considering the influence of human eye can be shown according to giving the variable quantity that the frequency component of importing image limits the time space frequency characteristic that causes.

[execution mode 3]

Then, embodiments of the present invention 3 are described.The formation block diagram that Figure 10 constitutes for the function of image evaluation device 80 that execution mode 3 is shown for example.This image evaluation device 80 is to have storage devices such as CPU (central processing unit), memory physically and be used for the computer of input/output unit (communicator etc.) of input and output image.Such as, image evaluation device 80 are devices of the information processing carried out of mobile communication terminals such as the fixed communication terminal of PC terminal etc. or portable telephone etc.

With reference to Figure 10 the function formation of image evaluation device 80 is described.As shown in figure 10, the formation of image evaluation device 80 comprises buffer part 801, frequency component limits job evaluation portion 802, evaluation of estimate detection unit 803 and frequency component limit amount adjustment part 804 (frequency component adjustment unit, frequency component correction adjustment unit).

Buffer part 801, be to have the memory that temporarily to preserve the input signal of video signal of importing as motion image signal from the outside 805, can will import signal of video signal 805 and output to frequency component restriction job evaluation portion 802 part of the read-write of this memory control data.In addition, after, the input signal of video signal 806 that temporarily is kept in the buffer part 801 outputs to frequency component restriction job evaluation portion 802 by revaluing indication 809.

Frequency component restriction job evaluation portion 802 has the same function of function that has with the image evaluation device 50 described in the above-mentioned execution mode 2.In other words, frequency component restriction job evaluation portion 802 has and adds with above-mentioned image characteristics analysis portion 501, by the frequency component restriction that the frequency component restriction image that the Ministry of worker 503 and image characteristics analysis portion 504 constitute adds the Ministry of worker 502, time space characteristic variations amount is calculated the same function of function that portion 505, image evaluation determination portion 506 and datum characteristic storage part 507 have.Frequency component restriction job evaluation portion 802, according to the input signal of video signal 806 from buffer part 801 outputs, the frequency component of calculating above-mentioned time space frequency characteristic limits the evaluation of estimate 807 that causes.In addition, having the unit that adds the Ministry of worker's 503 said functions with frequency component restriction, is can be set by the user and have can be to the part of the function of outside input and output for the frequency component limit amount B that makes regulation and about the information (determining the information of any two frame) of specific frame.

Frequency component restriction job evaluation portion 802 outputs to evaluation of estimate detection unit 803 with the evaluation of estimate of calculating 807.In addition, have with frequency component restriction job evaluation portion 802 in the frequency component restriction unit that adds the Ministry of worker's 503 said functions fixed frequency component limit amount B and fixed two field picture are outputed to frequency component limit amount adjustment part 804.

Evaluation of estimate detection unit 803 is judged from frequency component and whether is limited the evaluation of estimate 807 of job evaluation portion 802 outputs more than or equal to setting.In addition, the afore mentioned rules value both can be predefined, also can give from the outside.Evaluation of estimate detection unit 803 outputs to frequency component limit amount adjustment part 804 with result of determination 808.

Frequency component limit amount adjustment part 804 is the result of determination 808 that receive from 803 outputs of evaluation of estimate detection unit, carries out the part from the adjustment of the frequency component limit amount of the input signal of video signal 806 of buffer part 801 outputs.In addition, the size of the frequency component limit amount B that receives by frequency component restriction job evaluation portion 802 of storage and about the information (determining the information of any two two field pictures) of the two field picture determined.Frequency component limit amount adjustment part 804, under the situation of result of determination 808 more than or equal to setting, according to the frequency component limit amount B of storage and about the information of two field picture, adjust so that frequency component limit amount B and become greatly as the N value that is clipped in two two field picture numbers between the two field picture.Therefore, frequency component limit amount adjustment part 804 will output to frequency component restriction job evaluation portion 802 through the frequency component limit amount Bn that adjusts with about the information (signal 810) of the frame determined.Frequency component limit amount adjustment part 804,801 outputs revalue indication 809 to buffer part simultaneously.

In addition, in the present embodiment, carry out both adjustment of the size of frequency component limit amount B and N value, but also can be according to the information of the size of the frequency component limit amount B that receives by frequency component restriction job evaluation portion 802 and the two field picture of determining, at least one of the information of the size of frequency component limit amount B and the two field picture determined modulated.

In addition, in the present embodiment, be to adjust so that frequency component limit amount B becomes big, but also can adjust frequency component limit amount Bn is become less than frequency component limit amount B.Equally, also can adjust the N value is diminished.

Frequency component limit amount adjustment part 804, receive result of determination 808 from evaluation of estimate detection unit 803, and before result of determination 808 becomes smaller or equal to setting, adjust to the size of the frequency component limit amount B of frequency component limit amount and as the two field picture number N of processing object.Frequency component limit amount adjustment part 804 becomes occasion less than setting in result of determination 808, finishes the size of frequency component limit amount B and as the adjustment of the two field picture number N of processing object.

In addition, in the present embodiment, what illustrate is to finish the example of adjustment when result of determination 808 becomes smaller or equal to setting, but also can be the benchmark that the adjustment beyond it finishes.Such as, also can be until the size of frequency component limit amount B becomes more than or equal to the method for setting and the method for the end setup beyond adjusting this evaluation of estimate that finishes to adjust when number of times surpasses certain number of times.

Frequency component limit amount adjustment part 804, during in the size that finishes frequency component limit amount B with as the adjustment of the two field picture number N of processing object, with the information of the size of the frequency component limit amount Bn that finishes to adjust and the two field picture determined as adjusting result's 811 outputs.

In addition, in the present embodiment, be will result 811 finishes the size of the frequency component limit amount Bn that adjusts and the information of the two field picture determined be exported as adjusting, but also can will export through the information of the size of the frequency component limit amount Bn of any amount adjusted and the two field picture determined.

In addition, frequency component limits job evaluation portion 802, also can be used as the image evaluation device 10 of execution mode 1.In this case, frequency component restriction job evaluation portion 802, can be according to from the characteristic quantity 105 of image characteristics analysis portion 101 (with reference to Fig. 1) output be stored in the characteristic quantity 106 of becoming of datum characteristic storage part 103 (with reference to Fig. 1) about the time space frequency characteristic of the human vision property of the benchmark of relative evaluation, calculate the evaluation of estimate 807 of relativeness of the time space frequency characteristic of the time space frequency characteristic that has based on image and human eye, output to evaluation of estimate detection unit 803.

So whether evaluation of estimate detection unit 803 judges the evaluation of estimate 807 that receives more than or equal to setting, in the occasion more than or equal to setting, evaluation of estimate detection unit 803 outputs to frequency component limit amount adjustment part 804 as result of determination 808 with judged result.Frequency component limit amount adjustment part 804, reception is from the result of determination 808 of evaluation of estimate detection unit 803 outputs, carry out adjustment, will adjust indication and output to frequency component restriction job evaluation portion 802 (image evaluation device 10) and buffer part 801 the frequency component limit amount B of input signal of video signal 806.

Herein, frequency component restriction job evaluation portion 802 (image evaluation devices 10) when receiving the adjustment indication, are being selected any two two field pictures from the input signal of video signal of buffer part 801 outputs.So, frequency component restriction job evaluation portion 802, control image characteristics analysis portion 101 generates the frequency component limited images of the predetermined frequency of restriction so that the N that clips between selected two two field pictures is opened (N is the integer more than or equal to 1) image.The restriction of this frequency is handled, and handles identical with the restriction of frequency component limit amount B in using execution mode 2.In addition, the image evaluation device 10 of execution mode 1 must have the function of carrying out this restriction processing.In addition, herein based on the evaluation of estimate of the frequency component limited images of estimating once again, adjusting, both can be evaluation to whole image, also can be the evaluation that the N that clips between any two two field pictures is opened two field picture and these any two two field pictures.

Flow process to the image processing in the image processor 80 of execution mode 3 is illustrated below.Figure 11 is the flow chart that the image processing of image processor 80 is shown.

At first, buffer part 801 when will outputing to frequency component restriction job evaluation portion 802 from the input signal of video signal 805 that the outside is imported as motion image signal, is preserved this input signal of video signal 805 (step S901) temporarily.

Afterwards, frequency component restriction job evaluation portion 802 uses from the input signal of video signal 805 of buffer part 801 outputs, calculates the evaluation of estimate 807 (step S902) of the variable quantity of the time space frequency characteristic that restriction causes based on frequency component.In other words, frequency component restriction job evaluation portion 802, by using input signal of video signal 805, the time space resolution changable amount of carrying out in above-mentioned execution mode 2 till the step S701～S707 of explanation is determined to handle (with reference to Fig. 9), calculates the evaluation of estimate 807 of the relativeness of the variable quantity of the time space frequency characteristic that restriction causes based on frequency component and human vision property.

Afterwards, evaluation of estimate detection unit 803, carry out evaluation of estimate 807 and whether be regulation evaluation of estimate judgement and result of determination outputed to frequency component limit amount adjustment part 804 (step S903).Frequency component limit amount adjustment part 804, when judging result of determination 808 (S904) greater than setting, the component limit amount of adjusting frequency will output to frequency component restriction job evaluation 802 (the step S905) of portion through the frequency component limit amount adjusted and the information (signal 810) of two field picture.

Frequency component limit amount adjustment part 804, before receiving the result of determination 808 of result of determination 808 less than setting, component limit amount and output to step (the step S902～S905) of frequency component restriction job evaluation portion 802 repeats to adjust frequency.

Frequency component limit amount adjustment part 804, the size of the frequency component limit amount Bn that finishes to adjust when receiving result of determination 808 less than the judgement of setting and will finish to adjust and specific frame image information are exported (step S906) as adjusting result 811.

In addition, when frequency component being limited job evaluation portion 802 as the image evaluation device 10 of execution mode 1, as S902～S905, frequency component restriction job evaluation portion 802, before evaluation of estimate D becomes smaller or equal to setting, repeat from the input signal of video signal of input, to select any two two field pictures, the N that clips in selected two field picture is opened the processing that two field picture is calculated evaluation of estimate D.In this case, control N is changed towards the direction that reduces.

As mentioned above, utilize the image processor 80 in the execution mode 3, can generate the frequency component restriction processing image that generates by the processing that limits a part of signal frequency component a certain two field picture in the input image.So image processor 80 is calculated image that is restricted about a part of frequency component and the characteristic quantity of importing the time space frequency characteristic of image.Image processor 80, characteristic quantity by the time space frequency characteristic that has to the variable quantity of the characteristic quantity of calculating with about human vision property compares, and calculates the evaluation of estimate of the relativeness that the time space frequency characteristic (with the corresponding fiducial value of variable quantity) of the variable quantity of the time space frequency characteristic that causes based on the frequency component restriction that is applied to the input image and human eye causes.Image processor 80, the judgement of the evaluation of estimate of calculating, utilizing result of determination to adjust frequency the component limit amount also will be through the frequency component limit amount output of adjusting.

So, image processor 80, can limit the variable quantity of the time space frequency characteristic that causes according to the frequency component of importing image, judgement considers that the relative value to the influence of human eye is a setting, definite frequency component limit amount that can keep more than or equal to the value of setting.

In addition, above-mentioned image processor 80, such as, can be used for image acquisition device (such as, camera), image transfer apparatus, image coding device or image decoding device.

[execution mode 4]

Then, embodiments of the present invention 4 are described.Time space frequency-response analysis device 100 in this execution mode 4, be with execution mode 1 in the image characteristics analysis portion 101, the image characteristics analysis portion 501 of image evaluation device 50 in the execution mode 2 of image evaluation device 10 or the suitable part of image characteristics analysis portion 504.The formation block diagram that Figure 12 constitutes for the function of time space frequency-response analysis device 100 that execution mode 4 is shown for example.This moment, spatial frequency characteristic analytical equipment 100 was to have storage device such as CPU (central processing unit), memory physically and be used for the computer of input/output unit (communicator etc.) of input and output image.Such as, time space frequency-response analysis device 100 are devices of the information processing carried out of mobile communication terminals such as the fixed communication terminal of PC terminal etc. or portable telephone etc.

With reference to Figure 12 the function formation of time space frequency-response analysis device 100 is described.As shown in figure 12, the formation of time space frequency-response analysis device 100 comprises spatial frequency and calculates portion 1001 (spatial frequency is calculated the unit), spatial frequency component arrangement portion 1002 (spatial frequency component arrangement unit), datum characteristic storage part 1003, time space frequency and calculate portion 1004 (the time space frequency is calculated the unit) and time space frequency characteristic performance portion 1005 (characteristic arithmetic element).Below each formation is described.

Time space frequency-response analysis device 100, order read in the frame that constitutes image (such as, two field picture shown in Figure 2), are stored in storage part (not shown).

Spatial frequency is calculated portion 1001, is the two field picture that reads in storage, calculates the part of size of spatial frequency amplitude of the spatial frequency of two field picture.In specific words, at first, spatial frequency is calculated portion 1001, carries out discrete Fourier transform by the entire image to two field picture, carries out the conversion from the pixel space to the frequency space.In the present embodiment, spatial frequency is calculated portion 1001, and two field picture is carried out two dimensional discrete Fourier transform, two field picture is transformed to the data mode with spatial frequency zone.Spatial frequency is calculated portion 1001, according to the Fourier coefficient through conversion, also calculates the amplitude of the spatial frequency component that has carried out Fourier transform.In specific words, object images is applied the two-dimensional Fourier transform generated frequency determinant of coefficient, calculate the energy level of the square root of power spectrum as the coefficient in this coefficient of frequency determinant.In other words, spatial frequency is calculated portion 1001, calculates the size of each square root sum square of the real number item of each Fourier coefficient and imaginary terms as the spatial frequency amplitude of each spatial frequency.Spatial frequency is calculated portion 1001, and the amplitude of the spatial frequency component of calculating is sent to spatial frequency component arrangement portion 1002.

In addition, when constituting time space frequency-response analysis device 100 as image characteristics analysis portion 504, spatial frequency is calculated portion 1001, and input can limit the frequency component restriction image of a part of frequency component to object images, and calculates the amplitude of spatial frequency component in view of the above.

In addition, the manifestation of spatial frequency amplitude is not limited thereto, and also can be other manifestations.In addition, in the present embodiment, be to carry out the amplitude that Fourier transform is calculated spatial frequency component, the method for utilizing of spatial frequency component be not limited thereto.Such as, also can calculate phase place and use amplitude and phase place.

In addition, in the present embodiment, what illustrate is the example of carrying out discrete Fourier transform in the calculating of spatial frequency in the entire frame image, but also can use the one-dimensional discrete Fourier transform on all directions of image, is divided into repeatedly being transformed to frequency field.In addition, the method for frequency translation is not limited to discrete Fourier transform, also can use discrete cosine transform and wavelet transform or the like.

The amplitude 1007 of being calculated the spatial frequency component that portion 1001 calculates by spatial frequency is read in spatial frequency component arrangement portion 1002 from the low frequency band component of spatial frequency to the high band component order, by being transformed to the one-dimensional data row, carry out the arrangement of spatial frequency component.In specific words, spatial frequency component arrangement portion 1002, the amplitude of the spatial frequency component that will obtain through two-dimensional Fourier transform according to zigzag scanning and predefined order from the DC component until till frequency component last (i.e. a row size) read in proper order and be transformed to one-dimensional data.

In addition, in the present embodiment, be the data that spatial frequency component are transformed to one dimension, but also can just put in order according to two-dimentional former state.

In addition, in the present embodiment, be the spatial frequency component of a part of the amplitude of finish space frequency component, but so long as at least a portion frequency component of spatial frequency component, can to any spatial frequency component arrangement.Such as, both can be whole amplitudes of finish space frequency component, also can be only arrangement by datum characteristic storage part 1003 that deposit with the corresponding frequency component that obtains of datum characteristic.

In addition, when reading in, be and corresponding the obtaining of depositing in the datum characteristic storage part 1003 of datum characteristic in proper order for preferred with predefined order.In addition, the method for the amplitude of finish space frequency is not limited thereto.

In addition, spatial frequency component arrangement portion 1002, storage is the number of the frame rate of image through the quantity of the amplitude of the spatial frequency component of arrangement.At this, for convenience of description, the frame rate of setting of image is 30.Spatial frequency component arrangement portion 1002 is transformed to 2-D data 1009 with the one-dimensional data of the amplitude of the spatial frequency component of 30 number of frames.In other words, for the spatial frequency component that makes one-dimensional data becomes same row, become 2-D data 1009 by will on column direction, arranging to put in order by each one-dimensional data that delegation constitutes.Spatial frequency component arrangement portion 1002 will send to the time space frequency through the 2-D data of arrangement and calculate portion 1004.

In addition, in the present embodiment, the number quantity of the frame rate by stored images can make the unit of frame rate and temporal frequency consistent, but the number quantity of the frame rate of image is not limited thereto.

The time space frequency is calculated portion 1004, will carry out one-dimensional Fourier transform to expression towards the row of the variation of the time orientation of each spatial frequency component through the 2-D data of arrangement, calculates the amplitude 1010 of time space frequency.In specific words, the time space frequency is calculated portion 1004, and 30 amplitude datas of each spatial frequency component are carried out Fourier transform in proper order, uses the amplitude of calculating the time space frequency component through the Fourier coefficient of conversion.In other words, for the square root of calculating power spectrum energy level, calculate the size of each square root sum square of the real number item of each Fourier coefficient and imaginary terms as the spatial frequency amplitude of each spatial frequency as the one dimension Fourier coefficient of each spatial frequency component.In addition, the manifestation of spatial frequency amplitude is not limited thereto, and also can be other manifestations.

So the time space frequency is calculated portion 1004, the amplitude 1010 of the time space frequency calculated is sent to time space frequency characteristic performance portion 1005.

In addition, in the present embodiment, what illustrate for example is to carry out Fourier transform with regard to the data rows of total space frequency component, but also can be at least carries out Fourier transform with regard to the amplitude data of spatial frequency component.In addition, in the present embodiment, be to use one-dimensional Fourier transform to calculate temporal frequency, but also can use two-dimensional Fourier transform to calculate the time space frequency.

In addition, in the present embodiment, be to carry out the amplitude that Fourier transform is calculated spatial frequency component, but the method for utilizing of spatial frequency component is not limited thereto.Such as, also can calculate phase place and use amplitude and phase place.

In addition, the method for frequency translation is not limited to discrete Fourier transform, also can use discrete cosine transform and wavelet transform or the like.

Time space frequency characteristic performance portion 1005, extraction and application time space frequency is calculated the amplitude of the specific frequency component of the temporal frequency that portion 1004 calculates.In specific words, time space frequency characteristic performance portion 1005 extracts spatial high-frequency band x[Hz] specific temporal frequency y[Hz] big or small P1 and the low high frequency band z[Hz in space] specific temporal frequency v[Hz] the big or small P2 of component.Use the P1 and the P2 that extract, utilize the time space frequency characteristic frame rate f of following numerical expression (11) performance image.

[formula 11]

Time and space frequency characteristic F J (P1, P2)=k1*P1+k2*P2

…(11)

In addition, k1 and k2 represent the weight coefficient to P1 and P2 respectively.

The amount that specific frequency x, y, z, the v of the time space frequency that illustrates in the present embodiment in addition, is based on the datum characteristic of datum characteristic storage part 1003 is preferred.In addition, also can use predetermined coefficient x, y, z, v.

In addition, in the present embodiment, be to set specific spatial frequency and time frequency, but also can set bandwidth each frequency.In addition, also can extract the size of time space frequency component of maximum, minimum value or mean value of the size of the time space frequency that use comprises in the bandwidth of each frequency that this situation is set.

In addition, in the present embodiment, what illustrate is the example of extracting the amplitude of two frequency components, but so long as uses the size performance time space frequency characteristic F of the amplitude of high fdrequency component and low frequency component to get final product at least.In addition, in the present embodiment, use be high fdrequency component and low frequency component, but also can use intermediate frequency component expressive space frequency characteristic F.In addition, time space frequency characteristic F calculates, and is to utilize the ranking operation of P1 and P2 to obtain, but so long as the computing of use P1 and P2 also can use any computing to obtain time space frequency characteristic F.

Afterwards, processing describes to the image evaluation in the time space frequency-response analysis device 100 of execution mode 4.Figure 13 is the flow chart that the image evaluation processing of time space frequency-response analysis device 100 is shown.

Spatial frequency is calculated portion 1001, reads in the input signal of video signal 1006 (step S1101) of consecutive image data.Afterwards, spatial frequency is calculated portion 1001, obtains frame number (S1102), will be transformed to the amplitude 1007 (step S1103) of the spatial frequency that image has as the input signal of video signal 1006 of two field picture.

Afterwards, spatial frequency component arrangement portion 1002 is transformed to one-dimensional data by the amplitude 1007 with spatial frequency and puts (step S1104) in order, such as, the spatial frequency data of storing 30 frame signs generates 2-D data 1009 (step S1105).

Afterwards, the time space frequency is calculated portion 1004, the row of variation to the time orientation of each spatial frequency of expression 2-D data 1009 carry out one-dimensional Fourier transform in proper order, by obtaining the amplitude 1010 (step S1106) of calculating the time space frequency through the Fourier transform coefficient of conversion.

So, time space frequency characteristic performance portion 1005, extract the time space frequency amplitude 1010 at least more than or equal to one specific frequency components, the time space frequency characteristic 1011 (step S1107) of performance image.

As mentioned above, utilize the time space frequency-response analysis device 100 in the execution mode 4, to a certain two field picture of input in the image, the performance by the arrangement carrying out being consistent with the datum characteristic of time space frequency component can show the time space frequency characteristic that exists in the datum characteristic.In addition, by datum characteristic is become visual characteristic, can consider performance to the time space frequency of the influence of human eye.

[execution mode 5]

For the image evaluation device in above execution mode 1～execution mode 4 10,50 and image processor 80 and time space frequency-response analysis device 100, can utilize program to realize identical functions.In other words, can by program is recorded recording medium (such as, floppy disk, CD-ROM, DVD etc.), the computer that is made of CPU, RAM, ROM, auxilary unit etc. is read in, by computer realization and image evaluation device 10,50 and image processor 80 and time space frequency-response analysis device 100 identical functions.Below these programs are described.

At first, describe with the program of image evaluation device 10 identical functions realizing.Figure 14 is for realizing the module structure drafting with the program P10 of image evaluation device 10 identical functions.The formation of program P10 comprises image characteristics analysis module P101, image evaluation determination module P102 and datum characteristic storage module P103.

By carrying out the function that these each modules realize, the same with the function of each component parts of image evaluation device 10.In other words, above-mentioned each module can realize and a component parts identical functions respectively accordingly with image characteristics analysis portion 101, image evaluation determination portion 102 and datum characteristic storage part 103.

Afterwards, describe with the program of image evaluation device 50 identical functions realizing.Figure 15 is for realizing the module structure drafting with the program P50 of image evaluation device 50 identical functions.The formation of program P50 comprises image characteristics analysis module P501, frequency component restriction processing module P503, image characteristics analysis module P504, time space characteristic variations amount is calculated module P505, image evaluation determination module P506 and datum characteristic storage module frequency component P507.

The same by carrying out the function that these each modules realize with the function of each component parts of image evaluation device 50.In other words, above-mentioned each module, add the Ministry of worker 502 (the frequency component restriction adds the Ministry of worker 503 and image characteristics analysis portion 504) with image characteristics analysis portion 501, frequency component restriction image respectively, that time space characteristic variations amount is calculated portion 505, image evaluation determination portion 506 and datum characteristic storage part 507 is corresponding, can realize and each component parts identical functions.

Afterwards, describe with the program of image processor 80 identical functions realizing.Figure 16 is for realizing the module structure drafting with the program P80 of image processor 80 identical functions.The formation of program P80 comprises buffer module P801, frequency component restriction job evaluation module P802, evaluation of estimate determination module P803 and frequency component limit amount adjusting module P804.In addition, frequency component restriction job evaluation module P802 is and image evaluation program P10 or the identical module of P50.

The same by carrying out the function that these each modules realize with the function of each component parts of image processor 80.In other words, above-mentioned each module, corresponding with buffer part 801, frequency component restriction job evaluation portion 802, evaluation of estimate detection unit 803 and frequency component limit amount adjustment part 804, frequency component limit amount adjustment part respectively, can realize and each component parts identical functions.

Afterwards, describe with the program of time space frequency-response analysis device 100 identical functions realizing.Figure 17 is for realizing the module structure drafting with the program P1000 of time space frequency-response analysis device 100 identical functions.The formation of the module of program P1000 comprises spatial frequency module P1001, spatial frequency component sorting module P1002, datum characteristic storage module P1003, the time space frequency is calculated module P1004 and time space frequency characteristic functional modules P1005.

The same by carrying out the function that these each modules realize with the function of each component parts of time space frequency-response analysis device 100.In other words, above-mentioned each module, calculating portion 1001, spatial frequency component arrangement portion 1002, datum characteristic storage part 1003, time space frequency with spatial frequency respectively, to calculate timely spatial frequency characteristic performance portion 1005 of portion 1004 corresponding, can realize and each component parts identical functions.

Action effect to image evaluation device 10, image evaluation device 50 and the image processor 80 of present embodiment describes below.Image evaluation device 10 will be stored in the datum characteristic storage part 103 as the datum characteristic about the predetermined time space frequency characteristic of image.So in image evaluation device 10,101 inputs of image characteristics analysis portion are calculated the time space frequency characteristic of object image by the object image of a plurality of image constructions.Image evaluation determination portion 102 according to the time space frequency characteristic of calculating and the relativeness of datum characteristic, is determined the evaluation of estimate of the object image of input.Thus, can carry out, and the image evaluation in the image be compared, can carry out appropriate evaluation image based on evaluation as the characteristic of the image of consecutive image.Therefore,, obtain high praise sometimes, can estimate rightly as image as image even exist more than of images in the image owing to reasons such as resolution is low are estimated the occasions of step-down.

In addition, the image characteristics analysis portion 101 of this image evaluation device 10 has spatial frequency to be calculated portion 1001, spatial frequency component arrangement portion 1002, datum characteristic storage part 1003, time space frequency and calculates timely spatial frequency characteristic performance portion of portion 1,004 1005, can carry out following action.In other words, spatial frequency is calculated the spatial frequency that a plurality of images are calculated by portion 1001, spatial frequency component arrangement portion 1002 with the component transformation of the spatial frequency calculated for and the corresponding data of datum characteristic that are stored in the datum characteristic storage part 1003.So, the time space frequency is calculated portion 1004, to carrying out frequency translation through changing more than or equal to time of one spatial frequency component at least of the spatial frequency component of conversion, calculate the time space frequency, time space frequency characteristic performance portion 1005 can carry out computing to the data of using high fdrequency component and low frequency component to calculate the time space frequency characteristic that can show image among the time space frequency of calculating.Therefore, can show the time space frequency characteristic of image rightly, can carry out, and the image evaluation in the image be compared, can carry out appropriate evaluation image based on evaluation as the characteristic of the image of consecutive image.

In addition, in image processor 80, frequency component limit amount adjustment part 804, according to the evaluation of estimate of determining by frequency component restriction job evaluation portion 802 (suitable) with image evaluation device 10, N between any two images in the object image is opened (N is the integer more than or equal to 1) image, adjust the frequency component of regulation.So the evaluation of estimate that above-mentioned N that signal frequency component is adjusted opens image can be calculated by frequency component restriction job evaluation portion 802.Therefore, can the image of a part of having adjusted frequency component be revalued, the image all frequency components is estimated from various angles.

Image evaluation device 50 will be stored in the datum characteristic storage part 507 as the datum characteristic about the predetermined time space frequency characteristic of image.So, the image characteristics analysis portion 501 of image evaluation device 50, input is calculated the 1st image characteristics based on the time space frequency characteristic of object image by the object image of a plurality of image constructions.In addition, frequency component restriction image adds the Ministry of worker 502, calculates the 2nd image characteristics of the time space frequency characteristic in the frequency component restriction image that obtains based on object images is limited a part of frequency component.Time space characteristic variations amount is calculated portion 505, calculates the variable quantity for the 2nd image characteristics of the 1st image characteristics, and image evaluation determination portion 506 according to the relativeness of this variable quantity and datum characteristic, is calculated the evaluation of estimate of the object image of input.

Therefore, a part of frequency component of restriction can be used and the frequency component that obtains restriction image carries out the evaluation to the object image, the difference on the frequency component in the image that the frequency component restriction of time space frequency characteristic causes can be distinguished.Therefore, can limit image according to frequency component estimates rightly to the characteristic of the frequency component of the image that will estimate.Such as, make frequency component restriction image become the occasion of the image of the appreciable frequency component of vision that limits the people, by estimating, can carry out the evaluation of image of characteristic of relevant people's the imperceptible frequency component of vision according to the variable quantity between this frequency component restriction image and the object image and the relativeness of datum characteristic.

In addition, image evaluation device 50, add in the Ministry of worker 502 at frequency component restriction image, a part of frequency component that can open image by the image restriction N that the N between any two images in the object image is opened (N is the integer more than or equal to 1) obtains frequency component restriction image.Therefore, can carry out the evaluation of object image, can consider that a part of N in the object image opens the evaluation of characteristic of the frequency component of image according to this frequency component that is restricted restriction image.

In addition, the image characteristics analysis portion 501 of image evaluation device 50, have spatial frequency and calculate portion 1001, spatial frequency component arrangement portion 1002, datum characteristic storage part 1003, time space frequency and calculate portion 1004 and time space frequency characteristic performance portion 1005, can carry out following action.In other words, spatial frequency is calculated portion 1001, calculates the spatial frequency of a plurality of images, spatial frequency component arrangement portion 1002 with the component transformation of the spatial frequency calculated for and the corresponding data of datum characteristic that are stored in the datum characteristic storage part 1003.So, the time space frequency is calculated portion 1004, time through at least one or more space frequency component in the spatial frequency component of conversion changed carry out frequency translation, calculate the time space frequency, time space frequency characteristic performance portion 1005 uses high fdrequency component and low frequency component to calculate the data of the time space frequency characteristic that can show image among the time space frequency of calculating.Therefore, can show the time space frequency characteristic of image rightly, can carry out, and the image evaluation in the image be compared, can carry out appropriate evaluation image based on evaluation as the characteristic of the image of consecutive image.

In addition, the image characteristics analysis portion 504 of image evaluation device 50, have spatial frequency and calculate portion 1001, spatial frequency component arrangement portion 1002, datum characteristic storage part 1003, time space frequency and calculate portion 1004 and time space frequency characteristic performance portion 1005, can carry out following action.In other words, spatial frequency is calculated portion 1001, calculate spatial frequency from the frequency component restriction image that a part of frequency component of object images restriction is obtained, spatial frequency component arrangement portion 1002, with the component transformation of the spatial frequency calculated for and the corresponding data of datum characteristic that are stored in the datum characteristic storage part 1003.So, the time space frequency is calculated portion 1004, time through at least one or more space frequency component in the spatial frequency component of conversion changed carry out frequency translation, calculate the time space frequency, time space frequency characteristic performance portion 1005 uses high fdrequency component and low frequency component can carry out computing to the data of the time space frequency characteristic that can show image among the time space frequency of calculating.Therefore, can show the time space frequency characteristic of image rightly, can carry out, and the image evaluation in the image be compared, can carry out appropriate evaluation image based on evaluation as the characteristic of the image of consecutive image.

In addition, the frequency component limit amount adjustment part 804 of image processor 80 can be adjusted the characteristic of the frequency component that limited according to evaluation of estimate, can utilize the characteristic through the frequency component adjusted to revalue.

In addition, the frequency component limit amount adjustment part 804 of image processor 80 can change any two images as the restriction object of frequency component according to evaluation of estimate, can revalue according to the image that the restriction object is changed.

In addition, in image evaluation device 10, image evaluation device 50 and image processor 80, by the frequency characteristic that the datum characteristic conduct that will be stored in the datum characteristic storage part 103,507 is determined according to human vision property, can carry out evaluation based on human vision property.

In addition, above-mentioned image characteristics analysis portion 101, image characteristics analysis portion 501 and image characteristics analysis portion 504 can constitute time space frequency-response analysis device 100, reach following action effect.In other words, spatial frequency is calculated portion 1001, calculates the spatial frequency of a plurality of images, and spatial frequency component arrangement portion 1002 puts the component of the above-mentioned spatial frequency calculated in order according to being transformed to the corresponding one-dimensional data of predetermined datum characteristic.So, the time space frequency is calculated portion 1004, in the spatial frequency component that conversion is put in order to process at least one or the time of more space frequency component change carries out frequency translation, calculate the time space frequency, time space frequency characteristic performance portion 1005, the frequency characteristic data that among the time space frequency of calculating, uses high fdrequency component and low frequency component to calculate the time space frequency characteristic that can show image from the component of at least one or more specific time space frequency.Therefore, can show the time space frequency characteristic of image rightly, can carry out, and the image evaluation in the image be compared, can carry out appropriate evaluation image based on evaluation as the characteristic of the image of consecutive image.

Claims (17)

1. image evaluation device is characterized in that comprising:

Memory cell, storage is about the predetermined datum characteristic as the time space frequency characteristic of image;

The image characteristics analytic unit is calculated the time space frequency characteristic by the object image of a plurality of image constructions; And

Calculate the image evaluation determining unit of the evaluation of estimate of above-mentioned object image based on time space frequency characteristic of calculating by above-mentioned image characteristics analytic unit and the relativeness that is stored in the datum characteristic in the said memory cells.

2. image evaluation device as claimed in claim 1 is characterized in that also comprising:

The frequency component adjustment unit based on the evaluation of estimate of being determined by above-mentioned image evaluation determining unit, to the image that the N between any two images in above-mentioned object image opens, is adjusted the frequency component of regulation, and wherein, N is the integer more than or equal to 1,

Above-mentioned image evaluation determining unit is calculated and is utilized said frequencies component adjustment unit signal frequency component to be carried out the evaluation of estimate of adjusted image.

3. image evaluation device as claimed in claim 1 or 2 is characterized in that:

The formation of above-mentioned image characteristics analytic unit comprises:

Spatial frequency is calculated the unit, calculates the spatial frequency of a plurality of images;

Spatial frequency component arrangement unit, the component transformation that will calculate the spatial frequency of calculating the unit by above-mentioned spatial frequency for and the corresponding data of predetermined benchmark;

The time space frequency is calculated the unit, the time by at least one or more space frequency component in the spatial frequency component of above-mentioned spatial frequency component arrangement unit conversion is changed carry out frequency translation and calculate the time space frequency; And

Characteristic is calculated the unit, utilizes by above-mentioned time space frequency and calculates high fdrequency component and low frequency component among the time space frequency of calculating the unit, calculates the data of the time space frequency characteristic that can show image.

4. image evaluation device is characterized in that comprising:

Memory cell, storage is about the predetermined datum characteristic as the time space frequency characteristic of image;

The 1st image characteristics analytic unit is calculated based on the 1st image characteristics by the time space frequency characteristic of the object image of a plurality of image constructions;

The 2nd image characteristics analytic unit is calculated based on the 2nd image characteristics that above-mentioned object images is limited the time space frequency characteristic in the resulting frequency component restriction of a part of frequency component image; And

The image evaluation determining unit, based on the variable quantity of the 2nd image characteristics of calculating by above-mentioned the 2nd image characteristics analytic unit of above-mentioned the 1st image characteristics of relatively calculating by above-mentioned the 1st image characteristics analytic unit be stored in the relativeness of the datum characteristic in the said memory cells, calculate the evaluation of estimate of the above-mentioned object image of being imported.

5. image evaluation device as claimed in claim 4 is characterized in that:

Above-mentioned the 2nd image characteristics analytic unit limits a part of frequency component by the image that the N between any two images in above-mentioned object image is opened and obtains frequency component restriction image, and wherein, N is the integer more than or equal to 1.

6. as claim 4 or 5 described image evaluation devices, it is characterized in that also comprising:

The frequency component adjustment unit is adjusted the frequency component that is limited by above-mentioned the 2nd image characteristics analytic unit based on above-mentioned evaluation of estimate;

Above-mentioned image evaluation determining unit is calculated the evaluation of estimate of signal frequency component being carried out adjusted image by said frequencies component adjustment unit.

7. image evaluation device as claimed in claim 5 is characterized in that also comprising:

Frequency component limit amount adjustment unit based on above-mentioned evaluation of estimate, changes any two images of being determined the restriction object of frequency component by above-mentioned the 2nd image characteristics analytic unit.

8. as each the described image evaluation device in the claim 3 to 7, it is characterized in that:

Above-mentioned the 1st image characteristics analytic unit comprises:

Spatial frequency is calculated the unit, calculates the spatial frequency of a plurality of images;

Spatial frequency component arrangement unit, the component transformation that will calculate the spatial frequency of calculating the unit by above-mentioned spatial frequency for and the corresponding data of predetermined benchmark;

The time space frequency is calculated the unit, the time by at least one or more space frequency component in the spatial frequency component of above-mentioned spatial frequency component arrangement unit conversion is changed carry out frequency translation and calculate the time space frequency; And

Characteristic is calculated the unit, utilizes by above-mentioned time space frequency and calculates high fdrequency component and low frequency component among the time space frequency of calculating the unit, calculates the data of the time space frequency characteristic that can show image.

9. as each the described image evaluation device in the claim 4 to 6, it is characterized in that:

Above-mentioned the 2nd image characteristics analytic unit comprises:

Spatial frequency is calculated the unit, from the frequency component restriction image that a part of frequency component of object images restriction is obtained, calculates spatial frequency;

Spatial frequency component arrangement unit, the component transformation that will calculate the spatial frequency of calculating the unit by above-mentioned spatial frequency for and the corresponding data of predetermined benchmark;

The time space frequency is calculated the unit, the time by at least one or more space frequency component in the spatial frequency component of above-mentioned spatial frequency component arrangement unit conversion is changed carry out frequency translation and calculate the time space frequency; And

Characteristic is calculated the unit, utilizes by above-mentioned time space frequency and calculates high fdrequency component and low frequency component among the time space frequency of calculating the unit, calculates the data of the time space frequency characteristic that can show image.

10. as each the described image evaluation device in the claim 1 to 9, it is characterized in that:

The said reference characteristic is based on the frequency characteristic that human vision property is determined.

11. a time space frequency-response analysis device is characterized in that comprising:

Spatial frequency is calculated the unit, calculates the spatial frequency of a plurality of images;

Spatial frequency component arrangement unit, the component transformation that will calculate the spatial frequency of calculating the unit by above-mentioned spatial frequency for and the corresponding data of predetermined benchmark;

The time space frequency is calculated the unit, the time by at least one or more space frequency component in the spatial frequency component of above-mentioned spatial frequency component arrangement unit conversion is changed carry out frequency translation and calculate the time space frequency; And

Frequency characteristic is calculated the unit, utilizes by above-mentioned time space frequency and calculates at least one or more high fdrequency component and low frequency component among the time space frequency of calculating the unit, calculates the data of the time space frequency characteristic that can show image.

12. an image evaluation method is characterized in that comprising:

The image characteristics analytical procedure is calculated the time space frequency characteristic by the object image of a plurality of image constructions;

The image evaluation determining step is calculated the evaluation of estimate of above-mentioned object image based on the time space frequency characteristic of being calculated by above-mentioned image characteristics analytical procedure with about the relativeness of the predetermined datum characteristic as the time space frequency characteristic of image.

13. an image evaluation method is characterized in that also comprising:

The 1st image characteristics analytical procedure is calculated based on the 1st image characteristics by the time space frequency characteristic of the object image of a plurality of image constructions;

The 2nd image characteristics analytical procedure, the 2nd image characteristics of calculating the time space frequency characteristic in the frequency component restriction image that obtains based on above-mentioned object images is limited a part of frequency component; And

The image evaluation determining step based on the variable quantity of the 2nd image characteristics of above-mentioned relatively the 1st image characteristics with about the relativeness of the predetermined datum characteristic as the time space frequency characteristic of image, is calculated the evaluation of estimate of the above-mentioned object image of input.

14. a time space frequency-response analysis method is characterized in that comprising:

Spatial frequency is calculated step, calculates the spatial frequency of a plurality of images;

Spatial frequency component arrangement step, the component transformation that will calculate the spatial frequency that step calculates by above-mentioned spatial frequency for and the corresponding data of predetermined benchmark;

The time space frequency is calculated step, the time by at least one or more space frequency component in the spatial frequency component of above-mentioned spatial frequency component arrangement step conversion is changed carry out frequency translation and calculate the time space frequency; And

Frequency characteristic is calculated step, utilizes by above-mentioned time space frequency and calculates high fdrequency component and low frequency component among the time space frequency that step calculates, calculates the data of the time space frequency characteristic that can show image.

15. an image evaluation program is characterized in that comprising:

Memory module, storage is about the predetermined datum characteristic as the time space frequency characteristic of image;

The image characteristics analysis module is calculated the time space frequency characteristic by the object image of a plurality of image constructions; And

The image evaluation determination module based on time space frequency characteristic of being calculated by above-mentioned image characteristics analysis module and the relativeness that is stored in the datum characteristic in the above-mentioned memory module, is calculated the evaluation of estimate of above-mentioned object image.

16. an image evaluation program is characterized in that also comprising:

Memory module, storage is about the predetermined datum characteristic as the time space frequency characteristic of image;

The 1st image characteristics analysis module is calculated based on the 1st image characteristics by the time space frequency characteristic of the object image of a plurality of image constructions;

The 2nd image characteristics analysis module, the 2nd image characteristics of calculating the time space frequency characteristic in the frequency component restriction image that obtains based on above-mentioned object images is limited a part of frequency component; And

The image evaluation determination module, based on the variable quantity of the 2nd image characteristics of calculating by above-mentioned the 2nd image characteristics analysis module of the 1st image characteristics of relatively calculating by above-mentioned the 1st image characteristics analysis module be stored in the relativeness of the datum characteristic in the above-mentioned memory module, calculate the evaluation of estimate of the above-mentioned object image of input.

17. a time space frequency-response analysis program is characterized in that comprising:

Spatial frequency is calculated module, calculates the spatial frequency of a plurality of images;

The spatial frequency component sorting module, the component transformation that will calculate the spatial frequency that module calculates by above-mentioned spatial frequency for and the corresponding data of predetermined benchmark;

The time space frequency is calculated module, the time by at least one or more space frequency component in the spatial frequency component of above-mentioned spatial frequency component sorting module conversion is changed carry out frequency translation and calculate the time space frequency; And

Frequency characteristic is calculated module, utilizes by above-mentioned time space frequency and calculates high fdrequency component and low frequency component among the time space frequency that module calculates, calculates the data of the time space frequency characteristic that can show image.

Images