JP3772514B2 - Image evaluation system - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
  The present invention relates to an image evaluation apparatus, and more particularly to an image evaluation apparatus that evaluates how a person feels information visually given from a design image.
[0002]
[Prior art]
  In recent years, research on image processing and retrieval based on human sensitivity has been actively conducted, and in particular, there have been many proposals of sensitivity evaluation devices for evaluating human sensitivity and image analysis devices for analyzing physical features of images. Has been made.
[0003]
  In the sensitivity evaluation performed by the sensitivity evaluation device, a psychological evaluation method is mainly used. In particular, when investigating how users can accept design expressions, especially the combination of colors and shapes for expressing intentions in images, the Semantic Differential Method, which has been almost established as a psychological evaluation method. (Hereinafter referred to as the SD method) is often used.
[0004]
  The SD method is a method for analyzing an image given by an evaluation object (object or word) by allowing a subject to evaluate a scale composed of many adjective pairs.
  FIG. 13 is a diagram showing a part of the result of evaluating a certain image by a conventional general SD method. In the conventional general SD method, the reproducibility of evaluation of each evaluation object is not so good as shown in the figure.
[0005]
  That is, the evaluation results for subject A (represented by a solid line) and subject B (represented by a dotted line) are significantly different, and for example, it cannot be determined whether the evaluation target has a clear impression or a vague impression. .
[0006]
  This is very important when, for example, determining the quality of the evaluation when determining the design concept. The main cause of this poor reproducibility is that the subjective judgment criteria set before each subject starts evaluation differ from person to person.
[0007]
  As countermeasures, procedures such as showing the entire sample to be evaluated to the subject in advance for evaluation or preparing an evaluation scale and performing evaluation by comparing with the evaluation scale are generally used.
[0008]
  However, if the sample group to be evaluated has a distribution that is biased in terms of sensibility or physical properties, or if there is no significant difference, that is, the sample distribution space is narrow, this problem is difficult to avoid. I cannot help it.
[0009]
  In addition, for this problem, a method of improving accuracy by applying a combination of a plurality of psychological evaluation methods may be used. However, with such a psychological method alone, the process of evaluation of what and how the subject sees and evaluates is hardly known, and thus the accuracy cannot always be improved.
[0010]
  For this reason, in JP-A-5-108002, means for detecting the position of the viewpoint of a subject who performs sensitivity evaluation together with the SD method is provided, and the accuracy is improved by combining viewpoint data.
[0011]
[Problems to be solved by the invention]
  However, in the conventional technology as described above, since only the position of the subject's viewpoint is detected, relatively simple image information and individual objects with no overlapping content are separated and independent, and the position information and the target It can be said that it is effective in a real space where the relationship with the object is clear, but when multiple objects are located in a complex image or narrow space where the contents overlap, and the relationship between the position information and the object is difficult to attach However, there was a problem that improvement in evaluation accuracy could not be expected.
[0012]
  Moreover, in the sensitivity evaluation with respect to the image etc. as described above, the psychological method alone, as represented by the SD method, hardly reveals the evaluation process of what the subject sees and how to evaluate. There was a problem.
[0013]
  Furthermore, even when the viewpoint information is used together, for example, when the background image and the main motif overlap in a specific area of the evaluation target image, whether the subject's psychological evaluation result comes from the background image or receives from the main motif It was impossible to judge whether it was an impression, and there was a problem that an accurate evaluation result could not be obtained.
[0014]
  The present invention has been made in view of these points, and is necessary for searching and processing images and illustrations based on people's preferences and sensibilities, and for supporting the design production of documents and web pages. It is an object of the present invention to provide an image evaluation apparatus capable of improving the accuracy of sensitivity evaluation for image information and obtaining sensitivity information efficiently.
[0015]
[Means for Solving the Problems]
  In the present invention, in order to solve the above problems, in an image image evaluation apparatus for evaluating how a person feels for information visually given from a design image, image information display means for displaying image information, Sensitivity evaluation data input means in which a subject performs sensitivity evaluation on the image information and inputs sensitivity evaluation data, visual characteristic detection means for detecting a visual characteristic that is a visual function state of the subject performing the sensitivity evaluation, and detection And analyzing the observation state of the subject with respect to the image information based on the visual characteristics thus obtained,SetGenerate visual characteristic data that is informationAndSaidThe position information of the viewpointFrom the visual characteristic data, the subject focuses on which part of the image informationTo generate observation position analysis informationAnalysis means;From the sensitivity evaluation data and the observation position analysis information, an image image calculation means for associating the image information with an impression when the subject observes the image information,There is provided an image evaluation apparatus characterized by comprising:
[0016]
  Here, the image information display means displays the image information. The sensitivity evaluation data input means allows the subject to perform sensitivity evaluation on the image information and inputs sensitivity evaluation data. The visual characteristic detection means detects a visual characteristic that is a visual function state of a subject who performs sensitivity evaluation. The analysis means analyzes the observation state of the subject's image information based on the detected visual characteristics and determines the position of the subject's viewpoint.SetGenerate visual characteristic data that is informationAnd the position information of the viewpointFrom the visual characteristic data, the subject pays attention to which part of the image information.To analyze observation position and generate observation position analysis information. The image image calculation means associates the image information with the impression when the subject observes the image information from the sensitivity evaluation data and the observation position analysis information.
[0017]
DETAILED DESCRIPTION OF THE INVENTION
  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG.Is a pictureIt is a principle figure of an image image evaluation apparatus. The image image evaluation apparatus 10 evaluates how a person feels with respect to information visually given from a design image.
[0018]
  The image information display means 1 displays image information (evaluated image). The sensitivity evaluation data input unit 2 inputs sensitivity evaluation data by subject A performing sensitivity evaluation on image information.
[0019]
  The visual characteristic detection means 3 detects the visual characteristic that is the visual function state of the subject A who performs the sensitivity evaluation.
  Based on the detected visual characteristic, the analysis unit 4 performs an analysis process of the characteristic of the observation position and the image information, and generates observation data and characteristic data (evaluated image characteristic data).
[0020]
  The image image calculation means 5 calculates the impression of the image information from the sensitivity evaluation data, the observation data, and the characteristic data. The image image calculation unit 5 includes a function of a determination unit.
[0021]
  NextPaintingThe configuration and operation of the image evaluation apparatus 10 will be described in detail. FIG. 2 is a diagram illustrating a basic hardware configuration of the image image evaluation apparatus 10, and FIG. 3 is a system diagram illustrating a relationship between operation modules of the image image evaluation apparatus 10.
[0022]
  The hardware of the image image evaluation apparatus 10 includes an image information display means 1, a sensitivity evaluation data input means 2, a visual characteristic detection means 3, an analysis means 4, an image image calculation means 5, a central processing unit 6, and an image information input means 7. The image information storage means 8 is configured as a computer system.
[0023]
  The central processing unit 6 is configured mainly with an MPU, RAM, ROM, and the like, and performs overall control of the above system elements via a system bus. The analysis unit 4 and the image image calculation unit 5 are included.
[0024]
  The image information input means 7 includes an image scanner for converting an arbitrary hard-copy image displayed in color into image data, a video interface for capturing image data from the outside in the form of a video signal, and the like.
[0025]
  Note that an apparatus of a type that reads a reflection type image typified by a print image or the like and a transmission type image typified by a photographic slide or the like is connected to the image scanner as necessary.
[0026]
  In addition, a VCR, a TV, and the like are similarly connected to the video interface as necessary.
  The image information storage means 8 includes a general hard disk device and a large-capacity storage device such as a write-once optical disk and a rewritable optical disk. The image information storage means 8 includes a large number of evaluation objects. Images and documents are stored.
[0027]
  Then, observation data for an evaluation image to be described later, characteristic data of the evaluation image analyzed based on them, sensitivity evaluation data, and image image data which is the final output of the apparatus are stored.
[0028]
  The image information display means 1 is composed of a CRT display, LCD, etc., and at least one of the image to be evaluated and the visual characteristic detection information inputted as described later is displayed in color on the display screen simultaneously or sequentially. The
[0029]
  The visual characteristic detection means 3 is a corneal reflection type or scleral reflection type eye camera, and detects the reflected light of the eye with respect to an irradiation light source such as a field-of-view imaging device and an infrared LED mounted on the subject's head. Configured by the device.
[0030]
  As a basic function of the visual characteristic detection means 3, the eye movement of the subject is measured, and the center point of the visual point perceived by the subject (hereinafter referred to as a gaze point) is used as a coordinate.Arithmetic processingThis is output to the unit 6.
[0031]
  The central processing unit 6 determines the position of the image information area as the evaluation target presented by the image information display means 1 on the coordinates in the visual field of the subject corresponding to the position of the head. Is preset.
[0032]
  Therefore, it is possible to know which image information the subject is viewing based on the coordinate data of the viewpoint output from the visual characteristic detection means 3 and the coordinate data of the image information area preset in the central processing unit 6.
[0033]
  Further, as the visual characteristic detection information, in addition to the position coordinate detection of the gazing point described above, information on the degree of adjustment for the brightness of the input information and information on the degree of adjustment for the depth of the input information are used.
[0034]
  These visual characteristic information is detected, for example, by detecting the pupil contour coordinates from the eyeball image irradiated with the infrared LED, calculating the ellipse major axis by ellipse approximation, and determining the diameter of the pupil with respect to the brightness adjustment degree. can do.
[0035]
  Regarding the focus adjustment, the reflected light amount distribution from the fundus of the irradiated infrared LED can be detected by detecting the luminance by the knife edge method and obtaining the distribution gradient.
[0036]
  The visual characteristic detection method described here is merely an example, and other methods may be used. Further, in the present embodiment, the visual characteristic detection means 3 has been described with a goggle type in which a visual field imaging device and a device for detecting reflected light of an eye with respect to an irradiation light source such as an infrared LED are mounted on the subject's head. When the image to be evaluated is displayed in a fixed space such as a display, by using a contact-type visual characteristic detection device in which the visual field imaging device, the irradiation light source, and the detection device are fixed to the side of the display The burden on the subject can be reduced.
[0037]
  The sensibility evaluation data input means 2 is composed of, for example, a keyboard and a mouse connected to a computer. The subject's sensitivity evaluation data for the evaluation item displayed on the monitor or the like is input via a keyboard or the like.
[0038]
  In addition, the display of evaluation items may be displayed every time all sensitivity evaluation is input, or may be sequentially displayed by giving an instruction via a keyboard.
[0039]
  In addition, the sensitivity evaluation data may be input by the subject himself / herself by looking at the evaluation items displayed on the monitor, or the operator asks the subject about the evaluation items and the sensitivity evaluation answered by the subject is performed by the operator. May be entered.
[0040]
  The analysis means 4 is information such as the observation position of the subject based on the visual characteristic data such as the gaze point, brightness, and focus adjustment degree of the subject obtained by the visual characteristic detection means 3 described above and the image data to be evaluated. Observation position analyzing means 4b for analyzing the image and image characteristic analyzing means 4a for analyzing the characteristics of the image to be evaluated.
[0041]
  The image image calculation means 5 calculates the image of the image information based on the sensitivity evaluation data obtained by the sensitivity evaluation data input means 2, the observation data obtained by the analysis means 4, and the evaluated image characteristic data. It is configured.
[0042]
  Next, the operation of the image evaluation apparatus 10 will be described below. If the image to be evaluated is stored in the image information storage means 8 in advance, it is called from there, and if the form is hard copy, it is also taken into the apparatus from the image information input means 7 and the image information is stored as necessary. Stored in the storage means 8.
[0043]
  Next, sensory evaluation by a subject is performed based on these image groups to obtain sensitivity evaluation data. Here, a sensory evaluation method will be described.
  FIG. 4 is a diagram showing an outline of the operation status of the image evaluation apparatus. The evaluated image S stored in the image information storage means 8 is sent to the image information display means 1 and presented to the subject.
[0044]
  At this time, when comparing the impressions of a plurality of images to be evaluated, if the evaluation target images are a small number of about 2 to 3 types, they are presented simultaneously, and if there are more than 2 types, 2 to 3 types are presented. Or it is desirable to present it in several groups for convenience.
[0045]
  The visual function state of the subject is measured by the visual characteristic detection means 3 from the time when the image to be evaluated is disclosed to the subject. In the figure, the gaze point E of the subject is shown. Next, image information as an evaluation target is presented one by one.
[0046]
  Then, the subject evaluates the presented image information by the SD method, and inputs sensitivity evaluation data for each evaluation item D from the keyboard in five stages.
  When the subject performs sensitivity evaluation for all evaluation targets, the input of sensitivity evaluation data is terminated by pressing a predetermined key on the keyboard.
[0047]
  In this embodiment, the evaluation item is displayed on another monitor from the viewpoint of accuracy. However, the image display and the evaluation item display can be simply performed on the same display.
[0048]
  Next, a method for calculating the observation state of the subject by the analyzing unit 4 will be described. First, regarding the detection of the gazing point as to which part of the image the subject is observing is derived from the relationship between the eye movement of the subject and time.
[0049]
  Specifically, it is performed by detecting whether or not the subject's viewpoint stays at 100 msec or more within a predetermined range (within a radius of 1 °) centering on the corresponding position from the position at a certain time.
[0050]
  This is because it is considered that the subject cannot recognize the information represented by the image in a short time of 100 msec or less.
  From the coordinate position obtained as a result and the data of the image to be evaluated, it is possible to capture which content of the image the subject is observing as an observation time axis.
[0051]
  Finally, by combining the gazing point data obtained as described above with the image image calculation means 5 and the sensitivity evaluation data obtained by the SD method or the like, when observing the specific content in the image to be evaluated It is possible to grasp the relationship between the image and the impression factor of the image to be evaluated.
[0052]
  Next, the operation of the image evaluation apparatus 10 will be described with reference to a flowchart. FIG. 5 is a diagram showing an operation flowchart of the image evaluation apparatus 10.
[S1] The image to be evaluated is displayed on a display device such as a calling display.
[S2] Sensitivity evaluation of the image to be evaluated is performed by the subject to obtain sensitivity evaluation data.
[S3a] At the same time as step S2, the visual characteristic detecting means 3 monitors the visual function state.
[S3b] At least one of observation data and evaluated image characteristic data is calculated based on the monitor information.
[S4] It is determined whether there are other sensitivity evaluation items. If there is, return to step S2, otherwise go to step S5.
[S5] The impression of the evaluated image is calculated based on the sensitivity evaluation data, the observation data, and the evaluated image characteristic data.
[S6] It is determined whether there is another image to be evaluated. If there is, return to step S1 and repeat for the number of images to be evaluated. If not, exit.
[0053]
  Here, the accuracy of the sensitivity evaluation result by the SD method of the present embodiment will be described in comparison with the evaluation result in the conventional general evaluation method. As described above, FIG. 13 shows a part of the result evaluated by a so-called general SD method in which a sensuous evaluation is performed while observing only an image to be evaluated of “a red balloon flying in the sky against a backdrop of mountains”. It is shown.
[0054]
  From this result, it can be seen that in the conventional evaluation method, the sensitivity evaluation result for this image has a large difference in the evaluation values of both the subjects A and B.
  From this evaluation result, for example, one person feels “warm” while another person feels “cold” very much.
[0055]
  Similarly, both “static-dynamic” and “unique-mediocre” have conflicting results.
  Next, FIG. 6 and FIG. 7 show the results of detecting the gazing point at the time of sensitivity evaluation of both subjects A and B according to the present embodiment. 6 and 7 are diagrams showing the results of detecting a gazing point at the time of sensitivity evaluation of subjects A and B. FIG.
[0056]
  When these detection results and the above-mentioned sensitivity evaluation results are collated, the following can be understood. First, since subject A mainly concentrates on the “red balloon” of the image to be evaluated, gaze point E1 concentrates on the subject, and the sensitivity evaluation results such as “soft”, “dynamic”, and “warm” of subject A are , It turns out that it is a result obtained by paying attention to the content of this “red balloon”.
[0057]
  On the other hand, the subject B has a comparatively gazing point E2 mainly in the background “mountain”, so that the sensitivity evaluation results of subject B such as “sharp”, “static”, “cold”, etc. Pays attention to the content of “Yamanami” and understands that it is the result obtained.
[0058]
  As described above, in this embodiment, it can be seen that there is no significant difference between the evaluation values of both subjects A and B. From this result, it is possible to detect visual characteristics at the time of sensitivity evaluation for a subject and take correspondence with sensitivity evaluation data obtained by the SD method or the like.TheIndividual differences in evaluation results can be reducedThe
[0059]
  Moreover, although the difference of the evaluation result of a different test subject was shown this time, in this embodiment, it can also be expected that the variation when one test subject repeatedly evaluates can be greatly reduced.
[0060]
  As explained above, PaintingimageImage evaluationThe device 10 is visually given image information to a subject, and a sensitivity evaluation is performed on this information.
  At this time, the visual function state of the subject is monitored, and this detection information is sent to the analysis means 4 to calculate the observation position and the characteristics of the image to be evaluated.
[0061]
  Then, the image image calculation means 5 calculates the impression of the evaluated image information based on the sensitivity evaluation data, the observation data, and the characteristic data of the evaluated image.
[0062]
  Therefore, by using the observation data of the subject and the characteristic data of the image to be evaluated in addition to the sensitivity evaluation data, the accuracy of the sensitivity evaluation for the image information can be improved and the sensitivity information can be obtained efficiently.
  NextOn the secondThe second embodiment will be described. FIG. 8 is a system diagram showing the relationship of the operation modules of the image / image evaluation apparatus 10 according to the second embodiment.
[0063]
  The difference from the first embodiment is that image region dividing means 9 for the image to be evaluated is added.
  The image area dividing means 9 is an arithmetic processing unit having a function of dividing the image data of the image to be evaluated into a plurality of areas such as a main motif area and a background area.
[0064]
  Generally, it is composed of retoucher software represented by Adobe's photo design production software Photoshop or the like, or a user interface necessary for operating it.
[0065]
  Then, the operator selects and divides an arbitrary area using a mouse or the like while viewing the image displayed on the display.
  The image area dividing means 9 is constituted by image processing hardware or software to which an image area dividing algorithm represented by clustering by the K-average algorithm is applied, and the area dividing is automatically executed according to the image feature amount. It has come to be.
[0066]
  In the analysis means 4, the subject can determine which of the images to be evaluated based on the region data of each content such as the main motif and the background divided by the image region dividing means 9 and the gazing point data of the subject monitored by the visual characteristic detection means 3. Whether the content has been observed for the longest time is calculated, and based on this result, the region of interest for the subject's evaluation image is predicted.
[0067]
  Then, the image image calculation means 5 receives the result and calculates the impression of the evaluated image information by taking correspondence with the sensitivity evaluation data of the subject.
  Therefore, by adding the region division information of the image to be evaluated, the factor of the sensitivity evaluation result is clarified, and effective sensitivity information can be obtained.
[0068]
  Next, a specific example of the second embodiment will be described. FIG. 9 is a diagram showing a result of dividing the image to be evaluated into contents.
  The “red balloon flying in the sky against the back of the mountain” is divided into three “red balloons” that are the main motif M0 by the image region dividing means 9, “yamanami” that is the sub-motif M1, and “sky” in the background M2. The result of dividing into contents is shown.
[0069]
  FIG. 10 is a diagram showing the rate of gaze time. The figure shows how much each subject pays attention to which region (content) based on the region in FIG. 9 and the result of detecting the gazing point at the time of sensitivity evaluation of subjects A and B (see FIGS. 6 and 7). Is expressed as a percentage of gaze time.
[0070]
  From this result, subject A spends most of his time on the “red balloon” that is the main motif M0, whereas subject B is focusing on the “mountain” that is the sub-motif M1. I understand.
[0071]
  When these analysis results are compared with the sensitivity evaluation results, the sensitivity evaluation results such as “soft”, “dynamic”, and “warm” of the subject A are mainly the contents of “red balloon” in the image to be evaluated. It turns out that it is a result obtained paying attention to.
[0072]
  On the other hand, it is understood that “sharp”, “static”, “cold”, etc., which are the sensitivity evaluation results of the subject B, are results obtained mainly by paying attention to “mountain” of the sub-motif M1.
[0073]
  Therefore, in order to confirm the above-mentioned result, the subject B was focused on the “red balloon” of the main motif M0 considered to have been focused by the subject A from the gazing point detection result, and the sensitivity evaluation was performed again. Results as shown in FIG. 11 were obtained.
[0074]
  FIG.Is a pictureIt is a figure which shows an image evaluation result. As shown in the figure, it can be seen that the result of the subject A has been approached in each item of sensitivity evaluation.
  As described above, by adding the detection information of the visual characteristics and the area division information of the image to be evaluated to the sensitivity evaluation such as the SD method, the factor of the sensitivity evaluation result becomes clear and effective sensitivity information can be obtained. it can.
[0075]
  Moreover, although the difference of the evaluation result of a different test subject was shown this time, in this embodiment, it can also be expected that the variation when one test subject repeatedly evaluates can be greatly reduced.
[0076]
  As explained aboveThe secondIn the second embodiment, a plurality of area division processes are performed on the image to be evaluated in advance, and the area division information is sent to the analysis means 4. Then, the analysis unit 4 predicts a region of interest for the subject's evaluation image based on the observation data of the subject monitored by the visual characteristic detection unit 3 and the evaluation image region division data.
[0077]
  ImageCalculationThe means 5 receives this result and calculates the impression of the evaluated image information by taking correspondence with the sensitivity evaluation data of the subject.
  Therefore, by adding the area division information of the image to be evaluated, the factor of the sensitivity evaluation result is clarified, and effective sensitivity information can be obtained.
[0078]
  NextOn the secondA third embodiment will be described. FIG. 12 is a system diagram showing the relationship among the operation modules of the visual characteristic detecting means 3, the analyzing means 4 and the image image calculating means 5 of the image image evaluation apparatus 10 according to the third embodiment.
[0079]
  The difference between the third embodiment and the second embodiment is that the visual characteristic detection means 3 has a brightness adjustment degree detection means 3a, a gaze point coordinate detection means 3b, and a focus adjustment degree detection means 3c. The subject's gazing point, brightness adjustment degree, and focus adjustment degree are detected, and the observation position of the subject is analyzed more accurately.
[0080]
  The analysis unit 4 includes a first observation position content prediction unit 4c-1, a second observation position content prediction unit 4c-2, and an evaluated image content prediction unit 4c-3.
[0081]
  The first observation position content prediction unit 4c-1 analyzes whether the observation position is a highlight portion or a shadow portion from the gazing point data and the brightness adjustment degree data, and the entire content expressed by the highlight portion or the shadow portion. Or whether or not attention is paid to a part of contents in a highlight part or a shadow part.
[0082]
  The second observation position content prediction unit 4c-2 predicts whether the observation position is a foreground content or a background based on the observation position data, the focus adjustment degree data, and the evaluated image area division data.
[0083]
  The to-be-evaluated image content prediction unit 4c-3 predicts the motif, content, situation, scene, and the like of the to-be-evaluated image combining them.
  The impression evaluation unit 5a of the image to be evaluated obtains the impression of the image to be evaluated by using the prediction result of the content thus obtained, the observation position of the subject, and the sensitivity evaluation data. It is possible.
[0084]
  As explained aboveThe secondIn the third embodiment, at least one characteristic value of the position of the subject's viewpoint, the brightness adjustment degree, and the focus adjustment degree is detected and used to derive the observation position of the subject and the characteristics of the image to be evaluated.
[0085]
  Therefore, even for more complex images and documents, the gaze point of the subject can be captured with high accuracy, and more characteristic values of the image to be evaluated can be predicted, so that detailed sensitivity evaluation can be performed.
[0086]
  Furthermore, in the third embodiment, the content of the content at the observation position is predicted based on the information detected by the visual characteristic detecting means 3 and the evaluated image data, and the result is evaluated based on the result and the sensitivity evaluation data of the subject. The impression of the evaluation image information is calculated.
[0087]
  Therefore, by adding the prediction information of the content of the image to be evaluated, the factor of the sensitivity evaluation result becomes clear, and more accurate sensitivity information can be obtained.
[0088]
【The invention's effect】
  As explained above, PaintingThe image image evaluation apparatus analyzes the observation state of the subject's image information, and generates visual characteristic data that is at least one or more of the position of the subject's viewpoint, the brightness adjustment degree and the focus adjustment degree for the image information. Then, from the visual characteristic data, it is configured to determine which part of the image information the subject has input the sensitivity evaluation data. Thereby, the accuracy of sensitivity evaluation for image information is improved, and it becomes possible to efficiently obtain sensitivity information.
[Brief description of the drawings]
[Figure 1]PictureIt is a principle figure of an image image evaluation apparatus.
FIG. 2 is a diagram showing a basic hardware configuration of an image evaluation apparatus.
FIG. 3 is a system diagram showing a relationship among operation modules of the image evaluation apparatus.
FIG. 4 is a diagram illustrating an outline of an operation state of the image evaluation apparatus.
FIG. 5 is a diagram showing an operation flowchart of the image evaluation apparatus.
FIG. 6 is a diagram illustrating a result of detecting a gazing point at the time of sensitivity evaluation of a subject A and a subject B.
FIG. 7 is a diagram showing a result of detecting a gazing point at the time of sensitivity evaluation of subjects A and B.
FIG. 8 is a system diagram showing the relationship of operation modules of the image evaluation apparatus according to the second embodiment.
FIG. 9 is a diagram illustrating a result of dividing an image to be evaluated into contents.
FIG. 10 is a diagram showing a ratio of gaze time.
FIG. 11PictureIt is a figure which shows an image evaluation result.
FIG. 12 is a system diagram showing a relationship between operation modules of visual characteristic detection means, analysis means, and image image calculation means of the image image evaluation apparatus according to the third embodiment;
FIG. 13 is a diagram showing a part of the result of evaluating a certain image by a conventional general SD method.
[Explanation of symbols]
  1 Image information display means
  2 Kansei evaluation data input means
  3 visual characteristic detection means
  4 Analysis means
  5 Image image calculation means
  10 Image evaluation device
  A Subject

Claims (4)

In an image evaluation device that evaluates how people feel about information given visually from design images,
Image information display means for displaying image information;
Sensitivity evaluation data input means for inputting sensitivity evaluation data by subject performing sensitivity evaluation on the image information,
Visual characteristic detection means for detecting a visual characteristic which is a visual function state of a subject performing the sensitivity evaluation;
Based on the detected visual characteristic, analyzes the observation state for said image information of said subject, generating a visual characteristic data is information position of the subject's viewpoint, visual characteristics is position information of the viewpoint Analyzing means for generating observation position analysis information by analyzing which part of the image information the subject focuses on from the data ,
From the sensitivity evaluation data and the observation position analysis information, an image image calculation means for associating the image information with an impression when the subject observes the image information,
The image evaluation apparatus characterized by having.   The image image evaluation apparatus according to claim 1, wherein the analysis unit includes a content prediction unit that predicts the content of the content at the observation position.   2. The image image evaluation apparatus according to claim 1, further comprising an image area dividing unit for dividing the image information into a plurality of areas to generate a divided area. It said analyzing means, for each divided image area, analyzes the observation state for said image information of said subject, picture image evaluation apparatus according to claim 3, wherein the generating the visual characteristic data.

Images