JP2007181071A - Apparatus and method for evaluating attention paid to contents - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an apparatus and method for evaluating attention paid to contents whereby effects of advertisement media and advertisement programs can more meticulously and accurately be evaluated. <P>SOLUTION: The apparatus for evaluating the attention paid to contents includes: a system server 200 provided with a passer detection section 202 including a moving region detection section 202a for detecting a moving region from an image photographed by a camera 150 and a person detection section 202b for detecting whether or not the moving region is a person, and provided with an attention paying person detection section 203 including a face detection processing section 203a for detecting the face of the person, and a face angle detection processing section 203b for detecting the angular direction of the face; and a collection center server 300 provided with a passing detection time registration section 302 for registering a face angle detection time of a passer detected by the passerby detection section 202 and a face angle registration processing section 303 for registering a face angle detected by the face angle detection processing section 203b. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a content attention evaluation device that evaluates the attention of a person existing in front of a content medium to the content based on an image captured by an imaging device of the person existing in front of the content medium. The present invention relates to a content attention evaluation apparatus and an evaluation method for storing a face angle of a person existing in front of a medium with respect to the content together with a change in time and evaluating the attention to the content.

  Originally, contents such as advertisements and TV broadcasts only show their significance after attracting more people's attention. Therefore, advertisement publishers, TV broadcast sponsors, etc. plan and install advertisements to attract people's attention efficiently, or become TV broadcast sponsors watched by many people. Therefore, they have a strong awareness of whether or not to increase the effectiveness of TV commercials. For this reason, advertisement publishers and TV broadcast sponsors are highly interested in evaluating how much people are interested in the content.

  Against this background, various methods for evaluating people's attention to content have been devised. For example, in Patent Document 1, a human face is detected from an image captured by a camera capable of capturing the front of an advertisement medium, and a face that can be evaluated as facing the advertisement medium is detected from the detected faces. An advertising media evaluation device is disclosed that evaluates how many people have focused on the advertising media by counting.

  Patent Document 2 discloses the following advertisement effect confirmation system. That is, the viewer's reaction to the advertisement program is photographed in an image, and data indicating the characteristics of the advertisement program when the image is photographed is added and transmitted to the analyzer together with the photographed image. In the analysis apparatus, the received images themselves are arranged and output in time series, so that not only the viewer's reaction to the advertising program can be visually and timely confirmed but also included in the received images. By detecting and counting the characteristic reactions of viewers, it is possible to check how many viewers are paying attention to the advertising program.

JP 2000-209578 A JP-A-11-153777

  However, the conventional techniques represented by Patent Documents 1 and 2 described above can grasp only the total number of people who are paying attention to advertising media and advertising programs. That is, it does not provide a material for determining how much attention a person who has passed through in front of an advertisement medium or an advertisement program has noticed the advertisement medium or the advertisement program with the passage of time. Therefore, there is a problem in that it is not possible to analyze in detail the attention of the person who has passed in front of the advertisement medium or the advertisement program to the advertisement medium or the advertisement program, and it is impossible to make a sufficiently meaningful evaluation. .

  Specifically, according to Patent Document 1, it is possible to ascertain the number of people who have paid attention to the advertising media, but at what depth each person who has paid attention to the advertising media has paid attention to the advertising media. It was not possible to grasp the time change of the depth of attention. For example, even if statistical data indicating that 10 people have paid attention to the advertising media is obtained, the evaluation results for the effectiveness of the advertising media differ depending on how much attention each person has shown to the advertisement. However, it was impossible to make such a strict evaluation.

  Moreover, in the said patent document 2, the objectivity of evaluation cannot be maintained including the process that the effect of an advertising program must still be evaluated visually. In addition, according to this conventional technique, it is possible to obtain statistics on how many viewers have been paying attention to the advertising program. Like the prior art of Patent Document 1, each person who has paid attention to the advertising program It was not possible to grasp the time change of the attention depth as to whether or not the advertising media was noticed at a certain depth, and the effect of the advertising program could not be accurately evaluated.

  For these reasons, it has become an extremely important issue to establish a method that can grasp the effects of advertising media and advertising programs in more detail and evaluate them more accurately. In particular, advertising media and advertising programs require huge advertising costs, and even if they are provided to viewers in an ineffective manner, the advertising costs are wasted. Have a very strong interest in what to offer. As a guideline for effectively providing advertisement media and advertisement programs, establishment of a more detailed and accurate evaluation method for the effects of advertisement media and advertisement programs has been awaited.

  The present invention has been made to solve the above problems (problems), and an object thereof is to provide a content evaluation apparatus and a content evaluation method capable of evaluating the effects of advertising media and advertising programs in more detail and accurately. And

  In order to solve the above-described problem and achieve the object, the present invention is based on an image obtained by capturing an image of a person existing in front of the content medium by the imaging apparatus, and attention is paid to the content of the person existing in front of the content medium. A content attention evaluation device for evaluating a person, a person detection unit for detecting a person imaged in the image, a face detection unit for detecting a face of a person detected by the person detection unit, and the face detection unit Face angle detection means for detecting the angle of the face direction of the person whose face is detected, and face angle storage means for storing the angle of the face direction detected by the face angle detection means as time passes. It is characterized by that.

  Further, the present invention is characterized in that, in the above invention, the time is a time when the angle of the face direction is detected by the face angle detecting means.

  Further, the present invention is the above invention, wherein the face angle storage means stores the angle of the face direction with the passage of time by the face angle detection means for each person detected by the person detection means. And

  In the present invention, the face direction angle stored in the face angle storage unit is summed up in units of persons detected by the person detection unit, and the sum is detected as the face direction angle. It further comprises gaze depth rate calculation means for calculating a gaze depth rate divided by the time from the start to the end and the angle threshold at which the angle of the face direction can be regarded as the angle at which the content is watched. Features.

  Further, in the present invention, in the above invention, the time during which the angle of the face direction stored in the face angle storage means is temporarily below an angle threshold that can be regarded as an angle at which the content is watched, The apparatus further comprises an empty gaze rate calculating means for calculating an empty gaze rate divided by a time from the start to the end of detection of the face direction angle.

  The present invention further includes a person position detecting means for detecting the position of the person detected by the person detecting means in the above invention, wherein the passage of time is the position of the person detected by the person position detecting means. It corresponds to the change of.

  Further, the present invention is the above invention, wherein the face angle storage means calculates the angle of the face direction detected by the face angle detection means for each person detected by the person detection means, and the person position detection means. This is stored together with the position of the person detected by.

  In the present invention, the face direction angle stored in the face angle storage unit is summed up in units of persons detected by the person detection unit, and the sum is detected as the face direction angle. A moving gaze depth rate calculating means for calculating a moving gaze depth rate by dividing the moving distance of the person from the start to the end and the angle threshold at which the angle of the face direction can be regarded as the angle at which the content is watched. Is further provided.

  Further, in the present invention according to the above invention, the person moves while the angle of the face direction stored in the face angle storage means is temporarily below an angle threshold that can be regarded as an angle at which the content is watched. A moving air gaze rate calculating means for calculating a moving air gaze rate that is obtained by dividing the movement distance by the movement distance that the person has moved from the start of the detection of the angle in the face direction to the end thereof. Features.

  Further, the present invention is the above invention, wherein the content attention evaluation device includes: an image analysis device connected to the imaging device; and a content attention evaluation analysis device connected to be communicable with at least one of the image analysis devices. The image analysis device is configured to transmit data that cannot identify a person included in the image to the content attention evaluation analysis device, and the content attention evaluation analysis device stores data that cannot identify the person. The face angle storage means is provided.

  Further, the present invention is a content attention evaluation method for evaluating the attention of a person existing in front of the content medium on the content based on an image obtained by imaging an image of the person existing in front of the content medium, A person detection step for detecting a person imaged in the image, a face detection step for detecting the face of the person detected by the person detection step, and an angle of the face direction of the person whose face is detected by the face detection step And a face angle storage step of storing the angle of the face direction detected by the face direction detection step together with the time of the detection.

  Further, the present invention is a content attention evaluation method for evaluating the attention of a person existing in front of the content medium to the content based on an image obtained by imaging the person existing in front of the content medium, A person detection step for detecting a person imaged in the image, a person position detection step for detecting the position of the person detected by the person detection step, and a face for detecting the face of the person detected by the person detection step A step of detecting, a face direction detecting step of detecting a face direction angle of the person whose face is detected by the face detecting step, and an angle of the face direction detected by the face direction detecting step. And a face angle storing step of storing together with the position of the person detected by the above.

  According to the present invention, since the face angle storage means for storing the face direction angle detected by the face angle detection means with the passage of time is provided, attention is paid to the content based on the face direction angle that changes with time. There is an effect that analysis is possible.

  In addition, according to the present invention, it is possible to analyze the attention to the content based on the face direction angle that changes with the time when the face direction angle is detected by the face angle detection unit.

  In addition, according to the present invention, it is possible to analyze the attention to the content based on the angle of the face direction that changes with time for each person.

  In addition, according to the present invention, by calculating the average ratio of the angle of the face direction that changes with time to the angle threshold that can be regarded as the angle at which the content is watched, according to the total gaze time, There is an effect that the gaze on the content can be comprehensively grasped and evaluated regardless of the time change.

  In addition, according to the present invention, since the gaze rate calculating means for calculating the gaze rate is provided, it can be regarded as a continuous gaze when the gaze rate is a predetermined value or less, There is an effect that the gaze evaluation can be performed in a form that more closely follows.

  In addition, according to the present invention, it is possible to analyze the attention to the content based on the angle of the face direction that changes with the position of the person detected by the person position detecting means.

  In addition, according to the present invention, it is possible to analyze the attention to content based on the angle of the face direction that changes with the position of the person for each person.

  Further, according to the present invention, by calculating the average ratio of the angle of the face direction that changes with the position of the person with respect to the angle threshold that can be regarded as the angle at which the content is watched, by the total moving distance, In this way, the gaze on the content in between can be grasped and evaluated as a whole, regardless of the change in the position of the person.

  In addition, according to the present invention, since the mobile air gaze rate calculating means for calculating the mobile air gaze rate is provided, when the mobile air gaze rate is equal to or less than a predetermined value, it is regarded as the state where gaze is continuous. It is possible to perform an evaluation of gaze accompanied by movement in a form that more closely follows.

  Further, the present invention has an effect that the content attention evaluation device can be constituted by a plurality of image analysis devices and a content attention evaluation analysis device that intensively analyzes data transmitted from the image analysis devices. Play.

  Exemplary embodiments of an evaluation apparatus and an evaluation method for content attention changing with time according to the present invention will be described below in detail with reference to the accompanying drawings. The content attention evaluation device and the content attention evaluation system of the present invention are suitable when installed in a place where the content can be noticed by a large number of people, for example, in a station facility. In the following Example 1, an example in which the angle of the detected face direction is stored along with the passage of the time at which the angle was detected is shown. In Example 2, the detected angle of the face direction is represented as a person. An example of storing together with the change of the position of will be shown. In the following embodiments, the detected angle in the face direction is expressed by an angle display θ [deg] based on the power method.

  FIG. 1 is an explanatory diagram for explaining the concept of the content attention evaluation device and the evaluation method according to the first embodiment. As shown in the figure, in the present apparatus and method, the direction of the line of sight of the person existing in the TV screen area in front of the TV screen 100t that changes with the lapse of time T to the content displayed on the TV screen is detected. It is memorized.

According to the figure, in the TV screen area, the line-of-sight direction where the angle formed with the parallel line of the TV screen was initially θ = θ ₁ [deg] is θ = θ ₂ [deg] after a predetermined time of time T has elapsed. It has changed to. As described above, by detecting and storing θ that changes with the elapse of time T, it is possible to analyze a change of a person existing in the TV screen area according to the time lapse of gaze on the content displayed on the TV screen.

  Conventionally, as a method for evaluating and analyzing a gaze on the content, a method of statistically counting the number of persons watching the content has been common. According to this method, when the face direction is even toward the content, it is considered that the user is gazing at the content, and is counted as a person gazing at the content. However, generally speaking, there are a case where a person gives deep attention to the content and a case where light attention is given to the content. For example, deep attention is when a person is paying attention to the content with a concentration of consciousness, and light attention is when the person is only watching with distraction to the content. Is the case. In this way, it is not easy to determine how much a person's consciousness is concentrating on the content, but for example, the gaze direction of the person with respect to the content is the concentration of consciousness about the content. There seems to be some correlation with the degree. The present invention has been made paying attention to this point, and by detecting and storing the angle formed by the line-of-sight direction of the person existing in front of the TV screen and the parallel lines of the TV screen, the TV of the person is obtained. This makes it possible to analyze and evaluate the degree of attention to the content displayed on the screen.

  Next, the system configuration of the content attention evaluation device and the evaluation method shown in FIG. 1 will be described. FIG. 2 is a system configuration diagram showing an outline of the system configuration of the content attention evaluation device and the evaluation method shown in FIG. According to the figure, one wide-angle camera for photographing a passing person passing through a passage in which a TV screen 100t is arranged beside the passage, that is, a passing person existing in the passage is photographed from the TV screen 100t side. A camera 150 is provided.

  The camera area r shown in the figure is an area where a camera 150 can capture a passerby. Passersby existing in this area can be detected as passersby by being photographed by the camera 150. In addition, it is detected whether or not the face direction of the passerby detected in the camera region r faces the direction of the advertising billboard 100t. This is because when the camera 150 is disposed in the same plane of the advertising billboard 100t or in the vicinity thereof, when the face direction is directed to the advertising billboard in order to direct the line of sight, the face facing the advertising billboard direction is caused by the camera 150. By being photographed.

  The camera 150 and the system server 200 are connected via a cable C. At least one system server 200 is connected to the aggregation center server 300 via the network N. The system server 200 and the aggregation center server 300 are installed on different floors, but may be installed on the same floor or may be mounted on the same computer system. The system server 200 is an analysis device for analyzing an image taken by the camera 150 and detecting a passerby and a face direction of the passerby from the image. The camera 150 captures an image including a passerby and transmits the image to the system server 200. The system server 200 analyzes the image, detects objective attribute data of a passerby included in the image, and transmits the attribute data to the aggregation center server 300. The aggregation center server 300 aggregates the attribute data and calculates various analysis indexes.

  Next, the configuration of the system server 200 shown in FIG. 2 will be described. FIG. 3 is a functional block diagram showing the configuration of the system server 200 shown in FIG. As shown in the figure, the system server 200 includes an operation to which a main control unit 201, a passer-by person detection unit 202, a gaze person detection processing unit 203, a storage unit 204, and an operation unit 206 such as a keyboard are connected. The control unit 205 includes a display control unit 207 to which a display unit 208 such as a display device is connected, and an I / F unit 209 that controls data exchange with the outside. The main control unit 201 is a control unit that controls processing of the entire system server 200. Specifically, the image data is received from the camera 150 via the I / F unit 209 and transferred to the passer-by detection unit 202 or the gaze detection unit 203, and the passer-by detection unit 202 or the gaze detection process The information output from the unit 203 is transferred to the storage unit 204 or transmitted to the aggregation center server 300 via the I / F unit 209, or the operation unit 206 receives a predetermined input signal from the reception operation control unit 205. The processing described above is executed in response to the input of the converted predetermined operation signal, and processing for displaying the processing result of each processing unit on the display unit 208 via the display control unit 207 is executed.

  The passer-by person detection unit 202a determines a moving region detection processing unit 202a that detects a moving region from the frame difference of an image captured by the camera 150, and determines whether or not the moving region is a rectangle that approximates a person. And a person detection processing unit 202b for detecting. The gaze person detection processing unit 203 extracts a rectangle assumed to correspond to the face portion from a rectangular moving region that approximates the person detected by the person detection processing unit 202b, and determines whether or not the rectangle is a face. A face detection processing unit 203a for detecting a face and a face angle detection processing unit 203b for detecting a face direction to which the face detected by the face detection processing unit 203a is facing. The person detection processing unit 202b and the face detection processing unit 203a set a flag on the image data of the detected person and prevent detection of the person with the flag in order to prevent the same person from being detected twice. Perform flag control. According to this flag control, the newly detected person is not flagged and the already detected person is flagged, so the newly detected person and the already detected person are Can be distinguished.

  The storage unit 204 is a non-volatile magnetic recording device, and stores a determination criterion for a face direction angle (face angle detection reference) determined by the face angle detection processing unit 203b as a face direction toward an advertising billboard. It has a gaze determination criterion 204a. Specifically, the aspect ratio of the face area is determined, the positional relationship between the camera 150 that captured the image and the TV screen 100t and the aspect ratio of the face included in the captured image are determined, and the face direction is determined. It is determined whether or not the angle is a face direction angle that can be regarded as facing the TV screen 100t. For example, in the case of a face photographed by the camera 150, since the camera 150 and the TV screen 100t are in a positional relationship facing the same direction, the closer the aspect ratio is to the ratio of a human face facing the front, the closer the face is to the TV. It is likely that you are watching the screen. When the face direction is an angle of the face direction that can be regarded as a gaze on the TV screen, the angle of the face direction becomes a detection target of the face angle detection processing unit 203b.

  Next, the configuration of the aggregation center server 300 shown in FIG. 2 will be described. FIG. 4 is a functional block diagram showing the configuration of the aggregation center server 300 shown in FIG. As shown in the figure, the aggregation center server 300 includes a main control unit 301, a passage detection time registration unit 302, a face angle registration processing unit 303, a gaze depth rate calculation processing unit 304, and an empty gaze rate calculation process. Control unit 305, storage unit 306, operation control unit 307 to which an operation unit 308 such as a keyboard is connected, display control unit 309 to which a display unit 310 such as a display device is connected, and external data transfer I / F unit 311. The main control unit 301 is a control unit that controls processing of the entire aggregation center server 300. Specifically, the main control unit 301 receives attribute data such as passage detection and a face direction angle from the system server 200 via the I / F unit 311, and passes to the passage detection time registration unit 302 or the face angle registration processing unit 303. Perform the transfer process. When the attribute data is exchanged, the passage detection time registration unit 302 registers the detection time of the person included in the image photographed by the camera 150 as a passerby for each person, and the face angle registration processing unit 303 The face direction angle (face angle) of the person included in the image as a person whose face direction angle is on the TV screen is registered in units of the person. Also, the gaze depth rate calculation processing unit 304 sums up the face angles that change with time for each person, and the sum is calculated based on the time at which the face angle is detected and the face angle (gaze angle threshold) that can be regarded as gaze. Calculate the gaze depth ratio divided. In addition, the empty gaze rate calculation processing unit 305, when the time when the face angle is equal to or greater than the gaze angle threshold, is adjacent to the time when the face angle temporarily falls below the gaze angle threshold, for each person. The ratio of the time when the face angle temporarily falls below the gaze angle threshold with respect to the total is calculated. This ratio is called the empty gaze rate, and when this value is below a predetermined value, it is considered that the TV screen is substantially watched even during the time when the face angle temporarily falls below the gaze angle threshold. . In this case, the time when the face angle temporarily falls below the gaze angle threshold is referred to as the empty gaze time. Further, the main control unit 301 performs a process of transferring information output from the passage detection time registration 302 or the face angle registration processing unit 303 to the storage unit 306, or the operation unit 308 receives a predetermined input signal by the operation control unit 307. The processing described above is executed in response to the input of the converted predetermined operation signal, and processing for displaying the processing result of each processing on the display unit 310 via the display control unit 309 is executed.

  The passage detection time registration unit 302 instructs to register the time when a passerby is detected based on the attribute data transmitted from the system server 200. This registration instruction is transferred to the storage unit 306. The face angle registration processing unit 303 instructs to register the face angle from the attribute data transmitted from the system server 200. This registration instruction is transferred to the storage unit 306. The passage detection time registration unit 302 and the face angle registration processing unit 303 issue a registration instruction for the number of people detection time and the face angle every 100 milliseconds. That is, it instructs to register the face angle of a person included in an image frame of 100 milliseconds along with the time. Receiving the registration instruction, the storage unit 306 registers the face angle of the person along with the time in the DB 306a.

  The gaze depth rate calculation processing unit 304 adds the face angles registered in the DB 306a in units of 100 milliseconds to the person unit from which the face angle is detected, and (total face angle) / (face angle detected). Time) ÷ (gazing angle threshold) is calculated to calculate the gaze depth rate. Note that the time when the face angle is detected is the time from the time when the face angle is equal to or greater than the face angle detection reference and the detection is started as the detection target to the time when the detection is no longer detected and the detection ends. Also, the gaze angle threshold is a minimum face angle at which the face angle can be regarded as gazing at the content displayed on the target TV screen. In the first embodiment, the total face angle is the sum of the discrete values when the face angle is a discrete value related to the time T, but the face angle can be expressed by a continuous function related to the time T. Is an integrated value obtained by integrating the continuous function in the interval of the time when the face angle is detected.

  Note that the gaze depth rate calculated by the above formula is as follows. That is, gaze depth rate = (total face angle) / (time when face angle is detected) / (gaze angle threshold) = (total face angle) / {(time when face angle is detected) × (gaze angle) Threshold)}, but {(time when face angle is detected) × (gaze angle threshold)} is the total value of face angles that can be considered to have been gaze on average over the time when the face angle was detected. It is the minimum limit. That is, if (total face angle) is equal to or greater than {(time when face angle is detected) × (gaze angle threshold)}, it is considered that content is being watched on average in the time when face angle is detected. Therefore, if (total of face angles) is less than {(time when face angles are detected) × (gaze angle threshold)}, it can be considered that content has not been watched on average. In other words, if the gaze depth rate is 1 or more, it can be considered that the content is watched on average, and if it is smaller than 1, it can be considered that the content is not watched on average.

  The storage unit 306 is a non-volatile magnetic recording device and includes a DB 306a that registers the face angle of a passerby every 100 milliseconds. One record is generated every 100 milliseconds, and the face angle of a passerby photographed in a frame of an image included in 100 milliseconds corresponding to the one record and the time at that time are registered.

  Next, a face angle registration table stored in the DB 306a of the storage unit 306 of the aggregation center server 300 illustrated in FIG. 4 will be described. FIG. 5 is a table image diagram of a face angle registration table stored in the DB 306a of the storage unit 306 of the aggregation center server 300 shown in FIG. As shown in the figure, the database table has columns of “person ID”, “date”, “time (T)”, and “face angle (Z)”. Records are generated for each person photographed by the camera 150 and regarded as the same person. For example, a record whose “person ID” is “1” includes a record of a person whose face angle of 10 degrees is detected at 16: 55: 20: 100 ms on December 16, 2005. After this time, face angles of 0 to 90 degrees are detected and registered every 100 milliseconds. That is, from the start to the end of the presence and face angle detection of the person in the TV screen area, the face angle is detected and registered together with the detection time every 100 milliseconds.

  By configuring the database table in this way, storing information on “date”, “time”, and “face angle” for each person, and analyzing the stored data, the following effects can be expected. That is, it is possible to statistically grasp the influence of the surrounding environment such as the ambient illuminance and the TV screen area that changes with the “date and time” on the gaze on the TV screen. In addition, it is possible to grasp the influence of external factors such as temperature, weather, and social conditions (reporting, conflicts, etc.) that can be specified by “date and time” on the TV screen. In addition, it is possible to grasp the influence of the appearance and design of the content itself displayed on the TV screen, or the planning and configuration of the content on the watching on the TV screen. With such gaze evaluation, it is possible to establish a method for presenting content that further encourages attention to the TV screen and to verify the effect.

  Next, a processing procedure in the content attention evaluation device and the evaluation method shown in FIGS. 1 and 2 will be described. FIG. 6 is a flowchart showing a processing procedure in the content attention evaluation device and the evaluation method shown in FIGS. As shown in the figure, the system server 200 first detects a moving area (step S101). Subsequently, it is determined whether or not the detected moving area is a person (step S102). When it is determined that the detected moving area is a person (Yes in step S102), face area detection processing is executed (step S103). Subsequently, it is determined whether or not the face area is detected (step S104). When it is determined that the face area is detected (Yes in step S104), face angle detection processing is executed (step S105). If the moving area is not determined to be a person (No at Step S102) or if it is not determined that a face area is detected (No at Step S104), the process proceeds to Step S107. Subsequent to step S105, information indicating the detected face angle is transmitted to the aggregation center server 300 (step S106). Subsequently, it is determined whether or not tracking is completed (step S107). When it is determined that tracking is completed (step S107 affirmative), the processing in the system server 200 is terminated, and when it is not determined that tracking is completed (No in step S107). ) The process is moved to step S101.

  On the other hand, the aggregation center server 300 first performs a timing start process (step S201) that starts counting a predetermined time prior to all the processes, and when face angle information is transmitted from the system server 200 by the process of step S106. Then, detection time / face angle registration processing is executed (step S202). Subsequently, it is determined whether or not a predetermined time (for example, 100 milliseconds) has elapsed (step S203). If it is determined that the predetermined time has elapsed (Yes in step S203), the process proceeds to step S204, and it is determined that the predetermined time has elapsed. If not (No at Step S203), the process proceeds to Step S202. In step S204, it is determined whether or not the detection time / face angle registration process is finished. If it is determined that the detection time / face angle registration process is finished (Yes in step S204), the process in the aggregation center server 300 is finished and detected. If it is not determined that the time / face angle registration process has ended (No at step S204), the process proceeds to step S201.

Next, an outline of face angle detection by the content attention evaluation device and the evaluation method shown in FIGS. 1 and 2 will be described. FIG. 7 is an explanatory diagram in which the outline of the face angle detection by the content attention evaluation device and the evaluation method shown in FIGS. Here, it is assumed that one frame of an image taken by the camera 150 corresponds to 100 milliseconds. FIG. 7 shows the correlation between the face angle detection time (T [sec]) and the face angle (Z [deg]) with respect to the face angle at which the content displayed on the TV screen 100t is watched. According to the figure, the face angle is 0 degree in the frame 1 (first 100 milliseconds), and the face angle Z ₁ in the frame 2. Then, the face angle Z ₂ in the frame 3-6. Note that since the face angle becomes Z ₂ and becomes the gaze angle threshold value Z ₀ or more for the first time, it can be considered that the content displayed on the TV screen 100t is being watched. The gaze angle threshold value Z ₀ is the minimum value of the face angle that can be regarded as content gaze. In frames 7 to 10, the face angle is Z ₃ , and Z ₃ is the maximum value in the graph, so that it can be considered that the content is most watched. The face angle in the frame 11 Z _4, and the face angle in the frame 12 is in the Z _5. Since Z ₅ is below the gaze angle threshold value Z ₀ , it can be considered that the gaze on the content is temporarily interrupted in the frame 12. However, the face angle is again Z ₃ in the frame 13, and it can be considered that the gaze on the content has been resumed. Then the frame 14, the face angle is transition to Z _6.

In the graph of FIG. 7, all face angles are continuously equal to or greater than the gaze angle threshold value Z ₀ for a seconds from frame 3 to frame 11, and can be regarded as one continuous gaze period. . The face angle Z ₅ is below the gaze angle threshold value Z ₀ in d seconds of the frame 12, but the face angle Z ₃ is again greater than or equal to the gaze angle threshold value Z _{0 in} b seconds of the frame 13. Here, c = d / (a + b + d) is referred to as “empty gaze rate”. Here, “d” is referred to as empty gaze time, which is a time when the content is conscious but the line of sight is not in the content. That is, the “empty gazing rate” indicates a ratio of time that can be considered that the content is temporarily being watched, but is continuously watched. If the gaze angle threshold value Z ₀ is not reached, continuous gaze and face angles are not considered, but if the above gaze depth is taken into account, it can be regarded as continuous gaze even in the empty gaze time.

  If the content attention evaluation can be performed as in the first embodiment, it is possible to analyze which of the contents displayed on the TV screen that changes with time has collected higher attention. For example, when displaying a specific content, it is possible to determine how the gaze on the content varies depending on the time zone. In addition, it is possible to analyze the degree of gaze on the CM that flows between TV broadcast programs. From these facts, according to the present invention, it is possible to analyze how the content displayed on the TV screen can be collected in terms of time so that a higher gaze can be collected.

  In the first embodiment, the time transition of gaze on the TV screen of a person existing in front of the TV screen is detected and stored. However, the present invention is not limited to this, and is installed in a vehicle of public transportation such as a train. It can also be applied to a moving image display device. Regardless of whether it is a TV screen or a moving image display device, if a continuous gaze effect can be expected under consciousness even if the face angle temporarily falls below the gaze angle threshold due to a mobile phone or other factors, the gaze It is possible to quantitatively grasp the effect, and it is possible to present an effective placement method to the content creator of the TV screen or the moving image display device. In addition, it is possible to calculate an integrated value of the attention rate under the consciousness in consideration of the empty gaze time. In this way, content that is displayed on the TV screen by introducing the concept of “empty gaze” that indicates a state in which the gaze angle threshold of the face angle that is gaze judgment is not exceeded but that the gaze is kept under consciousness is introduced. It is possible to perform a gaze evaluation on a more realistic form.

  FIG. 8 is an explanatory diagram for explaining the concept of the content attention evaluation device and the evaluation method according to the second embodiment. As shown in the figure, in the present apparatus and method, a person existing in the advertising billboard watching area in front of the advertising billboard 100a that changes as the moving distance (person position) L in front of the advertising billboard changes to the advertising billboard. It is characterized by detecting and memorizing the line-of-sight direction. The person position L is a coordinate representing the position of the person with respect to the advertisement signboard. In the first embodiment, the origin is a fixed point at or near the left end of the advertisement signboard, and represents a distance [m] from the origin. Coordinates.

According to the figure, in the advertising sign watching area, the direction of the line of sight that was initially θ = θ ₁ [deg] with the parallel line of the advertising sign is θ = θ ₂ [deg] after the person moves. It has changed. As described above, when θ that changes every time the person position L elapses is detected and stored, it is possible to analyze a change in accordance with the moving distance of the person present in the advertisement signboard gaze area to the advertisement signboard.

  Conventionally, as a method for evaluating and analyzing a gaze on the content, a method of statistically counting the number of persons watching the content has been common. According to this method, when the face direction is even toward the content, it is considered that the user is gazing at the content, and is counted as a person gazing at the content. However, generally speaking, there are a case where a person gives deep attention to the content and a case where light attention is given to the content. For example, deep attention is when a person is paying attention to the content with a concentration of consciousness, and light attention is when the person is only watching with distraction to the content. Is the case. In particular, when the content is installed on the assumption that it can be viewed while moving, it is considered that the degree of attention can also change with the movement. In this way, it is not easy to determine how much a person's consciousness is directed toward the content as the person moves, but for example, the gaze direction of the person with respect to the content is conscious of the content. It seems that there is some correlation with the degree of concentration. The present invention has been made paying attention to this point, and by detecting and storing the angle formed by the line-of-sight direction of the person moving in front of the advertising billboard and the parallel line of the advertising billboard, the movement of the person It is possible to analyze and evaluate the degree of attention to advertising billboards, which changes with time.

  Next, the system configuration of the content attention evaluation device and the evaluation method shown in FIG. 8 will be described. FIG. 9 is an explanatory diagram for explaining a system configuration of the content attention evaluation device and the content attention evaluation system according to the first embodiment. According to the figure, three wide-angle cameras for photographing a passing person passing through a passage where an advertising billboard 100a is arranged beside the passage are arranged at predetermined positions in the passage. That is, the camera 150a that captures a passerby present in the passage from the advertisement signboard 100a side, and the camera 150b that captures a passerby present in the passage from the position on the passage where the advertisement signboard 100a is located on the left side. On the right side, there are cameras 150c that shoot passersby in the passage from the placement position on the passage where the advertising billboard 100a is located.

  The camera area r1 shown in the figure is an area where a camera person can be photographed by the camera 150b and the camera 150c. A passerby existing in this area can be detected as a passerby by being photographed by the camera 150b or the camera 150c. Further, it is detected whether or not the face direction of the passerby detected in the camera region r1 is directed to the advertising billboard 100a. Similarly, the camera area r2 is an area where a passerby can be photographed by the camera 150a. In addition to detecting the presence of a passerby in the camera area r2 by photographing with the camera 150a, it is detected whether or not the face of the passerby faces the direction of the advertising billboard 100a. This is because when the camera 150a is arranged in the same plane of the advertising billboard 100a or in the vicinity thereof, when the face direction is directed to the advertising billboard in order to turn the line of sight, the face facing the advertising billboard direction is caused by the camera 150a. By being photographed.

  The cameras 150a, 150b, 150c and the system server 200 are connected via a cable C. At least one system server 200 is connected to the aggregation center server 300 via the network N. The system server 200 and the aggregation center server 300 are installed on different floors, but may be installed on the same floor or may be mounted on the same computer system. The system server 200 is an analysis device that analyzes an image captured by the camera 150a, 150b, or 150c and detects a passerby and a face direction of the passerby from the image. The cameras 150a, 150b, and 150c capture an image including a passerby and transmit the image to the system server 200. The system server 200 analyzes the image, detects objective attribute data of a passerby included in the image, and transmits the attribute data to the aggregation center server 300. The aggregation center server 300 aggregates the attribute data and calculates various analysis indexes. That is, the image information that can identify an individual is closed between the system server 200 and the cameras 150a, 150b, and 150c, and is not transmitted or acquired other than these. Therefore, the advertisement attention evaluation system according to the first embodiment is a system that can prevent leakage and diffusion of personal information only by intensively protecting personal information in the system server 200 and the cameras 150a, 150b, and 150c.

  Next, the configuration of the system server 200 shown in FIG. 9 will be described. FIG. 10 is a functional block diagram showing the configuration of the system server 200 shown in FIG. As shown in the figure, the system server 200 includes an operation to which a main control unit 201, a passer-by person detection unit 202, a gaze person detection processing unit 203, a storage unit 204, and an operation unit 206 such as a keyboard are connected. The control unit 205 includes a display control unit 207 to which a display unit 208 such as a display device is connected, and an I / F unit 209 that controls data exchange with the outside. The main control unit 201 is a control unit that controls processing of the entire system server 200. Specifically, the image data is received from the camera 150 via the I / F unit 209 and transferred to the passer-by detection unit 202 or the gaze detection unit 203, and the passer-by detection unit 202 or the gaze detection process The information output from the unit 203 is transferred to the storage unit 204 or transmitted to the aggregation center server 300 via the I / F unit 209, or the operation unit 206 receives a predetermined input signal from the reception operation control unit 205. The processing described above is executed in response to the input of the converted predetermined operation signal, and processing for displaying the processing result of each processing unit on the display unit 208 via the display control unit 207 is executed.

  The passer-by person detection unit 202a determines a moving region detection processing unit 202a that detects a moving region from the frame difference of an image captured by the camera 150, and determines whether or not the moving region is a rectangle that approximates a person. A person detection processing unit 202b for detecting the position and a person position detection processing unit 202c for detecting the position of the person with respect to the advertising billboard 100a. The gaze person detection processing unit 203 extracts a rectangle assumed to correspond to the face portion from a rectangular moving region that approximates the person detected by the person detection processing unit 202b, and determines whether or not the rectangle is a face. A face detection processing unit 203a for detecting a face and a face angle detection processing unit 203b for detecting a face direction to which the face detected by the face detection processing unit 203a is facing. The person detection processing unit 202b and the face detection processing unit 203a set a flag on the image data of the detected person and prevent detection of the person with the flag in order to prevent the same person from being detected twice. Perform flag control. According to this flag control, the newly detected person is not flagged and the already detected person is flagged, so the newly detected person and the already detected person are Can be distinguished.

  The storage unit 204 is a non-volatile magnetic recording device, and stores a determination criterion for a face direction angle (face angle detection reference) determined by the face angle detection processing unit 203b as a face direction toward an advertising billboard. It has a gaze determination criterion 204a. Specifically, the aspect ratio of the face area is determined, and the positional relationship between the camera 150a, 150b or 150c that captured the image and the advertising billboard 100a and the aspect ratio of the face included in the captured image are determined. Then, it is determined whether or not the face direction is an angle of the face direction that can be regarded as facing the advertising billboard 100a. For example, in the case of a face photographed by the camera 150a, since the camera 150a and the advertising billboard 100a are in a positional relationship facing the same direction, the face becomes closer to the proportion of a human face facing the front as the face becomes closer. There is a high possibility of watching. Then, when the face direction is an angle of the face direction that can be regarded as gazing at the advertising billboard, the angle of the face direction becomes a detection target of the face angle detection processing unit 203b.

  Next, the configuration of the aggregation center server 300 shown in FIG. 9 will be described. FIG. 11 is a functional block diagram showing the configuration of the aggregation center server 300 shown in FIG. As shown in the figure, the aggregation center server 300 includes a main control unit 301, a passage detection time registration unit 302, a face angle registration processing unit 303, a person position registration processing unit 312, and a gaze depth rate calculation processing unit. 304, an empty gaze rate calculation processing unit 305, a storage unit 306, an operation control unit 307 to which an operation unit 308 such as a keyboard is connected, a display control unit 309 to which a display unit 310 such as a display device is connected, And an I / F unit 311 for controlling data exchange with the outside. The main control unit 301 is a control unit that controls processing of the entire aggregation center server 300. Specifically, the main control unit 301 receives attribute data such as passage detection, position information of the person, and angle of face direction from the system server 200 via the I / F unit 311, and receives the passage detection time registration unit 302, person position Processing to deliver to the registration processing unit 312 or the face angle registration processing unit 303 is performed. When the attribute data is exchanged, the passage detection time registration unit 302 newly adds a record corresponding to a person included as a passerby in the image taken by the camera 150a, 150b, or 150c to the DB 306a, and also registers the person position. The processing unit 312 performs processing for registering the position of the person with respect to the advertising billboard for each person, and the face angle registration processing unit 303 includes the face of the person included in the image as a person whose face direction angle is facing the advertising billboard. The direction angle (face angle) is registered in the person unit. Also, the gaze depth rate calculation processing unit 304 sums up the face angles that change with time for each person, and the sum is calculated based on the time at which the face angle is detected and the face angle (gaze angle threshold) that can be regarded as gaze. Calculate the gaze depth ratio divided. In addition, the empty gaze rate calculation processing unit 305, when the section where the face angle is equal to or greater than the gaze angle threshold is adjacent to the section where the face angle temporarily falls below the gaze angle threshold, in the unit of person, A ratio of a section in which the face angle temporarily falls below the gaze angle threshold with respect to the total is calculated. This ratio is referred to as the empty gaze rate, and when this value is less than a predetermined value, it is considered that the advertisement signboard is substantially watched even in a section where the face angle temporarily falls below the gaze angle threshold. . In this case, a section in which the face angle temporarily falls below the gaze angle threshold is referred to as an empty gaze section. Further, the main control unit 301 performs processing for transferring information output from the passage detection time registration 302, the face angle registration processing unit 303, or the person position registration processing unit 312 to the storage unit 306, or the operation unit 308 performs reception operation control. The above-described processes are executed in response to the input of the predetermined operation signal converted into the predetermined input signal by the unit 307, and the process of displaying the processing results of the respective processes on the display unit 310 via the display control unit 309 is executed. To do.

  The passage detection time registration unit 302 instructs the DB 306a to perform a new record addition process when a passerby is detected based on the attribute data transmitted from the system server 200. This additional processing instruction is transferred to the storage unit 306. The face angle registration processing unit 303 instructs to register the face angle from the attribute data transmitted from the system server 200. This registration instruction is transferred to the storage unit 306. The person position registration processing unit 312 instructs to register the position of the person with respect to the advertisement signboard from the attribute data transmitted from the system server 200. This registration instruction is transferred to the storage unit 306. Note that the face angle registration processing unit 303 and the person position registration processing unit 312 issue a registration instruction for the face angle and the person position every 100 milliseconds. That is, the user instructs registration of the face angle of the person included in the image frame of 100 milliseconds together with the person position. Receiving the registration instruction, the storage unit 306 registers the face angle of the person together with the person position in the DB 306a.

  The passage detection time registration processing unit 302 according to the second embodiment performs only new record addition processing in the DB 306a and omits passage detection time registration. However, the present invention is not limited to this, and a new record in the DB 306a is performed. The passage detection time registration process may be performed during the additional process.

  The gaze depth rate calculation processing unit 304 adds the face angles registered in the DB 306a in units of 100 milliseconds to the person unit from which the face angle is detected, and (total face angle) / (face angle detected). The gaze depth ratio is calculated by calculating (positional difference) / (gazing angle threshold). The difference in the position at which the face angle is detected is the distance from the position at which the face angle is equal to or greater than the face angle detection reference and the detection start as a detection target to the position at which the detection is no longer the detection target. . The gaze angle threshold is a minimum face angle at which the face angle can be regarded as gazing at the target advertising billboard. In the second embodiment, the total face angle is the sum of the discrete values when the face angle is a discrete value related to the person position L, but the face angle can be expressed by a continuous function related to the person position L. In this case, it is assumed that the continuous function is an integrated value obtained by integrating the continuous function in the section of the person position where the face angle is detected.

  Note that the gaze depth rate calculated by the above formula is as follows. That is, gaze depth ratio = (total face angle) / (difference in position at which face angle is detected) / (gaze angle threshold) = (total face angle) / {(difference in position at which face angle is detected) X (gazing angle threshold)}, but {(difference in position where the face angle was detected) × (gazing angle threshold)} was averaged over the difference in the position where the face angle was detected. This is the minimum degree of the total face angle that can be considered. That is, if (the sum of face angles) is equal to or more than {(difference of position at which face angle is detected) × (gaze angle threshold)}, the content is watched on average in the difference in position at which face angle is detected. If (total face angle) is less than {(difference in position where face angle is detected) x (gazing angle threshold)}, the content is not watched on average. It can be considered. In other words, if the gaze depth rate is 1 or more, it can be considered that the content is watched on average, and if it is smaller than 1, it can be considered that the content is not watched on average.

  The storage unit 306 is a non-volatile magnetic recording device and includes a DB 306a that registers the face angle of a passerby every 100 milliseconds. One record is generated every 100 milliseconds, and the face angle of the passerby and the person position at that time are registered in an image frame included in 100 milliseconds corresponding to the one record.

  Next, a face angle registration table stored in the DB 306a of the storage unit 306 of the aggregation center server 300 illustrated in FIG. 11 will be described. FIG. 12 is a table image diagram of a face angle registration table stored in the DB 306a of the storage unit 306 of the aggregation center server 300 shown in FIG. As shown in the figure, the database table has columns of “person ID”, “person position (L)”, and “face angle (Z)”. A record is generated for each person photographed by the camera 150a, 150b or 150c and regarded as the same person. For example, a record whose “person ID” is “1” includes a record of a person whose face angle of 10 degrees is detected when the person is “1” [m]. After this time, face angles of 0 to 90 degrees are detected and registered every 100 milliseconds. That is, from the start to the end of the presence and face angle detection of the person in the advertising signboard gaze area, the face angle is detected and registered every 100 milliseconds.

  By configuring the database table in this way, storing information on “person position” and “face angle” for each person, and analyzing the stored data, the following effects can be expected. That is, it is possible to statistically grasp the influence of the information provided by the advertising billboard that differs depending on the “person position” on the attention to the advertising billboard. In addition, it is possible to grasp the influence of the appearance and design of the advertisement signboard itself or the planning / configuration of the advertisement signboard on the attention to the advertisement signboard. By such gaze evaluation, it is possible to establish a method of presenting an advertisement that further encourages attention to the advertising billboard, and to prove the effect.

  Note that when the passage detection time registration processing unit 302 performs the passage detection time registration processing at the time of new record addition processing to the DB 306a, the database table shown in FIG. "Time") column will be added. In this way, it is possible to grasp when the person detected and registered in the DB 306a is gazing at what position and at what depth the person is gazing at the advertisement signboard.

  Next, processing procedures in the content attention evaluation device and the evaluation method shown in FIGS. 8 and 9 will be described. FIG. 13 is a flowchart showing a processing procedure in the content attention evaluation device and the evaluation method shown in FIGS. As shown in the figure, the system server 200 first detects a moving area (step S111). Subsequently, it is determined whether or not the detected moving area is a person (step S112). If it is determined that the detected moving area is a person (Yes in step S112), face area detection processing is executed (step S113). Subsequently, it is determined whether or not the face area is detected (step S114). When it is determined that the face area is detected (Yes in step S114), face angle detection processing is executed (step S115). If the moving area is not determined to be a person (No at Step S112) and if it is not determined that a face area has been detected (No at Step S114), the process proceeds to Step S117. Subsequent to step S115, the person position (L) is detected (step S116), and information indicating the detected face angle is transmitted to the aggregation center server 300 (step S117). Subsequently, it is determined whether or not the tracking is finished (step S118). If it is determined that the tracking is finished (step S118 affirmative), the processing in the system server 200 is finished, and if the tracking is not judged to be finished (No in step S118). ) The process is moved to step S111.

  On the other hand, the aggregation center server 300 first performs a timing start process (step S211) that starts timing for a predetermined time prior to all the processes, and when face angle information is transmitted from the system server 200 by the process of step S117. Then, a person position / face angle registration process is executed (step S212). Subsequently, it is determined whether or not a predetermined time (for example, 100 milliseconds) has elapsed (step S213). If it is determined that the predetermined time has elapsed (Yes in step S213), the process proceeds to step S214, and it is determined that the predetermined time has elapsed. If not (No at Step S213), the process proceeds to Step S212. In step S214, it is determined whether or not the detection time / face angle registration process is completed. If it is determined that the detection time / face angle registration process is completed (Yes in step S214), the process in the aggregation center server 300 is terminated and detected. If it is not determined that the time / face angle registration process has ended (No at step S214), the process proceeds to step S201.

Next, an outline of face angle detection by the content attention evaluation device and the evaluation method shown in FIGS. 8 and 9 will be described. FIG. 14 is an explanatory diagram in which the outline of the face angle detection by the content attention evaluation device and the evaluation method shown in FIGS. Here, it is assumed that one frame of an image taken by the camera 150a, 150b, or 150c corresponds to 100 milliseconds. FIG. 14 shows the correlation between the person position (L [m]) and the face angle (Z [deg]) with respect to the face angle at which the content displayed on the advertising billboard 100a is watched. According to the figure, the face angle is 0 degree in the frame 1 (first 100 milliseconds), and the face angle Z ₁ in the frame 2. Then, the face angle Z ₂ in the frame 3-6. Note that since the face angle becomes Z ₂ and becomes the gaze angle threshold value Z ₀ or more for the first time, it can be considered that the advertising signboard 100a is being watched. The gaze angle threshold value Z ₀ is the minimum value of the face angle that can be regarded as an advertisement signboard gaze. In frames 7 to 10, the face angle is Z ₃ , and Z ₃ is the maximum value in the graph, so that it can be considered that the advertising signboard is most watched. The face angle in the frame 11 Z _4, and the face angle in the frame 12 is in the Z _5. Since Z ₅ is below the gaze angle threshold value Z ₀ , it can be considered that the gaze on the advertising signboard is temporarily interrupted in the frame 12. However, face angle in frame 13 has become a Z ₃ again, watch for the billboards can be considered to have resumed. Then the frame 14, the face angle is transition to Z _6.

In the graph of FIG. 14, a [m] from frame 3 to frame 11 has all face angles equal to or greater than the gaze angle threshold value Z ₀ , and should be regarded as one continuous gaze interval. Can do. The face angle Z ₅ is lower than the gaze angle threshold value Z ₀ at d [m] of the frame 12, but the face angle Z ₃ is again greater than or equal to the gaze angle threshold value Z ₀ at b [m] of the frame 13. . Here, c = d / (a + b + d) is referred to as “empty gaze rate”. Here, “d” is referred to as an empty gaze section, and refers to a section where there is a consciousness about the advertisement signboard but the line of sight is not on the advertisement signboard. That is, the “empty gazing rate” indicates a ratio of a section that can be regarded as being continuously watched, although the watch on the advertisement signboard is temporarily interrupted. Although continuous gaze and face angle and below the gaze angle threshold Z ₀ is not considered, when considering the above gaze depth, so that the can be regarded as gaze that even continuous in empty gaze interval.

  When content attention can be evaluated as in the second embodiment, it is possible to analyze which information gathered a higher gaze among the information displayed on different advertising signs depending on the person position. For example, when specific information is displayed, it can be determined how the gaze on the information differs depending on the person position. Therefore, according to the present invention, it is possible to analyze at which position the information to be displayed on the advertisement signboard can be collected to obtain a higher gaze. Further, when the advertising effect is remarkable, it can be determined that the information estimated to have attracted a high gaze has contributed to the advertising effect.

  When the advertising billboard of Example 2 is in the vicinity of an automatic ticket gate in a station, the face angle may temporarily fall below the gaze angle threshold when using the automatic ticket gate, but a continuous gaze effect can be expected under consciousness There is. According to the present invention, the gaze effect can be grasped quantitatively, and an effective placement method can be presented to the advertisement publisher. In addition, it is possible to calculate the integrated value of the attention rate under consideration, taking into account the empty gaze distance. In this way, by introducing the concept of `` empty gaze '' that indicates that the gaze angle threshold of the face angle that is the gaze determination does not exceed the consciousness but continues to gaze, the evaluation of gaze on the advertising signboard can be evaluated It can be done in a more realistic way.

  The first and second embodiments of the present invention have been described above. However, the present invention is not limited to these embodiments, and various different implementations can be made within the scope of the technical idea described in the claims. It may be implemented by way of example.

  INDUSTRIAL APPLICABILITY The present invention is effective in evaluating attention to contents such as advertisements and TV broadcast programs by a more objective method. In particular, the present invention can be efficiently designed to attract attention to advertisement media and TV content media. This is useful when you want to investigate the installation method.

It is explanatory drawing for demonstrating the concept of the content attention evaluation apparatus and evaluation method which concern on Example 1. FIG. FIG. 2 is a system configuration diagram illustrating an outline of a system configuration of a content attention evaluation device and an evaluation method illustrated in FIG. 1. It is a functional block diagram which shows the structure of the system server 200 shown in FIG. FIG. 3 is a functional block diagram illustrating a configuration of a totaling center server 300 illustrated in FIG. 2. FIG. 5 is a table image diagram of a face angle registration table stored in a DB 306a of the storage unit 306 of the aggregation center server 300 illustrated in FIG. It is a flowchart which shows the process sequence in the content attention evaluation apparatus and evaluation method shown in FIG. It is explanatory drawing which made the outline of the face angle detection by the content attention evaluation apparatus and evaluation method shown in FIG. It is explanatory drawing for demonstrating the concept of the content attention evaluation apparatus and evaluation method which concern on Example 2. FIG. It is explanatory drawing for demonstrating the system configuration | structure of the content attention evaluation apparatus and content attention evaluation system of Example 2. FIG. It is a functional block diagram which shows the structure of the system server 200 shown in FIG. FIG. 10 is a functional block diagram illustrating a configuration of the aggregation center server 300 illustrated in FIG. 9. It is a table image figure of the face angle registration table stored in DB306a of the memory | storage part 306 of the total center server 300 shown in FIG. 10 is a flowchart showing a processing procedure in the content attention evaluation device and the evaluation method shown in FIGS. 8 and 9. It is explanatory drawing which made the outline of the face angle detection by the content attention evaluation apparatus and evaluation method shown in FIG.

Explanation of symbols

100a advertising billboard 100t TV screen 150, 150a, 150b, 150c camera 200 system server 201 main control unit 202 passer-by person detection unit 202a moving area detection processing unit 202b person detection processing unit 202c person position detection processing unit 203 gaze person detection processing unit 203a Face detection processing unit 203b Face angle detection processing unit 204 Storage unit 300 Aggregation center server 301 Main control unit 302 Passing detection time registration unit 303 Face angle registration processing unit 304 Gaze depth rate calculation processing unit 305 Sky gaze rate calculation processing unit 306 Storage unit 306a DB
312 Person position registration processing unit

Claims (12)

A content attention evaluation device that evaluates the attention of a person existing in front of the content medium to the content based on an image captured by the imaging device of the person existing in front of the content medium,
Person detecting means for detecting a person imaged in the image;
Face detection means for detecting the face of the person detected by the person detection means;
Face angle detection means for detecting the angle of the face direction of the person whose face is detected by the face detection means;
A content attention evaluation apparatus comprising: a face angle storage unit that stores an angle of the face direction detected by the face angle detection unit as time passes.   2. The content attention evaluation apparatus according to claim 1, wherein the time is a time when the face direction angle is detected by the face angle detection unit.   The face angle storage unit stores the angle of the face direction with the passage of time by the face angle detection unit for each person detected by the person detection unit. Content attention evaluation device.   The face direction angle stored in the face angle storage means is summed up in units of persons detected by the person detection means, and the total time is a time from the start of detection of the face direction angle to the end. 4. The gaze depth rate calculating means for calculating a gaze depth rate by dividing the angle of the face direction by an angle threshold that can be regarded as an angle at which the content is gazeed. Content attention evaluation device.   The detection of the angle of the face direction is started during the time during which the angle of the face direction stored in the face angle storage means is temporarily below an angle threshold that can be regarded as an angle at which the content is watched. 5. The content attention evaluation device according to claim 2, further comprising: an empty gaze rate calculating unit that calculates an empty gaze rate divided by a time from the start to the end. Further comprising a person position detecting means for detecting the position of the person detected by the person detecting means,
2. The content attention evaluation apparatus according to claim 1, wherein the passage of time corresponds to a change in the position of a person detected by the person position detecting unit.   The face angle storage means stores, for each person detected by the person detection means, the angle of the face direction detected by the face angle detection means together with the position of the person detected by the person position detection means. The content attention evaluation device according to claim 6, wherein:   The face direction angle stored in the face angle storage means is summed up in units of persons detected by the person detection means, and the sum is detected until the face direction angle detection starts and ends. And a moving gaze depth rate calculating means for calculating a moving gaze depth rate divided by an angle threshold that can be considered that the angle of the face direction and the angle of the face direction is an angle at which the content is gazed. Item 6. The content attention evaluation device according to Item 6 or 7.   The movement distance that the person has moved while the angle of the face direction stored in the face angle storage means is temporarily below an angle threshold that can be regarded as the angle at which the content is watched is expressed as the angle of the face direction. 9. A moving air gaze rate calculating means for calculating a moving air gaze rate divided by a moving distance traveled by the person from the start to the end of detection of the image is further provided. Content attention evaluation device described in 1. The content attention evaluation device includes an image analysis device to which the imaging device is connected, and a content attention evaluation analysis device connected to be communicable with at least one of the image analysis devices,
The image analysis device transmits data that cannot identify a person included in the image to the content attention evaluation analysis device,
The content attention evaluation device according to claim 1, wherein the content attention evaluation analysis device includes the face angle storage unit that stores data in which the person cannot be specified. A content attention evaluation method for evaluating the attention of a person existing in front of a content medium to the content based on an image obtained by imaging an image of a person existing in front of the content medium,
A person detection step of detecting a person imaged in the image;
A face detection step of detecting the face of the person detected by the person detection step;
A face direction detecting step of detecting an angle of the face direction of the person whose face is detected by the face detecting step;
A content attention evaluation method comprising: a face angle storage step of storing the angle of the face direction detected by the face direction detection step together with a time of the detection. A content attention evaluation method for evaluating the attention of a person existing in front of the content medium to the content based on an image captured by the imaging method of a person existing in front of the content medium,
A person detection step of detecting a person imaged in the image;
A person position detecting step of detecting the position of the person detected by the person detecting step;
A face detection step of detecting the face of the person detected by the person detection step;
A face direction detecting step of detecting an angle of the face direction of the person whose face is detected by the face detecting step;
A content attention evaluation method comprising: a face angle storage step of storing the face direction angle detected by the face direction detection step together with the position of the person detected by the person position detection step.

Images