JP2009199610A - Merchandise interest degree measuring device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a device capable of quantitatively acquiring a shopper's degree of interest in individual merchandise, its distribution, changes with time, etc. by image processing. <P>SOLUTION: The device has: an image processor for tracking the movement of an individual on a picked-up image by an image processing technology and measuring the time of staying in an area set as within the limit of a close range of merchandise; a device for discriminating individual shoppers' degrees of interest in merchandise in several stages; and a data processor for processing data according to the degrees of interest registered in a device for storing data according to the degrees of interest. The image processor comprises a person recognizing means for recognizing the approximate contour of a person's image by a normal vector and tracking it with time. A criterion in the interest degree discriminating device has a speed reference compared with an instantaneous moving speed, and only when the speed reference is set to be not faster than a slow moving speed close to a stop state and the instantaneous moving speed corresponds to this speed reference, the degree of interest is discriminated by the long/short state of the staying time. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a device that measures the degree of interest of each shopper for an item displayed on a display product, such as a mannequin or a showcase, and subtotals it, that is, a device for measuring a product interest level. The level of interest of each shopper for each item can be quantified and aggregated.

Traditionally, it has been necessary to measure the degree to which a shopper is interested in products displayed on display shelves and showcases, and to measure the extent to which the shopper is interested. The behavior of shoppers in the vicinity of display shelves and showcases The actual situation is that the observer is directly and indirectly visually inspecting and judges the level of interest. In this judgment by the observer, the objectivity is poor, and it is difficult to standardize the evaluation results or the contents of the aggregated data. Therefore, it is difficult to share the interest level data in a plurality of product display locations. In addition, it is difficult to share the data.
If the evaluation of the interest level of each individual shopper is made objective, it is possible to share the data on the interest level at multiple product display locations and share the data. Is high and the usability is high. Therefore, it is required that the degree of interest of the shopper in the product is objectively evaluated and the evaluation result is quantified.

Japanese Patent No. 3800257 Japanese Patent No. 3406587 Japanese Patent No. 3390426 JP 2005-501348 A JP 2006-85366 A

In view of the above, it is necessary to realize a measuring device that can measure the degree of interest of a shopper for a product or the like with high reliability in a rational and objective manner.
In addition, it is effective for detailed analysis of the strength of the shopper's interest in each product to accurately grasp the quantitative measurement data and distribution of each shopper's interest. .

  Therefore, the accuracy of measurement of the degree of interest of the shopper for each product is improved, and the reliability of the aggregation result of the degree of interest is increased. It is an object of the present invention to make it possible to quantitatively grasp changes with time.

Means for solving the above problems are: (A) Based on a captured image in which an area where a product is arranged is always captured from above with one camera, the degree of interest of each shopper for the product is measured. About the merchandise interest level measuring apparatus which discriminate | determines the said measured value, registers, processes this, and outputs this, and outputs merchandise interest level information, it is based on following (a)-(l).
(A) Recognizing an individual person on the captured image, tracking the movement of the virtual centroid point of the person, and determining the virtual centroid point in the region of interest measurement set as the close range of the product Having an image processing device for measuring stay time;
(B) having an interest level determination device that determines the degree of interest of each shopper for the product in several stages based on the determination criteria based on the stay time from the image processing device;
(C) having an interest level-specific data storage device for registering interest level-specific data by the interest level determination device;
(D) having a data processing device that reads out the data by interest level registered in the data storage device by interest level and processes the data;
(E) having an output device for outputting interest level information from the data processing device;
(F) The image processing apparatus recognizes a rough outline of a person image in the captured image by a normal vector, and based on the normal vector, regards the person image as the virtual centroid point by the vector focus method. , Including human recognition means for tracking the virtual centroid point over time, and when the virtual centroid point crosses the boundary of the measurement area of interest from the outside to the inside of the measurement area, timing is started from the inside The image processing apparatus includes a stay time measuring means for stopping the timing when crossing the outside;
(G) The operation means sets the boundary of the product interest measurement area on the captured image for the image processing apparatus, and sets the interest level determination criterion for the interest level determination apparatus. That
(H) The determination criterion in the interest level determination device is a time reference to be compared with the stay time, and the determination criterion is set in multiple stages so that the time reference becomes longer as the interest level is higher.
(I) the image processing apparatus includes a speed measuring unit that measures an instantaneous moving speed of the virtual center of gravity within the region of interest measurement;
(Nu) The degree-of-interest discriminating device has a speed reference for comparison with the instantaneous moving speed as a discrimination reference,
(L) The degree of interest is determined based on the length of the staying time only when the speed reference is set to be equal to or less than the slow movement speed close to the stop state and the instantaneous movement speed is equal to the speed reference. That.

Note that the “product interest measurement area set as within the range of the product's close distance” is a range of the distance that the person approaches when the user collects information on the product, and the interest level measurement The boundary is set using operation means (such as a keyboard of the input device) according to the shape, size, etc. of the target.
“Individual shopper's interest level” means the shopper's individual interest level (strength of interest) that is presumed to have definitely shown interest in a specific product on display. The discriminant criteria for evaluation is set by the input device of the operation means in accordance with the type and size of the product.

By having the above requirements (a) and (b), only those who are interested in a specific product are subject to interest level measurement. Furthermore, by having the above requirements (f), (ri), (nu), and (le), the degree of interest is evaluated only for the shopper who has stopped moving in the product interest measurement area, and is discriminated in multiple stages. The
Then, the interest level data determined as described above is recorded by means of the above requirement (c), and the recorded interest level data is processed and processed into various interest level information by the means of (d). The interest level information is output by means of the requirement (e).
The above is the basic concept of the present invention.

According to the above-described means of the present invention, a shopper who is presumed to have shown a special interest in the displayed product is accurately distinguished and specified, and the individual shopper's interest level is used for observing the product. Based on the criteria for the length of time estimated to have been spent (time criteria), and based on a predetermined moving speed criterion (instantaneous moving speed criterion) indicating whether or not the exhibition was stopped, this evaluation result ( Multi-stage evaluation results based on the above time criteria) are recorded as measurement data for each stage of merchandise interest. Since the data classified and collected by the degree of interest in this way are objective and quantified, it is possible to quantitatively analyze and process the information on the merchandise interest level using this data. Therefore, the product interest information created by processing the data according to the degree of interest is extremely reliable.
It should be noted that the multi-stage discrimination criteria for the product interest measurement area boundary and the interest level as described above are easily set as appropriate according to the type and size of the interest level measurement target by the input unit of the input device.

According to the product interest level measurement device of the present invention, the accuracy of measurement of the individual interest level of the shopper for the specific product is enhanced, the reliability of the measured interest level according to the stage is increased, and the individual product is measured. It is possible to objectively and quantitatively grasp the degree of interest, its distribution, and changes with time.
The present invention has the above requirements (Li), (Nu), and (Lu), so that the measurement area of interest, that is, only when the instantaneous movement speed of the person represented by the virtual center of gravity is the speed reference, that is, Since the level of interest is determined and recorded according to the length of staying time within the closest distance of the product, the product interest level happens to approach the product by chance to avoid other pedestrians, etc. The shopper who has passed through the measurement area is accurately discriminated and excluded from the interest level measurement target. Therefore, the discrimination accuracy for discriminating only the shopper who is interested in a specific product is extremely high, and the accuracy of the collected interest level data is extremely high.

These are the schematic diagrams which show the picked-up image which imaged the place of the mannequin display which exhibits goods from upper direction in Embodiment 1. FIG. These are block diagrams which show the structure of the merchandise interest degree measuring apparatus by Embodiment 1. FIG. These are flowcharts which show operation | movement of the image processing apparatus of the merchandise interest degree measuring apparatus by Embodiment 1. FIG. These are flowcharts which show operation | movement of the interest degree discrimination | determination apparatus of the merchandise interest degree measuring apparatus by Embodiment 1. FIG. These are the schematic diagrams for demonstrating the circular Fourier method used in the image processing of person recognition. These are figures for demonstrating that the intensity | strength direction of an image is shown by performing a circular Fourier method, The figure (a) is a mode that performs fundamental wave Fourier transformation to the outline part of a person image in circular shape. (B) shows a waveform graph when the fundamental wave Fourier transform is executed. These are the schematic diagrams for demonstrating the vector focus method used in the image processing of a human recognition, The figure (a) shows a mode when a standard human-type image is arrange | positioned on an image, The figure (b). Shows a state when a person to be identified is shown on the image. These are the tables which show the interest degree parameter of the interest degree discrimination | determination apparatus in the merchandise interest degree measuring apparatus of Embodiment 1. FIG. These are the table | surfaces which show the interest degree memory | storage data of the data storage apparatus classified by interest degree in the merchandise interest degree measuring apparatus of Embodiment 1. FIG. These are the table | surface and graph which show the output result of the output device in the merchandise interest degree measuring apparatus of Embodiment 1. FIG.

Embodiments of the present invention will be described below with reference to the drawings.
(Embodiment 1)
In the first embodiment, as a simple application example of the merchandise interest level measuring device, a case where the merchandise interest level for the mannequin display is measured will be described as an example. And this Embodiment 1 is a thing of the range of requirements (A)-(li) of a solution means.

  FIG. 1 shows a place where a mannequin display in which clothing (product) is put on a mannequin M in a store S, such as a department store, and is displayed facing the passage T is displayed by a camera (imaging device) 9 from above. A captured image 1 being captured is shown. And the merchandise interest level measuring apparatus 2 of Embodiment 1 measures how much interest the shopper in the passage T shows with respect to the clothing of the merchandise worn on the mannequin M. .

  In FIG. 1, A is an interest measurement region set in the captured image 1, h is a person image of a shopper passing through the passage T, and G is a virtual center of gravity of the person image h. R1 is a movement trajectory showing the behavior of the person image h when the person image h stays in the interest measurement area A for a certain time (for example, 10 seconds) or more, and R2 is the person image h interested. It is a movement locus | trajectory which shows a person's behavior when not staying in the measurement area | region A more than the said fixed time.

  FIG. 2 shows a configuration diagram of the merchandise interest level measuring device 2. The merchandise interest level measuring device 2 includes a camera 9, an operating means 7, an image processing device 3, an interest level discriminating device 4, and an interest level-specific data storage device 5. , A data processing device 6 and an output device 8.

  The camera 9 is composed of, for example, a single CCD camera or the like, and is installed on the ceiling or wall at an angle at which the product and the floor near the product can be seen from above. 1 is continuously sent to the image processing apparatus 3. By installing the camera 9 at an angle at which the floor surface is reflected, it is possible to obtain information that makes it easy to grasp the course when the movement trajectory R of the person image h is displayed. When the ceiling is low, the range of the movement trajectory R of the person image h is narrowed. Therefore, it is preferable to use the camera 9 equipped with a wide-angle lens to widen the photographing range.

  The operation means 7 is composed of an input device such as a keyboard and a mouse, for example, and the interest measurement area A is set on the captured image 1 captured by the camera 9 with respect to the image processing apparatus 3 by the operation means 7. In addition, the operation means 7 sets an interest level determination criterion for the interest level determination device 4 or instructs the data processing device 6 to execute predetermined data processing and output it to the output device 8. To do.

  The interest measurement area A is an area within the range of the close distance of the product from the outer periphery of the product in real space, and its boundary is set on the captured image 1. Within the range of a product is when a person looks into the product, touches the product, or smells the product, and collects information from the product. It is within the range of the closest distance to approach. For example, when clothing is displayed on the mannequin M, it is a close distance when a specific part of the clothing is stared or touched to check the cloth, for example, the outer periphery of the mannequin M. Within the range of 50 cm to 60 cm from the part. In the case of foodstuffs, it is a close distance for visually checking the freshness of foodstuffs, or for approaching to confirm product display etc., for example, within a range of about 30 cm from the edge of the display shelf. is there. Furthermore, in the case of electrical appliances, it is a close distance to approach the electrical appliance for visual confirmation or to touch it for manipulation, for example, within a range of about 30 cm from the edge of the display shelf.

  The image processing device 3 recognizes an individual person on the captured image 1, tracks the movement of the virtual centroid G of the person, and analyzes the behavior of the virtual centroid G in the measurement area A of interest. It is. The image processing apparatus 3 includes a person recognition unit 31 and a stay time measurement unit 32.

  The human recognition means 31 executes fundamental wave Fourier transform on each pixel value on the circumference of a circle having the processing point as the center point for each processing point arranged at equal intervals in the captured image 1 and obtained. The outline of the person image is recognized by the normal vector for each processing point from the obtained phase, and the virtual center of gravity G of the person image is captured based on the normal vector, and the virtual center of gravity G is determined over time. This is a means provided with an image processing technique (Patent No. 3406587, No. 3390426) based on a so-called vector focus method for tracking.

  As the image processing of the vector focus method, first, as shown in FIG. 5, using the density difference between the object image B and the background image W in the captured image 1, they are arranged at equal intervals on the captured image 1. For each processing point P, a fundamental wave Fourier transform represented by the following formula (1) is performed on the pixel values of the pixels d1 to d12 on the circumference of the circle 11 having the processing point P as the center point ( Circular Fourier method). The fundamental wave means a sine wave and a cosine wave having a length corresponding to one cycle of the circumference centering on the processing point P.

上記数式中、ＳＶは画素のピクセル値にフーリエ変換の正弦波を積和した積和値、ＣＶは画素のピクセル値にフーリエ変換の余弦波を積和した積和値、ＡＴＡＮはアークタンジェント、ｓｉｔａはフーリエ変換で得られる位相値を示す。）

In the above equation, SV is a product-sum value obtained by multiplying and summing a pixel value of a pixel by a Fourier transform sine wave, CV is a product-sum value obtained by multiplying and summing a pixel value of a pixel by a cosine wave of Fourier transform, and ATA is an arc tangent, site Indicates a phase value obtained by Fourier transform. )

  The phase obtained by the Fourier transform indicates the light and shade direction of the image existing within the circumference, and becomes the light and shade direction of the image at the pixel at the processing point P at the center point. That is, this light and shade direction is recognized as a normal vector with respect to the boundary between the object B and the surrounding background W in the two-dimensional captured image 1, that is, the contour portion of the object B.

  For example, when the circle 11 is on the object B as shown in FIG. 6 (a), it is rotated counterclockwise from a predetermined reference position to perform a Fourier transform into a circle, and a sine wave and cosine wave are converted into a circle. When the pixel values of the pixels on the circumference are integrated, a waveform peak appears at around 90 degrees as shown in FIG. As a result, the phase obtained by the above circular Fourier method just indicates the normal direction with respect to the contour portion of the object B. In FIG. 6B, sin is a sine wave, cos is a cosine wave, s · p is a waveform obtained by integrating the sine wave and the pixel value, and c · p is a waveform obtained by integrating the cosine wave and the pixel value. ing.

  However, since the contour portion of the object image B is not included in the circumference and the image in the circumference has almost uniform density distribution, the SV and the formula shown in the following equation (2) are excluded. When the square root value (pow) of the square sum of CVs is less than a certain value, the processing point P is treated as having no shading.

  In order to extract the person image h, as shown in FIG. 7A, first, the standard human image o is arranged with reference to an arbitrary arrangement point Q on the captured image 1, and the circular Fourier method is used. Thus, the normal vector group b in the contour portion of the standard human figure image o is obtained, and the position information from the arrangement point Q to the normal vector group b and the angle information of the normal vector group b are used as standard data of the arrangement point Q. Remember.

  Next, as shown in FIG. 7B, a normal vector group b ′ at the contour portion of the person image h is obtained by the circular Fourier method with respect to the captured image 1 on which the person image h moving during monitoring is shown. Based on the standard data, a plurality of correct answer points corresponding to the arrangement point Q are obtained from the normal vector group b ′, and if the so-called vector focus G where these correct answer points are gathered is equal to or greater than a certain value, the person image Recognize that h is a person. This vector focus G becomes the virtual centroid G of the person, and the centroid G is tracked for each captured image 1 sent from the camera 9.

  In the first embodiment, the virtual barycentric point G is preferably arranged at the front (front) of the person image h, the center of the head, or the like. Thereby, the approach and exit of the person image h into the interest measurement area A can be reliably captured, and the reliability of the interest level measurement can be improved.

  According to the image processing technique based on the vector focus method described above, the captured image 1 sent from the camera 9 can be continuously processed, and the human image h can be defined by a virtual center of gravity G called a vector focus for each processing. The movement of the virtual barycentric point G can be tracked, and the instantaneous moving speed of the virtual barycentric point G can be measured by the tracking.

  Therefore, according to the person recognition means 31, each person appearing in the captured image 1 can be recognized one by one. For example, regarding the two persons crossing the shoulder, the individual person can be distinguished and recognized. Can be recognized. The human recognizing means 31 then moves the virtual centroid G of the person image h after the virtual centroid G of the person image h appears in the captured image 1 across the boundary, and then the captured image 1 across the boundary. Draw a tracking line (movement trajectory) until it disappears and track it.

The stay time measuring means 32 enters the inside of the measurement area A across the virtual center of gravity G of the person image h specified by the person recognition means 31, and then enters the inside of the measurement area A across the boundary. The time until leaving, that is, the staying time is measured and recorded.
When the virtual centroid G of the person image h specified by the person recognition means 31 crosses the boundary of the interest measurement area A and goes out of the boundary, or disappears outside the captured image 1, the person recognition means described above. 31, information recorded by the stay time measuring means 32 is output to the interest level determination device 4.

  The degree-of-interest discriminating device 4 evaluates the degree of interest of each shopper with respect to the product in several stages based on the discriminant criteria based on the stay time input from the image processing device 3, and discriminates the degree of interest in each stage. Is. The interest level determination device 4 includes an interest level determination means 41 and an interest level parameter 42.

  The discrimination criterion is registered in the interest level parameter 42. This discrimination criterion is for distinguishing the degree of interest into several levels, and a reference value corresponding to the level of each level is set. And this discrimination | determination reference | standard is made into the time reference for contrasting with the said stay time. As shown in the table of FIG. 8, for example, when the interest level is evaluated in four stages in ascending order, level 0 is 0 to 9 seconds, level 1 is 10 to 29 seconds, and level 2 is 30 seconds to 59 seconds, level 3 is set to 60 seconds or more, and so on.

  It should be noted that how to set the discrimination criteria is subject to the degree of closeness, gaze time, etc. of the interested person (shoppers) depending on the type and size of the displayed products. It should be decided as appropriate in consideration of the behavior of the shopper according to the type, size, and exhibition environment of the product.

  The degree-of-interest discriminating means 41 compares the staying time input from the image processing device 3 with the discrimination criterion set in the degree-of-interest parameter 42 to determine the degree of interest of each person at several levels. Sort into one.

  The degree-of-interest data storage device 5 is for registering the degree-of-interest data by the interest degree discriminating device 4. For example, as shown in FIG. 9, the data storage device 5 according to the degree of interest can have a data structure in a tabular format according to the degree of interest, the number of people, and the total interest time for each date and time. The interest level-specific data storage device 5 adds 1 to the corresponding level of interest level in the table of FIG. 9 corresponding to the interest level determined by the interest level determination device 4 to obtain the total interest time. The stay time is added and stored as a cumulative value.

  The data processing device 6 reads the interest level-specific data registered in the interest level-specific data storage device 5 and processes the data. The data processing device 6 processes the data stored in the interest degree-specific data storage device 5 into a predetermined output format for output by the output device 8. For example, this includes totaling by specific time zone, totaling by day, table creation, and graph creation.

  The output device 8 outputs the interest level information from the data processing device 6. The output device 8 is configured by, for example, a monitor, a printer, and the like, and the product interest level measurement result processed by the data processing device 6 into a predetermined output format is output by the output device 8. For example, the output device 8 is output as a table and a graph as shown in FIG.

Next, the operation of measuring the merchandise interest level for the mannequin display shown in FIG. 1 by the merchandise interest level measuring apparatus 2 of the first embodiment will be described.
First, the boundary of the measurement area A of interest is set on the captured image 1 captured by the camera 9 with respect to the image processing device 3 by the operation means 7 and the determination criterion of the interest level is set for the interest level determination device 4. To do. As shown in FIG. 1, the interest measurement area A is located in front of the mannequin M that can be approached by a shopper within a close range (50 cm) from which the mannequin M can come in contact with the product worn on the mannequin M. Is set on the captured image 1 in a fan shape so that the certain range is the interest measurement range. As a discrimination standard, a time standard is set. For example, as shown in the table of FIG. 8, four levels are set. Level 0 is 0 to 9 seconds, Level 1 is 10 to 29 seconds, Level 2 is set. Is set to 30 to 59 seconds, and level 3 is set to a time reference of 60 seconds or more.
When the above setting is completed, the degree of interest of each item of the shopper in the mannequin display is measured as follows, and the degree (level) is determined as follows.

Referring to the flowchart of the operation of the image processing apparatus 3 shown in FIG. 3, when the captured image 1 is continuously input from the camera 9 to the image processing apparatus 3 (S1), the image processing apparatus 3 When the person is recognized and the presence of the person is recognized, the person is captured by the virtual barycentric point G (S2), and the virtual barycentric point G is tracked over time on the continuously input captured image 1. (S3). Note that this person recognition uses the image processing technique based on the vector focus method as described above (see FIGS. 5 to 7).
Then, when the virtual centroid point G crosses the boundary of the measurement area A of interest and enters the inside thereof, timing starts, and the time until the virtual centroid point G crosses the boundary of the measurement area A of interest and goes outside thereof, That is, the stay time t in the region of interest A is measured (S4). Thereby, the time estimated that the person in the real space approaches the mannequin M displaying the product and shows interest in the product in the range of the closest distance of the product is measured.
Next, when the virtual center-of-gravity point G crosses the boundary of the region of interest measurement A and goes outside, that is, when a person in the real space (shopper) leaves the range of the mannequin M displaying the product. The image processing apparatus 3 outputs the measured staying time to the interest degree determination apparatus 4 (S5).

  Then, the interest level determination device 4 determines the level of interest (degree of interest) by comparing the stay time input from the image processing device 3 with a time reference as a determination reference. That is, referring to the flowchart of the operation of the degree-of-interest discriminating apparatus 4 shown in FIG. 4, when the stay time t is input from the image processing apparatus 3 (S10), the stay time t is level 0 (0 seconds to 9 seconds). Seconds), level 1 (10 seconds to 29 seconds), level 2 (30 seconds to 59 seconds), or level 3 (60 seconds or more) not corresponding to these (S11 to S13). Then, the interest level determination device 4 adds 1 to the data storage device 5 classified by interest level and adds the stay time t to the total interest time, and stores each as a cumulative value. (S14 to S17). In addition, the data storage device 5 classified by interest degree stores the date and time when the stay time t is measured.

Through the above operation, the captured images 1 sequentially sent from the camera 9 are continuously processed, and the degree of interest is measured and recorded. This interest level measurement is normally performed within the opening hours of the store where the mannequin display is performed.
Then, the interest level-specific data registered in the interest level-specific data storage device 5 is processed into a predetermined output format by the data processing device 6 by operating the operating means 7, and is displayed by the output device 8 in FIG. Are output in the form of a table or graph.

As described above, according to the product interest level measuring apparatus 2 according to the first embodiment, a shopper who is presumed to have exhibited special interest in the displayed product is identified, and the shopper observes the product. The time (stay time t) estimated to be spent on the customer is measured, and the level of interest (interest level) of the shopper is evaluated and discriminated based on the length of the stay time t. Since the measurement data is aggregated according to the stage, the collected data according to the degree of interest is made objective and quantified.
That is, not only the presence / absence of the shopper's interest in the product, but also the level of interest is determined in multiple stages with high accuracy, so the degree of interest in the product can be easily and quantitatively sensed by a numerical value.
Therefore, the degree of interest is highly reliable with respect to the aggregation result, and the data for each degree of interest can be processed into various types of interest information and output.
Therefore, it is possible to objectively and quantitatively grasp the degree of interest of each shopper for each product (the level of strength of the shopper's overall interest) with high accuracy. It is possible to grasp objectively and quantitatively.

Note that since a shopper who is determined to have a particularly high degree of interest is identified, it is possible to record the image and use it as information at a later date.
In addition, how to process and display the data by interest level should be appropriately devised and selected according to the product type, customer base, etc. In addition to time-dependent changes, data by time, etc., displaying by percentage by interest level, or displaying by percentage of the total number of shoppers, etc. is effective in analyzing and understanding the degree of overall interest. It is.

(Embodiment 2)
The merchandise interest level measurement device 2 according to the second embodiment (the embodiment of the present invention) is an example including all of the first embodiment and including the points (li) to (le) of the solving means.
That is, in the second embodiment, the image processing apparatus 3 is provided with speed measuring means for measuring the instantaneous moving speed of the virtual center of gravity G in the interest measurement area A, and the degree of interest determination apparatus 4 includes: The interest level parameter 42 is provided with a speed reference to be compared with the instantaneous movement speed as a discrimination reference, and the speed reference is set to be equal to or lower than the slow movement speed close to the stop state, and the stay is performed only when the instantaneous movement speed hits the speed reference. The degree of interest is evaluated based on the length of time.

The speed measuring means measures and records the instantaneous moving speed for every predetermined time while the virtual center of gravity G of the person image specified by the person recognizing means 31 is located in the measurement area of the measurement area of interest, and records the virtual speed. Whether or not the center of gravity G has stopped moving (or almost stopped) can be detected.
Further, the speed reference of the interest level parameter 42 is a reference for the instantaneous moving speed, not the average moving speed of the virtual center of gravity G, and is equal to or lower than the moving speed when the product is particularly paid attention, for example, 0. .1m / sec or less shall be specified.

  According to the second embodiment, only when the instantaneous movement speed of the person shown at the virtual center of gravity G is equal to the speed reference, the stay time in the measurement area A, that is, the close distance of the product The level of interest is determined and recorded according to the length of. As a result, people who are unrelated to the behavioral interest in the product, such as accidentally approaching the product due to other pedestrian's avoidance behavior, etc. are accurately distinguished from the target of interest level measurement and excluded. . Therefore, the accuracy and reliability of the aggregation result of the degree of interest according to the second embodiment are higher.

  The speed reference may be defined in two stages. In this way, the degree of interest of the shopper can be determined and analyzed in more detail by combining the instantaneous movement speed of the person and the staying time. it can.

The present invention is not limited to the above embodiment, and various design changes can be made.
For example, the operation means 7, the image processing device 3, the interest level determination device 4, the interest level-specific data storage device 5, the data processing device 6, and the output device 8 may each be composed of independent dedicated units, All or a part of these may be configured by a general-purpose computer such as a single personal computer.
When a plurality of products are displayed adjacent to each other and the degree of interest in each product is measured, a plurality of the interest measurement areas A are set for each product, and the degree of interest is measured for each product. You may do it.
The interest measurement area A can be set in various shapes such as a circle, a rectangle, and a polygon in accordance with the outer shape of the product, the mannequin M displaying the product, the display shelf, and the like.
In addition to determining the degree of interest (level of interest) of each shopper based on the staying time or a combination of staying time and moving speed, in addition to this, other data other than these, for example, the behavior of the shopper It is also possible to make a determination in consideration of the data (such as the course direction).

1: Captured image 2: Merchandise interest level measuring device 3: Image processing device 4: Interest level discriminating device 5: Data storage device classified by interest level 6: Data processing device 7: Operating means 8: Output device 9: Camera (imaging device) )
31: Person recognition means 32: Stay time measurement means 41: Interest degree determination means 42: Interest degree parameter A: Interest measurement area h: Person image M: Mannequin

Claims (2)

Based on a captured image in which the area where the product is placed is constantly imaged from above with a single camera, the degree of interest of each shopper for the product is measured and the measured value is determined in multiple stages, registered, A product interest level measuring device that processes this and outputs product interest level information,
Recognize each person on the captured image, track the movement of the virtual center of gravity of the person, and determine the stay time of the virtual center of gravity in the measurement area of interest set as the close range of the product. Having an image processing device to measure,
An interest level determination device for determining the degree of interest of each shopper for the product in several stages based on the determination criteria based on the stay time from the image processing device;
The image processing apparatus recognizes a rough outline of a person image in the captured image by a normal vector, and based on the normal vector, captures the person image as the virtual centroid by the vector focus method, Human recognition means for tracking the center of gravity point over time, and when the virtual center of gravity crosses the boundary of the measurement area of interest from the outside to the inside of the measurement area, the time measurement is started and the inside of the measurement area crosses from the inside to the outside Is an image processing device including a stay time measuring means that stops timing when
The image processing apparatus includes a speed measuring unit that measures an instantaneous moving speed within the measurement area of interest of the virtual center of gravity,
The degree-of-interest discrimination device has a speed reference for comparison with the instantaneous moving speed as a discrimination reference,
The degree of interest is determined based on the length of the staying time only when the speed reference is set to be equal to or less than the slow movement speed close to the stop state and the instantaneous movement speed hits this speed reference. Product interest level measuring device characterized by. Based on a captured image in which an area for placing a product is always captured from above with a single camera, the degree of interest of each shopper for the product is measured, and the measured value is determined in multiple stages and registered. , A product interest level measuring device that processes and outputs product interest level information,
An interest degree-specific data storage device for registering the interest degree-specific data by the interest degree determination device;
It has a data processing device that reads out the data by interest level registered in the data storage device by interest level and processes the data,
An output device for outputting interest level information from the data processing device;
The operation means sets the boundary of the product interest measurement region on the captured image for the image processing device, and sets the interest level determination criterion for the interest level determination device.
The discrimination criterion in the degree-of-interest discrimination device is a time criterion to be compared with the staying time, and the time criterion is a discrimination criterion set in multiple stages so that it becomes longer as the interest level is higher. Item 1. A commodity interest degree measuring device.

Images