JP2008102781A - Information guidance system and its program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an information guidance system and its program capable of discriminating types of a user with high accuracy and providing guidance information according to a discriminated type of the user without imposing no burden on the user to wear any special device and the like. <P>SOLUTION: The information guidance system photographs a moving object by a video photographing apparatus, extracts an image of the moving object as a moving image area, for example, a face and a hand of a person as a specific part of the person, discriminates a moving direction, a moving speed, and a moving range of the face or the hand, or a relative positional relation between the face and the hand, discriminates whether the person is a white cane user or a wheelchair user, and outputs guidance suited to the person concerned. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a system and program for guiding movement support information, disaster information, accident information, crime prevention information, event information, and the like to a person who moves or uses a public facility or the like in a city.

  Conventionally, in public places such as stations and parks, for example, a tactile board that displays braille descriptions of places and facilities such as toilets has been installed so that visually impaired persons can obtain information. ing. Also, by pressing the switch of the pendant type transmitter that can be worn down from the neck by hand, a predetermined weak electromagnetic wave is transmitted, and when the weak electromagnetic wave is received by the receiver, the sound reproduction device is activated and the equipment There is also a voice guidance system for obtaining the guidance information by voice.

  In Patent Document 1, only a card with a built-in transmitter is carried, no switch operation is required, and the voice guidance apparatus installed in the vicinity of various facilities such as a toilet is within a predetermined distance range. A voice guidance system has been described in which voice guidance guidance information is automatically obtained for the equipment at the time, thereby making it easier to use various facilities.

  In this way, as a system that changes the type of guidance service depending on the type of user, ID codes for blind people, elderly people, hearing impaired people, etc. are provided in the transmission signal of the portable wireless card, and guidance is provided according to the ID code. There is a system that changes contents and guidance means (voice / image).

In addition, as a system that does not require a special item such as a switch-operated radio device as a condition for starting a guidance service or selecting a type, there is a switch operation or wireless reception. And a system for providing guidance services for blind people is described.
Japanese Patent No. 2839797 JP 2003-168110 A

  The conventional guidance system has a problem that a burden on the user is heavy, such as a switch operation is required for starting a guidance service or selecting a type, or a special device is required to be carried. In addition, in the case of a system based on detection of a white cane by image, there is a problem that if the color and shape of the cane match, it is detected even if it is not a blind person.

  An object of the present invention is information that enables a user to determine the type of user with high accuracy without the burden of wearing a special device, and to guide information according to the determined type of user. It is to provide a guidance system and an information guidance program.

The present invention employs the following configuration in order to solve the above problems.
That is, according to one aspect of the present invention, an information guidance system and an information guidance program of the present invention extract image data receiving means for receiving image data, and a moving image area from image data received by the image data receiving means. A moving image region extracting unit; a specific portion extracting unit that extracts a specific portion from the moving image region extracted by the moving image region extracting unit; a case determining unit that determines information about the specific portion extracted by the specific portion extracting unit; A moving object discriminating unit that discriminates the type of the moving object based on a discrimination result by the case discriminating unit; and a guidance output unit that outputs a corresponding guide based on the discrimination result by the moving object discriminating unit.

Further, it is desirable that the moving object discriminating means discriminates the type of moving object by fuzzy reasoning or fuzzy associative reasoning.
In the information guidance system and the information guidance program according to the present invention, the image data receiving unit receives image data captured by an imaging device, and the moving image region extracting unit extracts a moving human image as a moving image region. The specific part extracting means extracts a human face part or a human hand part as a specific part, and the case determining means is a moving direction, a moving speed, a moving range or the face of the face part or the hand part. It is desirable to determine the relative positional relationship between the portion and the hand portion, and the moving object determining means determines whether the person is a white cane person or a wheelchair user.

  According to the present invention, since it is not necessary for the user to wear a special device or the like, the burden on the user can be reduced, and the type of user can be determined only by analyzing the operation status of the user. And can provide the most appropriate guidance to the user.

FIG. 1 is a functional block diagram of an information guidance system according to an embodiment of the present invention.
The information guidance system according to the embodiment of the present invention includes video photographing means 101, motion detection means 102, inference means 103, and output means 104. Motion detection means 102 includes video acquisition means 102-1 and video data input. The inference means 103 includes membership function means 103-1, case base means 103-2, case determination means 103-3, and moving object determination means 103-4.

The video photographing means 101 is, for example, a video camera and is installed at a specific location such as an entrance of a facility, for example, and photographs a person passing through that location.
When there is one video camera, the subject person is photographed from the horizontal direction. In this case, the relative distance between the target person and the video camera is preferably 3 to 5 m. In the case of the video photographing means 101 capable of performing stereo viewing, for example, a human face and hand are photographed from the front direction with two video cameras. In stereo view, the distance between the video camera and the target person can be calculated, so the distance between the target person and the video camera can be set freely.

FIG. 2 is a diagram for explaining stereo viewing.
In FIG. 2, the coordinates of the object 201 are P (x, y, z), the coordinates of the left video camera with respect to the object 201 are OL, the coordinates of the right video camera with respect to the object 201 are OR, and the left The coordinates of the image plane of the object 201 in the image photographed by the video camera are P _L (X _L , Y _L ), and the coordinates of the image plane of the object 201 in the image photographed by the right video camera are P _R ( X _R , Y _R ). These coordinates P _L and P _R are the coordinates in the image on the imaging surface.

  Here, d is the relative distance between the center coordinate OL of the left video camera image and the center coordinate OR of the right video camera image, and f is the focal length of the left video camera and the right video camera. It is.

Then, the coordinates P (x, y, z) of the object 201 are obtained by the following expressions (1) to (3).
x = d · X _L / (X _L −X _R ) (1)
y = d · Y _L / (X _L −X _R ) (2)
z = d · f / (X _L −X _R ) (3)
Returning to the description of FIG.

The video acquisition unit 102-1 acquires the video data output from the video shooting unit 101, extracts a moving object from the video data, and outputs the moving object to the video data input unit 102-2.
As a method for extracting a moving object, for example, there is a background difference method. The background difference method is a method of comparing the brightness and luminance of an input image and a background image and extracting a moving object based on the difference.

FIG. 3 shows an example of an image by the background subtraction method.
In this embodiment, the difference between the RGB value of the original background image and the input image data is taken, and if the difference is within the threshold range, it is displayed in black (no background change) and exceeds the threshold value (Not in the background image) outputs the video.

As another method for extracting a moving object, there is an inter-frame difference method or the like. In the present invention, any method can be used according to the surrounding situation.
Returning to the description of FIG.

In the video data input unit 102-2, the video acquisition unit tracks a portion designated in the display screen with respect to the moving object output from the unit 102-1.
FIG. 4 is a diagram illustrating face and hand tracking.

  In the present embodiment, the skin color is tracked as the color designated in the display screen, and is determined to be a part of a human image. A tracking target within a certain range is determined as one group, and the upper tracking portion 401 in the group is recognized as a face and the lower tracking portion 402 is recognized as a hand.

For example, the face and the hand are discriminated based on, for example, the size of the area of the skin color, and the larger one of the two is discriminated as the face, and the smaller one is discriminated as the hand.
Then, the center of gravity coordinates of the face and the center of gravity coordinates of the hand are calculated. Further, from the calculated center of gravity coordinates of the face and the center of gravity coordinates of the hand, the relative position of the hand with respect to the face and the movement speed of the person are calculated from the difference in the coordinates of the position of the face per predetermined time.

Returning to the description of FIG.
The video data input means 102-2 outputs each data calculated as described above to the membership function means 103-1.

The membership function means 103-1 stores a membership function, and outputs a membership value corresponding to the input value based on the membership function from the input value.
In the present embodiment, a plurality of characteristic motion membership functions are used.

  There are three types of input values: “hand movement”, “moving speed”, and “face and hand position”. Each membership function is created in advance and stored in membership function means 103-2. Keep it.

FIG. 5A is a diagram illustrating a hand movement graph according to the present embodiment. The vertical axis is the horizontal position of the hand relative to the face, and the horizontal axis is the time.
A graph 501 represented by a solid line in FIG. 5A is a graph representing the position of the hand of a healthy person, and has extreme points 506, 507, and the like. A graph 502 represented by a dotted line is a graph representing the position of the hand of a white cane who is a visually handicapped person using a white cane, and has extreme points 508, 509, and the like. Here, a healthy person refers to a person who has little trouble in walking, and refers to a person other than a white cane person or a wheelchair user who uses a wheelchair. Moreover, although the moving object in this Embodiment demonstrates using the example of a healthy person and a white cane person, a healthy person and a wheelchair user, or a healthy person, a white cane person, and a wheelchair user may be sufficient. .

FIG. 5B is a diagram illustrating a hand movement membership function according to the present embodiment. The vertical axis represents the membership value, and the horizontal axis represents the difference between the extreme values.
In FIG. 5B, membership functions indicating the degree of hand movement are μ _1S 503, μ _1M 504, and μ _1B 505, and the labels of the membership functions μ _1S 503, μ _1M 504, and μ _1B 505 are Each is small, medium and large.

FIGS. 6A and 6B are models representing hand movements. FIG. 6A is a diagram illustrating a model for a healthy person, and FIG. 6B is a diagram illustrating a model for a white cane person.
Since a healthy person moves his / her hand 602 back and forth with respect to the face 601 as shown in FIG. 6A when walking normally, a graph with a large amplitude like the graph 501 shown in FIG. 5A is obtained. . On the other hand, since the white cane person walks with the white cane in front when walking, the hand 604 is positioned forward with respect to the face 603 as shown in FIG. A graph with a small amplitude like the graph 502 shown in FIG.

  In the present embodiment, for example, when the graph 501 shown in FIG. 5A is obtained, the difference between the extreme values of any two arbitrary points is calculated. For example, if the upper extremum point 505 and the lower extremum point 506 are used, the respective extrema are 70 and -90, and therefore the difference is 70-(-90) = 160.

When this is applied to the membership function of FIG. 5B, μ _1S (160) = 0, μ _1M (160) = 0, and μ _1B (160) = 1. That is, the membership value of the membership function μ _1S 503 with a small label is 0, the membership value of the membership function μ _1M 504 in the label is 0, and the membership value of the membership function μ _1B 505 with a large label is 1 It will be.

Similarly, when the graph 502 shown in FIG. 5A is obtained, the difference between the extreme values of the extreme points 507 and 508 is calculated to be 10 − (− 10) = 20, which is shown in FIG. When applied to the membership function of b), μ _1S (20) = 1, μ _1M (20) = 0, and μ _1B (20) = 0. That is, the membership value of the membership function μ _1S 503 with the small label is 1, the membership value of the membership function μ _1M 504 in the label is 0, and the membership value of the membership function μ _1B 505 with the large label is 0. It will be.

FIG. 7A is a diagram illustrating a moving speed graph according to the present embodiment. The vertical axis is the moving speed, and the horizontal axis is the time.
A graph 701 represented by a solid line in FIG. 7A is a graph representing the movement speed of a healthy person, and a graph 702 represented by a dotted line is a graph representing the movement speed of a white cane person. Usually, a healthy person has a faster moving speed.

FIG. 7B is a diagram showing a membership function of the moving speed according to the present embodiment. The vertical axis represents the membership value, and the horizontal axis represents the moving speed.
Membership functions indicating the degree of moving speed are μ _2S 703 and μ _2B 704, and the respective labels are slow and normal.

  In the present embodiment, for example, when a graph 701 is obtained, an average of speeds within a predetermined time is calculated. For example, if the average of the speeds at times 2 to 4 seconds is calculated, 2000 is obtained.

When this is applied to the membership function of FIG. 7B, μ _2S (2000) = 0 and μ _2B (2000) = 1. In other words, the membership value of the late label membership function 703 is 0, and the membership value of the normal label membership function 704 is 1.

Similarly, when the graph 702 is obtained, the average of the velocities at times 2 to 4 seconds is calculated to be 500. When this is applied to the membership function of FIG. 7B, μ _2S (500) = 0. .8, μ _2B (500) = 0. That is, the membership value of the membership function 703 with the late label is 0.8, and the membership value of the membership function 704 with the normal label is 0.

FIG. 8A is a graph showing the positional relationship between the face and the hand according to the present embodiment. The vertical axis is the horizontal position of the hand relative to the face, and the horizontal axis is the time.
A graph 801 represented by a solid line in FIG. 8A is a graph representing the position of the hand of a healthy person, and has extreme points 805 and the like. A graph 802 represented by a dotted line is a graph representing the position of the hand of a white cane and has an extreme point 806 and the like.

FIG. 8B is a diagram showing the membership function of the positional relationship between the face and the hand according to the present embodiment. The vertical axis represents the membership value, and the horizontal axis represents the hand position.
Membership functions that indicate the degree to which the hand position changes only forward or back and forth are μ _3S 803 and μ _3B 804, and each label is constant.
And change.

  In this embodiment, for example, when a graph 801 is obtained, a membership value corresponding to the extreme value is obtained using a certain extreme value. For example, when an extreme point 805 is used, the extreme value is 35.

When this is applied to the membership function of FIG. 8B, μ _3S (35) = 0 and μ _3B (35) = 1. That is, the membership value of the constant label membership function 803 is 0, and the membership value of the label change membership function 804 is 1.

Similarly, when the graph 802 is obtained, the extreme value of the extreme point 806 is 4, and when this is applied to the membership function of FIG. 8B, μ _3S (4) = 0.6, μ _3B (4) = 0.4. That is, the membership value of the constant label membership function 803 is 0.6, and the membership value of the label change membership function 804 is 0.4.

Returning to the description of FIG.
The case base unit 103-2 stores a case base, and in this embodiment, inference is performed using the case base as a fuzzy rule.

FIG. 9 is a diagram illustrating an example of a case base according to the present embodiment.
In the example of the case base shown in FIG. 9, the following four rules are stored.
Rule A: if hand movement is small & speed is slow & face and hand position is constant then white cane Rule B: if hand movement is small & speed is normal & face and hand position is constant then white Cane Rule C: if hand movement is medium & speed is normal & face and hand position is change then normal person Rule D: if hand movement is large & speed is normal & face and hand position is change then Healthy Person Here, for example, the rule A is that if the hand movement is small, the moving speed is slow, and the positional relationship between the hand and the face is constant, it is the white cane that is operating. This is the rule.

Returning to the description of FIG.
The case determination unit 103-3 compares the input value with the case base value. That is, the fitness of each input value for each rule stored in the case base means 103-2 is calculated. Specifically, a fuzzy logic operation AND of membership values is performed. When the calculated fitness is larger than a threshold value (for example, 0.2. This value can be changed according to the situation), the process proceeds to the moving object determination unit 103-4. The threshold value is set in increments of 0.1 from 0 to 1. When it is 1, it depends on the membership function, and when it is 0, the case is not discriminated. The other part depends on the weight of the membership function, and the case is discriminated. If the matching degrees for each rule are all within the threshold value, that is, if there is no rule that matches the existing case base, the input value at that time is stored in the case base as a new case. At that time, the value stored in the case base is only the data of the case, and the user gives later what the output for the case is.

  The moving object discriminating means 103-4 infers what the moving object is based on the output of the membership function means 103-1 and the contents stored in the case base means 103-2. For example, it is inferred whether the moving object is a healthy person, a white cane person, or a wheelchair user.

The moving object discriminating means 103-4 can be inferred by fuzzy associative inference or fuzzy inference, and can be used by switching according to the situation.
When inferring by fuzzy associative reasoning, the moving object discriminating means 103-4 is constituted by a fuzzy associative reasoning system. A basic component of the fuzzy associative reasoning system is a bidirectional associative memory (BAM) of a neural network. In fuzzy associative reasoning, a network using a plurality of BAMs is constructed, and inference is performed by an echo operation that repeats propagation of active values. First, bidirectional associative memory will be described.

FIG. 10 is a diagram illustrating a network configuration of bidirectional associative memory.
As shown in FIG. 10, there is a network constituted by the layer _{L A} and layer _{L B,} the layer _L plurality of nodes to _{A a 1, a 2, ...} , there is _{a n,} a plurality in layers _{L B} Nodes b ₁ , b ₂ ,..., B _p exist. The connection weight between nodes is expressed by an association matrix MεR ^{n × p} . Pattern A∈R ⁿ for this association matrix M Layer _{L A} and layer _{L B,} when to be noted the pair B∈R ^p, given A'∈R ⁿ containing noise to the layer _{L A} as also, by the echo operation between layers L _a and layer L _B, the pattern B is recalled. Conversely, the pattern A will occur can give a pattern B'∈R ^p containing noise to the layer L _B. Association in this bidirectional associative memory is performed by the following equations (4) and (5).

Here, a _i (t) and b _j (t) are active value vectors of the respective nodes in step t and take values belonging to the section [0, 1]. It may be a value of {0, 1}. φ represents each node A (t) = (a ₁ (t), a ₂ (t),..., a _n (t)), B (t) = (b ₁ (t), b ₂ (t),. , B _p (t)), for example, a sigmoid function is used. The associative matrix M is obtained by the following equations (6) and (7), where the pattern pairs to be recorded are (A ₁ , B ₁ ),..., (A _m , B _m ).

Here, β _i is an associative coefficient. In the Kosko method, when an associative matrix M is obtained, a pattern pair obtained by bipolar conversion of a pattern pair of {0, 1} is used. However, if the association matrix M obtained by this Kosko technique is used, the network may cause a saturation state, and adjustment is necessary. In order to solve this problem, a method for normalizing the associative matrix M has been proposed. Therefore, in the fuzzy associative memory system used in the present embodiment, a normalized associative matrix obtained by this normalization method is used.

Next, a method for constructing a fuzzy associative memory system will be described.
FIG. 11 is a conceptual diagram of the fuzzy associative system of the present embodiment.
In the layer setting, the layer representing the antecedent part of the fuzzy rule is the If (X) layer (this corresponds to the membership function means 103-1 shown in FIG. 1), and the layer representing the consequent part is Then. (Y) Layer. A node in the If (X) layer represents a fuzzy set (membership function) of the antecedent part, and a node in the Then (Y) layer represents a fuzzy set (membership function) of the consequent part or an input / output function. . Further, since the correlation between the antecedent part and the consequent part exceeds the limit that BAM can store, one node sets a Rule (R) layer representing one rule. Therefore, the fuzzy rule can be expressed by configuring a BAM between the If (X) layer and the Rule (R) layer (associative matrix M _xr ). In addition, R
In the ule (R) layer, in order to prevent the explosion of ambiguity, a coordinator combination (M _rr ) having a function of suppressing activation of other nodes is set. Coordinator coupling means coupling to a node with a positive weight at an arbitrary node in the same layer and coupling to another node with a negative weight.

  Next, the concept of construction on the memory will be described by taking the network configuration of FIG. 11 as an example. First, in order to obtain an association matrix between each layer, the correlation of fuzzy rules is replaced with a value belonging to the interval [0, 1]. In some cases, a binary value of 0 or 1 indicated by {0, 1} may be substituted. For example, when the fuzzy rule in FIG. 9 is replaced, the result is as shown in the following equation (8).

In this equation (8), as an example, the correlation is represented by binary values of 0 and 1 represented by {0, 1}, but it is needless to say that the present invention is not limited to this.

  Using this calculation result as a pattern pair to be recorded, an associative matrix between each layer is obtained by the equations (6) and (7). However, the Kosko method obtained using a pattern pair obtained by bipolar conversion is not used.

As described above, the fuzzy associative memory system used in the present embodiment uses a normalized associative matrix at the time of inference. Thus, where an associative matrix obtained saves obtains the normalized associative matrix M _e by the following equation (9).

M _e = a (M + B) (9)
Here, each parameter is as follows.

Here, M, M _e , BεR ^{m × n} , and m _ij are M (i, j) elements.
In the embodiment, the associative matrix and the normalized associative matrix are calculated in advance before performing the inference, and are stored in, for example, the moving object discriminating unit 103-4.

  Here, when a plurality of rule sets are defined, the above processing is repeated by the number of rule sets to propose a plurality of networks. Then, information relating to the connection between the rule sets is read from the text file, and the respective networks are connected. At the time of inference, the weight (association matrix) of the connection between defined rule sets is also normalized. From here, the network constructed by the fuzzy rule R is referred to as a network R.

Next, the operation process of the fuzzy associative memory system will be described.
Inference is performed by repeating the propagation of active values while maintaining synchronization throughout the network. First, when an input is given, a value evaluated by each membership function (which is obtained by the membership function means 103-1 shown in FIG. 1) is output from the If (X) layer. Then, by applying the associative matrices M _xr , M _rr , and M _ry to the equations (4) and (5), the propagation of the active value is repeated by the set number of associative steps. For example, the output out _j of each node OUT layers connected by associative matrix _{M r} multiple INr layers can be determined by the following equation (13).

The proposed initial associative reasoning does not integrate reasoning results and can cause excessive ambiguity. In order to solve this, the formulas (14) to (16) are used to
(Y) Normalize the layer activation value.

Here, n is the number of nodes in the Then (Y) layer, γf _i is the activation value of node i, and γ _i is the contribution.

Thereby, the contribution degree γ _i is obtained. An inference result is obtained by interpolating the membership function of the consequent part or the output value of the input / output function with this contribution, and MAX-joining (summing) the obtained function (value).

Next, fuzzy inference will be described.
Fuzzy reasoning performs approximate reasoning based on fuzzy rules. In inference in fuzzy control that emphasizes real-time characteristics, the subjective ambiguity of an expert is expressed by a membership function as shown in FIGS. 5 (b), 7 (b), and 8 (b). House knowledge is expressed by fuzzy rules described by an antecedent part consisting of if ~ and a consequent part consisting of then ~. An example of fuzzy rule expression is shown below.

Rule A: if (x ₁ is PM) & (x ₂ is ZO) then y is ZO
Rule B: if (x ₁ is ZO) & (x ₂ is PM) then y is NM
At this time, the rule A expresses a fuzzy proposition that “if the input x ₁ is slightly larger on the plus side and the input x ₂ is near zero, the output is near zero”. In the rules, x ₁ and x ₂ indicate inputs and indicates an output. PB, PM, PS, and ZO mean positive big, positive medium, positive small, and zero, respectively, and are fuzzy sets that indicate the subjective ambiguity of experts. These are defined by membership functions.

An example of fuzzy inference will be described.
FIG. 12 is a diagram showing a process of fuzzy inference.
It is assumed that x ₀₁ = 0.7 and x ₀₂ = 0.2 are observed at inputs x ₁ and x ₂ in rules A and B, respectively. In the case of rules A and B, the membership functions g _a (y) and g _b (y) for x ₁ = x ₀₁ and x ₂ = x ₀₂ are
g _a (y) = (μPM (x ₀₁ ) ΔμZO (x ₀₂ )) ○ μZO (y)
g _b (y) = (μZO (x ₀₁ ) ΔμPM (x ₀₂ )) ○ μNM (y)
Is required. Here, Δ represents a fuzzy logic operation AND (min). ○ indicates the T-norm operator of the generalized synthesis rule. By selecting the T-norm operator appropriately, inference according to the situation is realized. Here, if the T-norm operator is set to · (multiplication),
g _a (y) = ω _a・ μZO (y)
g _b (y) = ω _b・ μNM (y)
It becomes. Where ω _a and ω _b are membership values,
ω _a = μPM (x ₀₁ ) ΔμZO (x ₀₂ ) = 0.7
ω _b = μZO (x ₀₁ ) ΔμPM (x ₀₂ ) = 0.2
It is. As a final conclusion, an output membership function is obtained by the following equation (17).

Here, ∇ indicates a fuzzy logic operation OR (max). In fuzzy inference, all the rules that have a relationship operate in parallel in proportion to their certainty.

  The moving object discriminating means 103-4 shown in FIG. 1 infers what the captured moving object is based on the above-described fuzzy associative reasoning or fuzzy reasoning. And the result is output to the output means 104 shown in FIG.

Returning to the description of FIG.
The output unit 104 outputs guidance based on the determination result of the moving object determination unit 103-4. For example, when the determination result is a white cane, the output unit 104 performs voice guidance. Thereby, the white cane can obtain necessary information by voice. For example, when the determination result is a healthy person, the output unit 104 can display guidance information such as characters.

FIG. 13 is a flowchart showing a process flow of the information guidance system according to the present embodiment.
The processing flow of the information guidance system will be described below with reference to FIGS. 1 and 13.

The membership function, case base, etc. used for inference are stored in advance in the information guidance system.
In step S1301, the moving image is captured by the image capturing unit 101, and the image acquisition unit 102-1 acquires the image.

In step S1302, a moving part, that is, a moving object is extracted from the acquired image by, for example, the background difference method, and is output to the video data input unit 102-2.
In step S1303, the video data input unit 102-2 tracks the specific part of the moving object, the face and the hand, calculates the center-of-gravity coordinates of the moving object, and further calculates the face based on the calculated center-of-gravity coordinates and hand-center-of-gravity coordinates. The moving speed of the person is calculated from the relative position of the hand to the face and the difference in the coordinates of the face position per predetermined time. Then, the calculated value is output to the inference means 103.

  In step S1304, using the value calculated in step S1303 as an input value, the fitness of each input value with respect to each rule stored in the case base unit 103-2 is calculated. If the degree of conformity to each rule is less than or equal to the threshold value, the process proceeds to step S1305, and if there is even one greater than the threshold value, the process proceeds to step S1306.

In step S1305, the input value used in step S1304 is stored in the case database unit 103-2 as a new case.
In step S1306, the moving unit determination unit 103-4 infers what the moving object is based on the input value.

In step S1307, guidance output is performed based on the result inferred in step S1306.
As described above, the embodiments of the present invention have been described with reference to the drawings. However, the above-described embodiments of the present invention are not limited to hardware, a DSP (Digital Signal Processor) board, or a CPU as a function of the information guidance system. It can be realized by firmware or software on the board.

  In addition, the information guidance system to which the present invention is applied is not limited to the above-described embodiment as long as the function is executed, and a system composed of a plurality of devices or a single device may be used. Needless to say, the integrated device may be a system in which processing is performed via a network such as a LAN or a WAN.

  It can also be realized by a system including a CPU, a ROM or RAM memory connected to a bus, an input device, an output device, an external recording device, a medium driving device, and a network connection device. That is, a ROM or RAM memory, an external recording device, or a portable recording medium that records a software program for realizing the system of the above-described embodiment is supplied to an information guidance system, and the information guidance system computer executes the program. Needless to say, this can also be achieved by reading and executing.

  In this case, the program itself read from the portable recording medium or the like realizes the novel function of the present invention, and the portable recording medium or the like on which the program is recorded constitutes the present invention.

  Examples of portable recording media for supplying the program include flexible disks, hard disks, optical disks, magneto-optical disks, CD-ROMs, CD-Rs, DVD-ROMs, DVD-RAMs, magnetic tapes, and nonvolatile memory cards. Various recording media recorded via a network connection device (in other words, a communication line) such as a ROM card, electronic mail or personal computer communication can be used.

  The computer (information processing apparatus) executes the program read out on the memory, thereby realizing the functions of the above-described embodiment, and an OS running on the computer based on the instructions of the program. Performs part or all of the actual processing, and the functions of the above-described embodiments are also realized by the processing.

  Furthermore, a program read from a portable recording medium or a program (data) provided by a program (data) provider is stored in a memory provided in a function expansion board inserted into the computer or a function expansion unit connected to the computer. After being written, the CPU of the function expansion board or function expansion unit performs part or all of the actual processing based on the instructions of the program, and the functions of the above-described embodiments are also realized by the processing. obtain.

  That is, the present invention is not limited to the embodiment described above, and can take various configurations or shapes without departing from the gist of the present invention.

It is a functional block diagram of the information guidance system concerning an embodiment of the invention. It is a figure for demonstrating a stereo vision. It is a figure which shows the example of the image | video by a background difference method. It is a figure which shows tracking of a face and a hand. (A) is a figure showing the graph of the hand movement which concerns on this Embodiment, (b) is a figure showing the membership function of the hand movement which concerns on this Embodiment. It is a model showing hand movement, (a) is a model in the case of a healthy person, and (b) is a figure showing a model in the case of a white cane person. (A) is a graph of movement speed, (b) is a figure showing a membership function. (A) is a graph showing the positional relationship between a face and a hand, and (b) is a diagram showing a membership function. It is a figure which shows the example of the example base which concerns on this Embodiment. It is a figure which shows the network structure of a bidirectional associative memory. It is a conceptual diagram of the fuzzy association system which concerns on this Embodiment. It is a figure which shows the process of a fuzzy reasoning. It is a flowchart which shows the flow of a process of the information guidance system which concerns on this Embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 Video imaging means 102 Motion detection means 102-1 Video acquisition means 102-2 Video data input means 103 Inference means 103-1 Membership function means 103-2 Case base means 103-3 Case discrimination means 103-4 Moving object discrimination means 104 Output means 201 Target 401, 402 Tracking portion 501 Graph (position of the hand of a healthy person)
502 graph (the position of the hand of a white cane)
503 Membership function μ _1S
504 Membership function μ _1M
505 Membership function μ _1B
506, 507, 508, 509 Extreme point 601 Face (healthy person)
602 hand (healthy person)
603 Face (white cane)
604 hand (white cane)
701 graph (moving speed of healthy person 702 graph (moving speed of white cane)
703 Membership function μ _2S
704 Membership function μ _2B
801 graph (position of the hand of a healthy person)
802 Graph (Position of a white cane's hand)
803 Membership function μ _3S
804 Membership function μ _3B
805,806 extreme points

Claims (4)

Image data receiving means for receiving image data;
A moving image region extracting unit that extracts a moving image region from the image data received by the image data receiving unit;
Specific part extracting means for extracting a specific part from the moving picture area extracted by the moving picture area extracting means;
Case discriminating means for discriminating information related to the specific part extracted by the specific part extracting means;
Moving object discriminating means for discriminating the type of moving object based on the discrimination result by the case discriminating means;
Guidance output means for outputting a corresponding guidance based on the determination result by the moving object determination means;
An information guidance system comprising:   2. The information guidance system according to claim 1, wherein the moving object discriminating means discriminates the type of moving object by fuzzy reasoning or fuzzy associative reasoning. The image data receiving means receives image data imaged by an imaging device,
The moving image region extracting means extracts a moving human image as a moving image region,
The specific part extracting means extracts a human face part or a human hand part as a specific part,
The case determination means determines a moving direction, a moving speed, a moving range, or a relative positional relationship between the face part and the hand part of the face part or the hand part,
The moving object determining means determines whether the human is a white cane or a wheelchair user;
The information guidance system according to claim 1 or 2, characterized by the above-mentioned. Computer
Image data receiving means for receiving image data;
A moving image region extracting unit that extracts a moving image region from the image data received by the image data receiving unit;
Specific part extracting means for extracting a specific part from the moving picture area extracted by the moving picture area extracting means;
Case discriminating means for discriminating information related to the specific part extracted by the specific part extracting means;
Moving object discriminating means for discriminating the type of moving object based on the discrimination result by the case discriminating means;
Guidance output means for outputting a corresponding guidance based on the discrimination result by the moving object discrimination means;
Information guidance program to function as.