This application is based upon and claims the benefit of priority from Japanese Patent Application No. 2008-311465, filed Dec. 5, 2008, the entire contents of which are incorporated herein by reference.
1. Field
The present invention relates to an information processing method and an information display apparatus utilizing face authentication. More specifically, the present invention relates to an information processing method that enables electronically exchanging (transmitting/receiving) video mail contents through a content collection device (e.g., a video recorder) or a content display device (e.g., a digital TV) without using a keyboard by recognizing a face of a user acquired by a camera, and to an information (message) display apparatus utilized in this method.
2. Description of the Related Art
Electronic mails have convenience that is not observed in phones (e.g., the other party does not have to be present in that location at that time, or a recipient side can see contents any time and anywhere). Therefore, mails have widely spread but have the following difficulties as compared with phones.
1> Using a keyboard to input information is general, but this is a large obstacle for elderly persons, young children, or disabled persons who are bad at using a keyboard.
2> Since a keyboard is required, there is a limit in an environment for transmitting or receiving mails (e.g., an environment where a mobile phone or a personal computer (PC) can be utilized is required).
3> It takes time to transmit a mail since a document text of the mail must be created, or a mail cannot be rapidly transmitted even though a user wants to inform the other party with something (the phone is easy since just making a call to talk can suffice, but the mail does not have such convenience).
4> The electronic mail is not oriented to a use application that a plurality of persons, e.g., family members see a mail and the respective persons offer responses.
It should be noted that a TV telephone dedicated device or a personal computer conventionally has a TV telephone function or a video chat function, but such a function is a continuation of the phone and does not have the above-explained convenience of the mail (e.g., the other person does not have to be present in that location at that time).
As a method of obtaining a mail environment where the difficulties of the mail are decreased as much as possible, a method utilizing face authentication can be considered. As a prior art utilizing a face authentication function, there is “Electronic Mail Creation Method and Apparatus Utilizing Image Recognition Function” disclosed in Japanese Patent Application Publication (KOKAI) No. 2005-267146.
In the Japanese Patent Application Publication, in a <mail address automatic acquisition process>, a facial image retrieval unit 104 (
According to the Japanese Patent Application Publication, in a <facial image retrieval process>, image analysis data received from an image analysis unit 205 (
That is, a face of a person who appears in a photograph taken by a camera of a portable communication terminal is recognized by using the technique disclosed in the above-described document, a “table in which faces and mail addresses of persons are registered in advance” is searched by using the recognized face to find out a mail address of the person who appears in the photograph, and the photograph is transmitted to this person. With such an operation, if contents transmitted through a mail can be created without a keyboard like a photograph, even a user who is unaccustomed to using a keyboard can readily transmit a mail.
However, when “using the face authentication function” disclosed in the document is adopted, there arises a problem that a mail may be possibly transmitted to a wrong person, for example. That is, face recognition that is 100% assured in processing using a machine alone is technologically difficult at present.
A general architecture that implements the various feature of the invention will now be described with reference to the drawings. The drawings and the associated descriptions are provided to illustrate embodiments of the invention and not to limit the scope of the invention.
Various embodiments according to the invention will be described hereinafter with reference to the accompanying drawings. In general, according to one embodiment of the invention, an information processing method comprising: capturing facial image data of a user; extracting a feature point of a face of the user from the facial image data and encoding the feature point; and associating the facial image data with a unique identifier and/or the encoded feature point to be registered in an address management table.
Embodiments of this invention will now be described hereinafter in detail with reference to the accompanying drawings.
The mail system depicted in
Besides, to the server 80 are appropriately connected a digital TV 20 having a camera 20a, an external personal computer (a PC) 30 having a camera 30a, a computer game console 40 having a camera 40a, a mobile phone 50 having a camera 50a, and others through the Internet.
A terminal of a contract firm 90 is connected with a provider who manages the server 80. Based on this provider, preference information of each terminal user can be collected through the Internet, and pinpoint information satisfying the preference of each terminal user can be provided to this user. For example, in the system depicted in
A face authentication/collation processing apparatus 100 is incorporated in each user terminal (20 to 60) so that features of a face of a user who is in front of the camera 60a can be encoded or features of a face of a user who is in front of the camera can be collated with features of registered faces in each user terminal. Image data of faces of users taken in by the cameras of the user terminals, feature point data obtained by encoding features of faces by the face authentication/collation processing apparatus 10, and others are registered in an address management table 800 provided in the server 80. A video mail itself (mail contents) from each user is associated with registered contents in the address management table 800 to be stored in a non-illustrated data storage in the server 80. Details of the address management table 800 and those of the face authentication/collation processing apparatus 100 will be described later.
In the system depicted in
At home, contents of the management information table 800 can be copied in a built-in HDD (not shown) of, e.g., the TV 10 or the recorder 60. Then, an in-family video mail (or message) can be exchanged by using a single device, e.g., the TV 10 or the recorder 60. In this case, the TV 10 or the recorder 60 has functions of the server 80 on the Internet.
In each row in the column of the facial image data 800b, bitmap data of a facial image (face image) of a user taken in by the cameral is registered. This bitmap data may be data obtained by eliminating background data around the facial image of the user (a background of the facial image may be set to a solid color like a blue background, a background image may be counterchanged with a still image additionally acquired by a digital camera, or a background image may be counterchanged with a moving picture additionally acquired by a digital video camera).
In each row in the column of the facial feature point 800c, each code that recognizes a feature point of each face of one or more users included in a registered facial image is written.
A unique ID number incremented in order of registration is written in each row in the column of the identifier 800a. This unique ID is assigned to each facial image in the same row.
In each row in the column of the identifier link 800d, one or more ID numbers assigned to faces of the same user can be written.
In each row in the column of the mail reception folder 800e, a folder name associated with the identifier 800a in the same row can be given. Specifically, a folder name “mail_receive_folder#1” is given to a folder belonging to the mail reception folder 800e in a row where the identifier 800a has an ID=1, a folder name “mail_receive_folder#2” is given to a folder in a row having an ID=2, a folder name “mail_receive_folder#3” is given to a folder in a row having an ID=3, and a folder name “mail_receive_folder#4” is given to a folder in a row having an ID=4.
Likewise, a folder name associated with an ID is written in each row in the column of the mail reception folder 800e.
In the mail reception folder 800e, not only video contents (a video mail) themselves are registered but also a shortcut (a pointer) to contents, an URL (a description that specifies a location where information is present on the Internet and a method of acquiring this information), and others can be written.
A folder in the mail reception folder 800e has a function like an electronic mail office box. For example, when a video mail from a grandchild is registered in a folder having a name “mail_receive_folder#1”, a user having an ID=1 (“grandpa” and “grandma”) is informed of reception of this mail. Additionally, when links to users having the same faces (“grandpa” having an ID=3 and “grandma” having an ID=4) are provided in the row of the ID=1, “grandpa” having the ID=3 and “grandma” having the ID=4 are individually informed of reception of the mail. “Grandpa” and/or “grandma” who have received this information can see the video mail from the grandchild registered in the folder having the name “mail_receive_folder#1”.
In the mail system using the address management table 800 having the above-described structure, a facial image corresponds to an address that can be seen by a user, and an ID corresponds to an address in a device. When one facial image includes faces of two or more users, addresses of the individual faces are also registered (feature points of the individual faces are separately registered). Then, mails for a plurality of faces (e.g., “grandpa” and “grandma”) are put into a folder of a user himself/herself alone (e.g., “grandpa”), thereby facilitating confirmation of mails.
For example, when “grandpa” registered in the address management table 800 with the ID=3 sits in front of the camera of the TV 10 that is in a standby mode (BLOCK [10]), the face authentication/collation processing apparatus 100 included in the TV 10 starts face authentication processing of “grandpa”. Further, a code associated with a feature point of a face obtained by this face authentication processing is sought from a code group of the feature point 800c in the address management table 800. If a corresponding code is found, it can be understood that “grandpa” registered with the ID=3 is in front of the TV 10. If a corresponding code cannot be found, the standby mode continues until another registered user (e.g., grandma having the ID=4) appears in front of the camera or a power supply of the TV 10 is turned on.
When it has been revealed that “grandpa” registered with the ID=3 is present, the ID=3 is linked with the ID=1 in the identifier link 800d in the address management table 800, and a video mail from grandchildren (e.g., “Kousuke” and “Ayana”) is registered in the video mail reception folder “mail_receive_folder#1” having the ID=1, and a ring tone “you have a mail from Kousuke and Ayana” is added to this video mail. In this case, an audio guidance message “you have a mail from Kousuke and Ayana” is output from a speaker of the TV 10 (BLOCK [13]), and the power supply of the TV 10 is turned on (BLOCK [14]).
In a screen of the TV 10 which has been turned on, a facial image of “grandpa and grandma” registered with the ID=1 in the address management table 800 and a facial image of “grandpa” registered with the ID=3 are displayed. At this time, a guidance text “Which addressee's mail do you want to see?” is also displayed by default (BLOCK [20]). “Grandpa” who is sitting in front of the TV 10 points a cursor on the facial image of “grandpa and grandma” by manipulating a non-illustrated cursor key in a remote controller which is in his hand, and presses a non-illustrated Enter key (a decision key) in this remote controller to decide the selection of “grandpa and grandma” (BLOCK [20]). Then, in the column of the video mail reception folder name 800e in the address management table 800, thumbnails of video mails from grandchildren (“Kousuke” and “Ayana”) registered in the folder “mail_receive_folder#1” in the row having the ID=1 are displayed in the TV 10.
At this time, when another video mail is present in the folder “mail_receive_folder#1”, a thumbnail of this video mail is displayed at the same time, and a guidance text “please select mail” is also displayed by default (BLOCK [26]). At this time, since there was the audio guidance message “you have mail from Kousuke and Ayana” in BLOCK [13], “grandpa” can select a thumbnail including faces of the grandchildren (“Kousuke” and “Ayana”) without question. That is, “grandpa” points the cursor on the facial image of “Kousuke and Ayana” by operating the cursor key in the remote controller which is in his hand and presses the Enter key in the same remote controller to decide the selection of “Kousuke and Ayana” (BLOCK [26]). Then, the video mail from “Kousuke and Ayana” registered in the folder “mail_receive_folder#1” can be reproduced in the TV 10 (BLOCK [28]).
When reproduction of the video mail is finished, the current screen is changed to an operation menu screen (BLOCK [30]). When a user again wants to see the same video mail or wants to see a different mail in the same folder, “(A) See mail” is selected from a displayed menu by manipulating the cursor key in the remote controller, and the Enter key is pressed. Then, the respective pieces of processing from BLOCKs [20] to [28] are again executed. In these pieces of processing, a user can see a different mail registered in, e.g., the folder “mail_receive_folder#3” in the row having the ID=3.
In the above-described operation, “grandpa” can see contents of the mail by operating the cursor key and the Enter key in the remote controller while confirming a facial image he knows, and a keyboard operation is no longer required.
In the operation menu screen (BLOCK [30]), when “(B) Transmit mail” is selected by operating the cursor key in the remote controller and the Enter key is pressed, respective pieces of processing from BLOCKs [40] to [49] are executed. That is, when the grandchildren as senders of a mail (“Kousuke” and “Ayana”) record a mail for “grandpa and grandma” in front of the camera and recording is finished, data of an image and audio acquired in this recording becomes one video mail file (BLOCK [40]). It should be noted that contents of this video mail file may be a still image acquired by a digital camera or a still image taken in by a scanner.
When this video mail file (contents of the mail) is brought to completion, thumbnails of faces registered in each column of the facial image 800b in the address management table 800 depicted in
In the above-described operation, “Kousuke” and “Ayana” can transmit the video mail by just operating the cursor key and the Enter key in the remote controller while confirming the facial images of “grandpa” and/or “grandma” that they know, and a keyboard operation is no longer required.
In the operation menu screen (BLOCK [30]), when “(C) Register mail address” is selected by operating the cursor key in the remote controller and the Enter key is pressed, respective pieces of processing from BLOCKs [60] to [64] are executed. That is, for example, two persons, i.e., “grandpa” and “grandma” are in front of the camera, and face recognition of both these persons is executed by the face authentication/collation processing apparatus 100. Then, the facial image of the two persons, i.e., “grandpa and grandma” is displayed in the screen of the TV 10, and a character guidance “It is registered when you press decision key” is displayed by default (BLOCK [60]). When “grandpa” or “grandma” presses the Enter key (the decision key) in the non-illustrated remote controller, a character guidance “These two persons are going to be registered. Is it OK?” is displayed for confirmation (BLOCK [62]). If there is no problem, the Enter key (the decision key) is again pressed. Then, facial bitmap data of “grandpa and grandma” is registered in the row having the ID=1 in the column of the facial image 800b depicted in
In the above-described operation, “grandpa” or “grandma” can register himself/herself by operating the cursor key and the Enter key in the remote controller while confirming the facial image that he/she knows, and a keyboard operation is no longer required.
A series of processing in BLOCKs [10] to [20] in
01) When a person who has received a mail (a user subjected to face recognition) is watching a TV screen, a TV power supply is automatically turned on, and a message associated with the incoming mail is displayed. This message is not restricted to an example in BLOCK [20] in
02) Before or after turning on the power supply, audio guidance associated with a person who is watching the TV screen (the user subjected to face recognition) is executed.
03) Specifically, a mail sender registers a calling voice in a video mail file of his/her face, and a call, e.g., “you have a message from Taro” is output in this voice when a mail recipient is in front of the TV.
04) At this time, when a plurality of persons are watching the TV screen (a plurality of persons are subjected to face authentication), an object person of the call can be selected. (For example, even if “grandpa” and “grandma” are in front of the TV, an object person of the call may be set to “grandpa” only. This can be realized by registering audio data “a calling voice for grandpa” in the mail reception folder having the ID=1 in the address management table 800 depicted in
05) When seeing contents of a message of a transmitted mail, a previously registered operation (e.g., movement of a hand or fingering of a user who has registered his/her facial image) can be utilized in place of a password. (For example, when a picture outline of a V-like shape formed by an index finger and a middle finger of the user reciprocates in the lateral direction for three times in one or two seconds, this motion is accepted as a password.) In this case, when a different person who should be prevented from seeing a message is also in front of the TV besides users subjected to face authentication, setting a password is helpful. When, e.g., a motion of a hand or fingers of a user is registered as a password, keyboard input is not required for setting the password.
06) It is possible to configure that contents of a message can be seen only when a person who is watching the screen has a wireless key. (In this example, the wireless key means a key having the same function as a keyless entry system of a car. A call using a mobile phone may be utilized in place of this key.) Alternatively, an operation of seeing contents of a message may be accepted only from a remote controller which is in a hand of a user subjected to face recognition and has a specific identification code set thereto (an operation from any other remote controller is not accepted).
07) As a list of received mails, faces extracted from received video mail may be aligned. When faces cannot be extracted from video mails, thumbnails of facial images of users subjected to face authentication can be output like the example of BLOCK [20] in
It should be noted that, when casting aside a possibility that strangers may see mails in a situation which is 100% unreliable based on face authentication using machine processing alone is desirable, a setting that another key is provided and a message or the like of a mail can be seen only in a situation where a this key is present can be formed in addition to the face authentication processing.
Specifically, a mail may be reproduced only when a user who is in front of the TV 10 has a wireless key (which is like a keyless entry system of a car, and a mobile phone or a remote controller having a special code set therein may be used in place of this wireless key) and a specific key code is accepted from this wireless key. In this case, when a receiver (not shown) of the wireless key is connected with the control module 120 depicted in
Alternatively, a previously registered operation (e.g., a motion of a hand of a user) can be utilized as another key in place of the wireless key (or with the wireless key). For example, when a picture of a hand of a user holding the wireless key or the remote controller reciprocates (in an up-and-down direction, a left-and-right direction, or an oblique direction) for a predetermined number of times within a fixed time, this can be accepted as key match.
A user who is in front of the camera 60a may be changed to a different person, another person may appear to increase the number of users, or some of a plurality of persons who are in front of the camera 60a may leave the camera to decrease the number of users. Therefore, even if face authentication is performed previously (e.g., BLOCK [100] in
When face authentication of a user (e.g., “grandpa”) who is in front of the camera 60a is finished and its feature point data is obtained, a facial image registered in the address management table 800 in
Likewise, in the next check, if a facial image having a matched feature point (“grandpa” having the ID=3 which matches with the feature point of “grandpa”) is found (BLOCK [19—YES]), a thumbnail of this facial image (a facial image having the ID=3 in the example depicted in
When the face retrieval BLOCK [18] and the registered face display BLOCK [20] are completed with respect to all the facial images 800b registered in the address management table 800 in
This video mail is partially reproduced (for example, a part corresponding to a period from 20 seconds to 30 seconds after start of reproduction of a video is reproduced at a rate of, e.g., 10 frames per second), and this reproduced video frame is searched for a facial image of a person. If a facial image can be found, this facial image data is taken out, and authentication processing is executed with respect to this facial image in BLOCK [24] to check whether this facial image matches with a feature point of any face registered in the address management table 800 in
When the number of the extracted face is one, a video mail file associated with this face is taken out from the address management table 800 to be reproduced (BLOCK [28]). When, a plurality of faces extracted, these respective faces are displayed in the TV, and a user (a mail recipient) is prompted to select one from the displayed faces (BLOCK [26a]). When the user specifies a specific face from the plurality of displayed faces by operating the cursor in the remote controller and presses the Enter key to select this face, a video mail file associated with this face is taken out from the address management table 800 to be reproduced (BLOCK [28]).
When any face cannot be selected from the video mail at all by the above-described operation (BLOCK [25—NO]), a thumbnail of a registered facial image having an ID associated with a file of this video mail (e.g., the facial image of “grandpa and grandma” having the ID=1 registered in the address management table 800 in
When one thumbnail alone is displayed, a video mail file associated with a face in this thumbnail is taken out from the address management table 800 to be displayed (BLOCK [28]). When there are a plurality of displayed thumbnails, they are displayed in the TV, and a user (a mail recipient) is prompted to select one from the displayed thumbnail (BLOCK [26b]). When the user specifies a specific thumbnail in the plurality of displayed thumbnails by operating the cursor in the remote controller and presses the Enter key to select this thumbnail, and a video mail file associated with a face in this thumbnail is taken out from the address management table 800 to be reproduced (BLOCK [28]).
It should be noted that, when facial image data is obtained by processing in BLOCK [24] in advance, a thumbnail to be displayed in BLOCK [26b] can be created from this facial image data.
If a link to an ID likewise indicative of the same individual face is provided to a row of this ID=1, an ID of this link is acquired (BLOCK [47]). For example, assuming that “grandpa and grandma” is registered with the ID=1, the same “grandpa” is registered with the ID=3, and the same “grandma” is registered with the ID=4, the ID=3 and the ID=4 are written in the row of the ID=1 in the column of the link 800d to the same individual face ID. In this case, when the ID=1 is acquired in BLOCK [46], the ID=3 and the ID=4 are obtained in BLOCK [47].
A shortcut to the registered video mail is provided to folders (“mail_receive_folder#3 and “mail_receive_folder#4”) associated with the thus obtained one or more IDs (the ID=3 and the ID=4 linked to the ID=1), (BLOCK [48]). This shortcut can be realized by setting to a corresponding folder a file pointer indicative of a file storage position of a video mail which should be transmitted or writing a URL which specifies a location where a file of the video mail is present on the Internet and an acquisition method for this file. Further, a mail is transmitted to recipients, i.e., “grandpa and grandma” having the ID=1, “grandpa” having the ID=3, and “grandma” having the ID=4 (BLOCK [49]).
Based on the above-described processing, a user (a mail transmitter) can transmit a mail without operating a keyboard.
When the feature point of the recognized face of “grandpa” coincides with a feature point of a face of “grandpa” in the faces of registered “grandpa and grandma” having the ID=1 (BLOCK [57—YES]), a facial image of registered “grandpa and grandma” having the ID=1 is displayed in the TV 10, and an inquiry “(do you belong to) the same group as this?” is output in the form of text and/or audio (BLOCK [58]). Here, a judgment can be made by the user without leaving the operation to the machine, thereby enhancing an accuracy of face registration.
When a user in front of the TV 10 (grandpa) selects a pictorial symbol of non-illustrated “YES” by operating the remote controller which is in his hand and presses the Enter key (BLOCK [58—YES]), an inquiry “do you want to carry out additional registration?” is output in the form of text and/or audio (BLOCK [62]). When the user (grandpa) determines that “additional registration is not required” (BLOCK [62—NO]), a non-illustrated End key in the remote controller in his hand is pressed to terminate the processing in
When the user (grandpa) determines that “additional registration is carried out” (BLOCK [62—YES]), “grandpa” subjected to face recognition is registered with the ID=3 in the address management table 800 in
It should be noted that a background image except a face is discriminated from a facial image when performing facial authentication in BLOCK [100], and hence the background image discriminated from the face can be cut at the time of registration of the facial image in BLOCK [64] (the background may be left in case of a face acquired outdoors having a good landscape, but cutting the background is desirable in case of a face acquired in a messy room). The cut background image region part can be filed with, e.g., pixel data of a blue color alone (in this case, the background is a solid blue background).
There is also a method of inserting an arbitrary image in a background image region to obtain facial image data. As this method, there is, e.g., a method of “extracting a facial image alone and combining the extracted facial image with a still picture additionally acquired by a digital camera or combining the same with a moving picture additionally acquired by a digital video camera”.
Specifically, for example, when a facial image is combined with bitmap data of an 8-bit plane to be brought to completion, a one-bit plane in the lowest layer is counterchanged with still picture data, and a 7-bit plane on a higher layer can be determined as facial image data of a user.
Alternatively, processing of counterchanging a background image with a moving picture additionally acquired by a digital video camera (or a computer graphic animation created by a personal computer) can be also executed. Specifically, when a facial image is formed of bitmap data of an 8-bit plane, a 1 to 4-bit plane in the lowest layer can be counterchanged with moving picture data, and a 5 to 7-bit plane in the higher layer can be determined as facial image data of a user.
When a feature point of a recognized face of “grandpa” does not coincide with a feature point of a face of “grandpa” in faces of registered “grandpa and grandma” having the ID=1 (BLOCK [57—NO]), or when a user himself/herself determines that they are strangers (e.g., brothers who resemble each other) by taking a closer look even though the machine recognizes that the feature points match with each other (BLOCK [58—NO]), this “grandpa” is a user who is not registered in the address management table 800, and hence a registering operation in BLOCK [64] is executed. When there is only one user to be registered, face registration is all terminated BLOCK [65—YES].
On the other hand, when a plurality of faces are included when face authentication is performed in BLOCK [100], the respective faces are registered one by one. That is, when all of the respective faces are not registered yet (BLOCK [65—NO]), whether each face is a face registered in the address management table 800 is checked (BLOCK [66]). When each face is not registered in the address management table 800 (BLOCK [66—NO]), an image, a feature point, a link ID, and a folder name of this face are registered in a row of a new incremented ID (BLOCK [68]).
When a user determines “strangers who resemble each other” by taking a closer look at a face image (BLOCK [67—NO]) even though the machine determines that this face is registered in the address management table 800 (BLOCK [66—YES]), an image, a feature point, a link ID, and a folder name of this face are registered in a row of a new incremented ID (BLOCK [68]). Here, the user can likewise make a judgment without leaving the operation to the machine, thereby improving an accuracy of face registration. It should be noted that a background can be appropriately cut in face registration in BLOCK [68].
When the machine determines that this face is registered in the address management table 800 (BLOCK [66 —YES]) and the user looks this facial image to determine that “this facial image is mine” (BLOCK [67—YES]), the processing in BLOCK [68] is skipped to avoid redundant registration. When there is no user whose face should be registered, face registration is all terminated (BLOCK [65—YES]).
Although not shown, the main control module 112 is formed of an MPU having, e.g., a program memory or a work memory, and the MPU executes a program (firmware) stored in the program memory to realize various functions.
The image interface unit 113 is an interface that is used to input an image (bitmap data having a resolution associated with the number of pixels of a frequently utilized camera) acquired by the camera 60a (or 20a, 30a, 40a and 50a) in
The face collation module 116 executes collation processing of a facial image detected by the face detection module 115 and a facial image of a specific person stored in the collation data base (a data base including information corresponding to the ID 800a, the facial image 800b, and the facial feature point 800c in
The face detection module 115 detects a facial image from a person from an image acquired by the camera 60a based on detection processing. At this time, the face detection module 115 appropriately changes a gray scale or contrast of acquired image data and executes, e.g., edge enhancement processing to sharply define edges (contours of face, contours of eyes, nose and mouth, and other features) of a facial image that appears in the image data, thereby facilitating detection of a pattern of the face including eyes, nose, mouth, and other features. As a result, a boundary between a region where the facial image is present and a peripheral region where a background image is present becomes clear, thus facilitating processing of deleting a background image portion.
Additionally, in, e.g., a two-dimensional coordinate value having an upper left corner of rectangular region including a facial image as an original point, a two-dimensional region FA including the facial image is determined. In this two-dimensional region FA of the facial image, the face detection module 115 obtains a correlation value of a facial pattern in the region FA and standard facial patterns prepared in advance while moving a face detection template (one or more standard facial patterns prepared in advance). Then, a correlation value of a feature point (e.g., eyes, nose, or mouth) of the facial pattern at a specific position (x, y) represented in the two-dimensional coordinate with respect to the standard pattern is encoded. A result of this encoding is written in the facial feature point 800c in the address management table 800 in
Now, it is assumed that a plurality of standard facial patterns of different sizes which are prepared in advance are gs(i, j) and a correlation of each standard facial pattern gs(i, j) and a local region (a region of, e.g., eyes, nose, or mouth) in an image f(x, y) in the two-dimensional region FA is Ms(x, y). In this case, for example, the correlation Ms(x, y) in the local region can be obtained as follows:
Ms(x,y)=Σf(x+i,y+i)gs(i,j/{Σf(x+i,y+i)·Σgs(i,j)} (1)
Here, the standard facial patterns gs(i, j) of different sizes can be created by, e.g., averaging previously collected facial patterns of a predetermined size.
As a method of obtaining the correlation Ms, a known method can be adopted. As an example, there is “a verification medium issuing apparatus and an identity collating apparatus” disclosed in Japanese Patent Application Publication (KOKAI) 2003-108935 (KOKAI) for reference (paragraphs 0043 to 0057 in this publication disclose a specific example that can be utilized in face authentication/collation processing).
The control module 120 of the video equipment includes a display 121 that reproduces a facial image or a video mail, i.e., displays a picture and character information, a speaker 122 that outputs audio guidance for mail arrival or audio output in a video mail, a voice synthetic module 123 that has a data memory for voice guidance output from the speaker 122 and combines an audio signal from audio data in this memory, a video memory module 124 in which thumbnails, dialogs, or pictorial symbols that are displayed in the display 121 are spread, a remote controller 125 that accepts user operations, a hard disk drive (HDD) 126 that stores video letters and any other information, a memory (an IC card) 127 that stores video letters and any other information, and others. This control module 120 instructs the face authentication/collation processing apparatus 100 to execute face authentication/collation processing, and sequentially carries out the “user operation”, the “arrival notification processing”, the “mail folder choice display processing”, the “received mail list display processing”, the “mail transmission processing”, and the “mail registration processing” explained in conjunction with
It should be noted that functions of the apparatus 100 and the control module 120 in
<[a]> Mail Address (=face) Registration Method 1a:
1a] A face acquired by a camera disposed to a device is registered as an address (a destination of registration is a device main body (e.g., an HDD) or a server on the Internet);
2a] When a plurality of faces are acquired by the camera at a time in address registration, individual faces are recognized, and the individual faces are registered as address data, respectively. However, as an address choice, an image including a plurality of faces may be used;
3a] When an image including a plurality of persons is acquired, it becomes an address for concurrent transmission for multiple persons; and
4a] When registering a facial image, an arbitrary background is used (a messy room as a background can be hidden).
<[b]> Mail Transmission Method (a face of a mail recipient is selected):
1b] A plurality of arbitrary faces are selected as addresses of video letters from a plurality of faces, and the selected faces are registered;
2b] A domestic device acquires an address list of video letters from a server on the Internet (when the address list is used by family members);
3b] A picture taken by a camera disposed to a device as well as other contents (e.g., a picture taken by a digital camera) can be added to each video letter; and
4b] Video letters are recorded in a device (e.g., an HDD) or a server on the Internet.
<[c]> Mail Arrival Notification Method (Face Recognition):
1c] When a person having an incoming mail is watching a screen (face recognition), a power supply is automatically turned on to display a corresponding message (The message means discovery of recommended contents, a video letter, or information from a site);
2c] When the power supply is turned on, a corresponding message is displayed for a person who is watching the screen (in face recognition);
3c] A transmitter registers a calling voice with respect to his/her face, and a mail recipient is called in this voice when he/she is in front of a TV (e.g., “you have a message from Taro”);
4c] At this time, if a plurality of persons are watching the screen, an object person can be selected (Alternatively, contents that can be comprehensively recommended, a relevant video letter, or information from a site are displayed for the persons who are watching the screen);
5c] When seeing contents of a message, an operation registered in advance (e.g., a motion of a hand) is used in place of a password;
6c] Contents of a message can be seen only when a person is watching the screen and has a wireless key (The wireless key: a key like a keyless entry system of a car or a mobile phone can be used in place of the wireless key); and
7c] Faces extracted from video mails are aligned as a list of received mails (When faces cannot be extracted, thumbnails are output).
The following effect can be provided by a camera disposed to, e.g., a TV, a video mail system, and face recognition.
11] A person who cannot use a keyboard can receive the full benefit of mails;
12] Mails can be seen or transmitted through not only a mobile phone or a PC but also a TV and others (Any environment can be adopted);
13] Mails can be transmitted easily like making phone calls (Both advantages of phones and advantages of mails can be provided); and
14] A plurality of persons, e.g., family members can see or transmit mails.
A camera and face recognition means can provide the following effects.
21] Even aged persons who are bad at typing can transmit a mail using a video to grandchildren (although mails have merits that cannot be observed in phones, aged persons have difficulty in using PCs), and even grandchildren or disabled persons who cannot use PCs can transmit mails using videos;
22] A message using a video can be readily left even in case of an urgent need;
23] Text does not have to be typed when registering an address (a face serves as an address); and
24] A feeling of security is provided since a face serves as an address, and address errors can be reduced because of high legibility.
31] Even aged persons who are bad at typing can transmit a mail using a video to grandchildren (although mails have merits that cannot be observed in phones, aged persons have difficulty in using PCs), and Even grandchildren or disabled persons who cannot use PCs can transmit mails using videos;
32] A message using a video can be readily left even in case of an urgent need;
33] Text does not have to be typed when registering an address (a face serves as an address);
34] A feeling of security is provided since a face serves as an address, and address errors can be reduced because of high legibility; and
35] A feature point of a face serves as a mail address (=a user himself/herself) and also as a key for a mail reader (a user himself/herself alone can read (a wireless key or an operation password can complement security as required)).
41] Since a face recognition ability of a person (an adult with considerable experience of life in particular) usually overtakes a face recognition capability based on mechanical processing alone, a mail can be prevented from being transmitted to a wrong person (a stranger unconnected with the intended mail recipient) except a case where each of twins who resemble each other is taken for the other (a problem like erroneous transmission to an unconnected stranger hardly occurs in case of a mistake of twins);
42] Users who have difficulty in operating a keyboard (aged persons who are bad at operating the latest electronic devices in particular) can receive the full benefit of the face recognition processing as long as they can perform simple operations like those of a remote controller; and
43] Further, since face registration can be carried out in actual use without operating a keyboard, convenience of the face recognition processing can be utilized for mail communication even if “a face and a mail address of a person” are not registered in advance.
<I> There is provided an information processing method including: taking facial image data of a user (BLOCK [56]); extracting a feature point of a face of the user from the facial image data and encoding the feature point (BLOCK [100]); and associating the facial image data (800b in
<II> There is provided the information processing method, wherein the address management table (800 in
<III> There is provided the information processing method, wherein the unique identifier (800a: e.g., the ID=1 or the ID=3), the encoded feature point (800c: e.g., “4, 242, C25, D . . . 8, 512, T23, E . . . ” or “4, 242, C25, D . . . ”), and/or the identifier (e.g., the ID=3 and the ID=4 referred in the row having the ID=1) which is referred in the identifier link (800d) is used as a data code (corresponding to a mail address) indicative of an address for a user (e.g., “grandpa and grandma” in case of a plurality of persons or “grandpa” in case of a single person) associated with the facial image data (800b).
<IV> There is provided the information processing method, wherein the address management table (800 in
<V> There is provided the information processing method, wherein a facial image (800b in
<VI> There is provided the information processing method, wherein, when an incoming mail is present, if a user is in front of a camera (BLOCK [11—YES] in
<VII> There is provided an information display apparatus comprising: means (60a, 113) for taking facial image data of a user; means (115) for extracting a feature point of a face of the user from the facial image data and encoding the feature point; and a display (121) which displays a picture and/or an image of a character including an image of the face of the user, wherein the facial image data (“grandpa and grandma”) is supplied to an address management table (800 in the video letter server 80) to be stored in association with a unique identifier (the ID=1) and/or the encoded feature point (800c).
<VIII> There is provided the information display apparatus, further comprising: storage means (the HDD 126 and/or the IC card 127) for storing a copy of contents in the address management table (800 in the video letter server 80).
As explained above, a user can directly select a registered face as an address of a mail by applying the embodiment according to the present invention (the machine does not recognize and select a face, but a user visually confirms and selects a face). Therefore, a mail can be prevented from being transmitted to an incorrect person (a stranger unconnected with the intended mail recipient). Furthermore, an address does not have to be input by operating a keyboard. Further, in an information processing method according to an embodiment of the present invention, facial image data of a user is taken in, and a feature point of a face of the user is extracted from this facial image data to be encoded. Furthermore, a unique identifier and/or the encoded feature point are associated with the facial image data to be registered in an address management table. Moreover, image data of faces are registered in the address management table, and a user can directly select a registered face as an address of a mail (a machine does not recognize a face to be selected, but the user can see a face to be selected). Therefore, a mail can be prevented from being transmitted to a wrong person (a stranger unconnected with a person as a target of transmission). Still over, an address does not have to be input by operating a keyboard.
While certain embodiments of the inventions have been described, these embodiments have been presented by way of example only, and are not intended to limit the scope of the inventions. Indeed, the novel methods and systems described herein may be embodied in a variety of other forms; furthermore, various omissions, substitutions and changes in the form of the methods and systems described herein may be made without departing from the spirit of the inventions. The accompanying claims and their equivalents are intended to cover such forms or modifications as would fall within the scope and spirit of the inventions.
Number | Date | Country | Kind |
---|---|---|---|
2008-311465 | Dec 2008 | JP | national |