JP2001524776A - Speech recognition unit for audience measurement system - Google Patents

Abstract

(57) Summary A viewer member of a program to which a receiver is tuned is identified by a system comprising first and second microphones, a noise cancellation unit, and a speech recognition unit. One of the first and second microphones captures a relatively large amount of audio signals from a member of the viewer, and the other of the first and second microphones captures a relatively large amount of acoustic signals from the receiver. The noise cancellation unit substantially removes the audio signal from the audio signal. The speech recognition unit identifies the member from the audio signal of the viewer member. The first and second microphones, the noise cancellation unit, and the speech recognition unit may be housed in a handheld or tabletop housing. The system may also have a passive human meter that passively identifies members and a program for identifying programs or channels to which the receiver is tuned.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

TECHNICAL FIELD OF THE INVENTION

The present invention relates generally to audience measurement systems, and more particularly, to recognize members of a viewer of a program being tuned by a receiver, such as a television, radio or computer, and / or to a television, radio or The present invention relates to a viewer measurement system for recognizing a program tuned by a receiver such as a computer.

[0002]

[Prior art]

Measuring viewers of a program has long been of interest to program suppliers and advertisers, as measuring viewers can evaluate the effectiveness of the program and advertisements. Many known methods have been employed to provide an overall audience rating for a program, portion of a program, and / or commercial. In addition to all viewers, these methods add statistically significant viewer segments (i.e., the number of men and / or women during a particular age watching a selected portion of a given program). Provide a positive evaluation. These methods measure manually and / or automatically the viewing and / or listening habits of members of a statistically selected home audience, and then, based on these measurements, the total audience Including performing an evaluation. Statistically selected home audience members are usually referred to as panelists.

[0003] Measuring a viewer's viewing and / or listening habits generally requires three separate decisions: (1) A receiver that is monitored in a statistically selected home. Determining which channels or stations are tuned; (2) determining which programs are available on the tuned channels at a time while the receiver is tuned to these channels or stations; and 3) To determine the home members and visitors present in the area of the receiver while the receiver is tuned to the selected program.

[0004] All three decisions can affect the accuracy of audience measurement. Of the three decisions discussed above, decision (3) most probably affects the accuracy of the audience measurement, since the receiver audience members typically require cooperation in the identification process.
When members are required to be involved in the identification process, they often generate errors, or simply forget to identify either themselves or other members of the viewer. For example, one system that allows audience members to identify themselves to indicate their presence is the "People Meter". This "human meter" is a remote control and / or multiple push buttons that are pressed by viewer members as they enter and exit the monitoring area. However, audience members often forget to press the button, or they sometimes mistakenly press the correct button.

[0005] In an attempt to solve this problem, several measurement systems have been developed. For example, in U.S. Pat.No. 3,056,135, Currey et al. Periodically record switches and switches configured to count the number of people entering, leaving, and remaining in a monitored area, and the composition of the viewer. And a system for strategically using a photo recorder. This approach requires a photo recorder that is monitored by the operator, which violates the privacy of the audience members and imposes unacceptable costs on the operation of the measurement.

[0006] These problems have led to many passive, non-compulsive methods of counting (not discriminating) viewers and tracking their movement about the area being monitored. For example, Kiewit et al., In U.S. Pat. No. 4,644,509, disclose an ultrasound detection system that counts and tracks audience members. However, the described system does not identify viewer members, and thus requires viewer members to manually identify themselves.

Other systems, such as those described by Kiewit et al. In US Pat. No. 4,930,011, are worn or carried by members of the audience and uniquely identify corresponding ones of these audience members. It relies on electronically active tags to identify. These systems respond to a tagged audience member whenever it is within the monitored area of the receiver. However, audience members often forget to wear their tags, in which case such audience members are not identified.

[0008] Lu, in US Pat. Nos. 4,858,000 and 5,031,228, disclose an image recognition system for identifying a viewer member of a receiver. The system includes a video camera for acquiring images of the faces of the audience members, and a computer subsystem for identifying these images by comparing these faces with reference faces stored in a database. Use The system also includes an infrared illuminator that locates and tracks the viewer and provides control of the level of undetectable illumination of the monitored area for the video camera. However, this system is generally expensive and difficult to install.

The system disclosed by Lu et al. In US Pat. No. 5,550,928 also uses image recognition to identify viewer members. The system incorporates an algorithm that detects the body of the audience member and then the head. Once the head is found, a face detection algorithm extracts the facial image representation that is provided to two face recognition routines, a native face recognition routine and a template matching routine. The system can identify viewer members who are moving around the area being monitored and / or not facing the camera. However, this system is also complicated and expensive.

[0010] Also, the determination of (1) and (2) discussed above is a measure of the viewer, since one of these determinations alone is usually not sufficient to identify the program to which the receiver is tuned. Can affect the accuracy of That is, the determination of the channel on which the statistically selected home receiver is tuned is to the determination of the program being tuned without knowing what is available on that tuned channel. Will not lead. For similar reasons, the determination of the programs available on that tuned channel leads to the determination of the program to be tuned without knowing which channel the statistically selected home receiver will tune to. Never. Obtaining information about available programs and entering this information into a database for later adaptation to the tuned channel,
This has led to adverse effects on viewer measurement and data entry errors.

[0011] Thomas et al., In US Pat. No. 5,481,294, address these problems with an audience measurement system that detects and records ancillary codes embedded and identified in programs tuned by receivers. are doing. Therefore, if there is an auxiliary code in the program being watched, only one decision needs to be made. If the supplementary code is not available in the tuned program, the program signature is extracted from the program and compared to the previously extracted reference signature.

The measurement device disclosed in this patent includes a non-interventional sensor such as a microphone near the receiver that detects an auxiliary code in the acoustic output of the receiver being monitored. If a microphone was used for auxiliary code detection, a second microphone is provided to capture background noise. The outputs of the auxiliary code detection microphone and the noise detection microphone are processed to subtract background noise from the signal provided by the auxiliary code detection microphone.

The audience measurement system includes a “human meter” as described above for additional backup.

Turner et al. In US Pat. No. 4,907,079 discloses a system for monitoring a viewer at a receiver. A viewer member provides control commands (such as channel selection) to a microprocessor having a keypad via an infrared unit. The information gathered by the microprocessor is stored and transmitted over a regular telephone line to a central computer. A movement detector connected to the microprocessor detects when a viewer member enters or leaves the monitored area. A character generator connected to the microprocessor allows for conversation with the audience members, for example, via alphanumeric messages displayed on a television screen. An utterance output unit connected to the microprocessor generates an utterance signal for talking to the audience members.

[0015] The Turner et al. System also includes an utterance output unit coupled to a microprocessor so that members of the audience can identify themselves audibly. Audible utterances from audience members can be detected by the microphone,
Amplified, digitized, and provided to a microprocessor. The microprocessor performs a power density spectral analysis on the signal to derive the voice signature of the particular audience member. These signatures are compared to previously stored reference signatures during system initialization to identify the viewer. The microprocessor also normalizes the digital audio data and removes small amplitude samples to minimize background noise. When the system disclosed in the Turner et al patent is activated, an initial display message is placed on the screen requesting the viewer to dictate their name and surname to provide a reference sign. However, speech recognition by itself does not allow a visible audience measurement system. Further, the disclosed signal processing that minimizes the effects of background noise is not sufficient under many conditions.

[0016]

[Problems to be solved by the invention]

The present invention is directed to a system for identifying the audience members of a receiver and solving one or more of the above problems.

[0017]

[Means for Solving the Problems]

According to a first aspect of the present invention, a system for identifying a viewer member of a receiver includes first and second microphones, a noise cancellation unit, and a speech recognition unit. The first and second microphones are such that one of the first and second microphones captures a relatively large amount of audio signals from a member of the viewer and the other of the first and second microphones is a receiver. Is configured to capture a relatively large amount of the acoustic signal from The noise cancellation unit is connected to the first and second microphones, and the noise cancellation unit is configured to substantially remove an acoustic signal from the audio signal. The speech recognition unit is connected to the noise cancellation unit, and the speech recognition unit is configured to identify a speech signal from a member of the viewer.

According to another embodiment of the present invention, a system includes a microphone, a voice recognition unit, and a program sensor. The system identifies the members of the viewer of the program to which the receiver is tuned. The voice recognition unit is connected to the microphone and identifies the member from the voice signal. The program sensor is configured to identify the program.

According to yet another embodiment of the present invention, a system includes a passive human meter, a microphone, a voice recognition unit, and a controller. The system identifies the members of the viewer of the program to which the receiver is tuned. The passive human meter is configured to provide an initial identification of the member by passively identifying the member. The microphone is configured to receive an audio signal from the member. The speech recognition unit is connected to the microphone and is configured to provide a second identification of the member by identifying the member from the audio signal. The controller is configured to select at least one of the first and second identities to identify the member.

According to yet another embodiment of the present invention, the viewer identification system comprises a housing,
It has a microphone supported by the housing and a voice recognition unit supported by the housing and connected to the microphone. The speech recognition unit identifies the member of the viewer from the identifier dictated by the member.

[0021]

BEST MODE FOR CARRYING OUT THE INVENTION

An audience measurement system 10 for determining the audience members to a program to which a monitored receiver is tuned is shown in FIG. The audience measurement system 10
It has a viewer recognition unit 12, which is shown as a handheld unit in FIG. 1, but may be other types of units, such as a tabletop type. A viewer recognition unit 12 may allow viewer members to be identified either passively or actively, and may interact with a receiver 14 during the identification process.

The program sensor 16 is a program (and / or channel) to which the receiver 14 is tuned.
Cooperate with the receiver 14 to determine Program sensor 16 may be any known sensor that determines a program (and / or channel). For example, the program sensor 16 detects an auxiliary code embedded in the signal of the program being tuned by the receiver 14 and specifically identifies the tuned program. Additionally or alternatively, the program sensor 16 extracts a signature signature from the signal of the program to which the receiver 14 is tuned, thereby signifying the signature (and / or) the program to which the receiver 14 is tuned. Channel) is compared to a reference signature to identify the signature. Additionally or alternatively, program sensor 16 may be a device that determines the channel to which receiver 14 is tuned by monitoring the tuner of receiver 14.

The viewer recognition unit 12 and the program sensor 16 communicate using a base unit 18 which is either a site unit or a home unit. As is known in the art, a site unit is provided for each viewer area in the home, and all site units in the home can communicate with the home unit on a wall power line in the home. it can. The home unit then communicates via a public telephone system with a central unit that gathers tuning and viewer information from other homes so that the information can be combined into reports useful to program providers, advertisers, etc. I do. In a home with only one viewing area, the base unit 18 is configured to receive data from the viewer recognition unit 12 and the program sensor 16 and to transmit this data to the central unit via a public telephone system. ing.

Accordingly, as shown in FIG. 2, the viewer recognition unit 12 may be, for example, infrared or R
A communication link 20, which may be an F communication link, provides the base unit 18 with its recognition data. The program sensor 16 provides the program identification data to the base unit 18 via a communication link 21, which may be, for example, a hard wire, but alternatively, for example, an infrared or RF communication link. Other communication links may be used. The base unit 18 gathers viewer recognition data from the viewer recognition unit 12 and communicates this data over a payphone system 22 to a central site 24 that stores similar data from other homes.

As shown in FIG. 2, the audience measurement system 10 also includes a passive human meter 26.
have. For example, the passive human meter 26 may be of a type that passively identifies the audience member being monitored, in accordance with the techniques of the aforementioned US Pat. No. 5,550,928. The passive human meter 26 is connected to the base unit 1 via a communication link 28.
8 to provide its personal identification data, and this communication link 28 may be, for example, a hard wire, but may alternatively be another communication link, which may be, for example, an infrared or RF communication link.

Thus, if the viewer measurement system 10 has both a viewer recognition unit 12 and a passive human meter 26, the passive human meter 26 will If not, it can be used to passively identify audience members. Alternatively, the viewer recognition unit 12 is used to actively or passively identify viewer members if the passive human meter 26 cannot passively identify viewer members. You. Further alternatively, the output of both the viewer recognition unit 12 and the passive human meter 26 may be simultaneously applied to soft decision logic identifying viewer members based on both of these outputs. In each of these alternatives, the certainty that the audience member is accurately recognized is increasing.

The viewer recognition unit 12 is shown in more detail in FIG. The viewer recognition unit 12 has first and second microphones 30 and 32, and the first and second microphones 30 and 32 are such that one of the first and second microphones 30 and 32 has a relatively large number. Viewer recognition is such that audio signals from the audience members tend to be incorporated, and the other of the first and second microphones 30 and 32 tends to acquire audio signals from a relatively large number of monitored receivers. It is arranged in the unit 12. For example, in most arrangements of the viewer recognition unit 12, the first and second microphones 30 and 32 may have one of the first and second microphones 30 and 32 near a viewer member, and The other of the first and second microphones 30 and 32 are spaced such that they are near the receiver being monitored. First
And the preferred separation of the second microphones 30 and 32 is 6.72 inches, although other suitable separations may be employed. In this embodiment of the invention, the first and second microphones 30 and 32 are located in opposite directions. Alternatively, the first and second microphones 30 and 32 may be closer to the audience member and the first of the first and second microphones 30 and 32 may be closer to the receiver being monitored. And the second microphones 30 and 3 may be arranged such that one or both of the two are remote from the viewer recognition unit 12.

The analog output of the first microphone 30 is amplified by an amplifier 34, and the amplified analog output is converted into a digital signal by an analog-to-digital converter 36. Similarly, the analog output of the second microphone 32 is
8, and the amplified analog output is converted into a digital signal by the analog-to-digital converter 40.

[0029] Even if the first and second microphones 30 and 32 have a tendency for one of the first and second microphones 30 and 32 to incorporate audio signals from a relatively large audience member, Even though the other of the two microphones 30 and 32 tends to pick up acoustic signals from a relatively large number of monitored receivers, the first and second microphones 30 and 32 may be capable of receiving audio signals from the audience members. It takes in both the acoustic signal from the receiver 14. Also, both the first and second microphones 30 and 32 take in background noise other than the acoustic output from the receiver 14. Analog-to-digital converters 36 and 40 are provided in noise cancellation arrangement 42 to reduce or eliminate the effect of background noise and the acoustic output from receiver 14. The noise canceling arrangement 42 converts the audio signal supplied from the audience member from the receiver 1 similarly to the background noise.
The outputs from the analog-to-digital converters 36 and 40 are processed so as to cancel the acoustic output of the fourth, so that these audio signals can be more easily identified.

The noise cancellation arrangement 42 uses signal subtraction to subtract background noise and the acoustic output of the receiver 14 from the audio signal generated by the audience members. For example, if the first microphone 30 is pointed at the audience member and the second microphone 32 is pointed toward the receiver 14, the signal from the first microphone 30 will be S _r (r (Meaning). In general, the signal _Sp is given by the following equation:

[0031]

(Equation 1)

Here, S _p is a signal from the first microphone 30, V _p is a contribution of the audio signal of the viewer member to the signal from the first microphone 30, and V _r is a receiver. First microphone 30 with 14 sound outputs and other background outputs
And a and b are the parameters discussed below, and
Is greater than b.

Similarly, the signal S _r is given by the following equation:

[0034]

(Equation 2)

Here, S _r is a signal from the second microphone 32, c and d are parameters discussed below, and c is smaller than d. Equations (1) and (2) are subtracted according to the following equation using the weight parameter e.

[0036]

(Equation 3)

Where D is the weighted difference between equations (1) and (2) and e is the weight parameter discussed below.

If a, b, c and d are fixed parameters determined by the environment and the characteristics of the first and second microphones 30 and 32, the weight parameter e is (
be-d) = 0 or e = d / b> 1. Therefore, if (b
If e-d) = 0, the difference D would be formed only by the signal V _p. In this case, the signal V _p from equation (3) is given by:

[0039]

(Equation 4)

Since a, e and c are constant in a given household, these constants can be treated as a single constant k according to the following equation:

[0041]

(Equation 5)

Thus, the signal V _p is given by the following equation.

[0043]

(Equation 6)

The weight parameter e is determined adaptively while the receiver 14 is turned on but no one of the viewer members speaks. During such a period, the viewer recognition unit 12 can automatically find the weight parameter e such that D = 0. The weight parameter e is then substituted into Equation (5) and (6) to determine V _p, V _p is processed by the speech recognition unit 44 as discussed below.

This noise cancellation approach is such that the first microphone 30 is directed or closer to the viewer than the second microphone 32, and the second microphone 32 is directed or closer to the receiver 14 than the first microphone 30. Assume that you are. For example, if the viewer recognition unit 12 is a handheld unit, the first microphone 30 and the second microphone 32 are oriented so that they are spaced apart and the user is facing the viewer recognition unit 12, By first
Is mounted on the viewer recognition unit 12 such that the microphone 30 of the second microphone 32 faces the direction of the viewer and the second microphone 32 faces the direction of the receiver 14. Similarly, if the viewer recognition unit 12 is a tabletop unit, the first microphone 30 and the second microphone 32 are oriented so that they are spaced apart and the user is facing the viewer recognition unit 12; Thereby, the first microphone 3
The second microphone 32 is mounted on the viewer recognition unit 12 so that 0 faces the viewer and the second microphone 32 faces the receiver 14. Alternatively, whether the viewer recognition unit 12 is a handheld unit or a tabletop unit, the first and second microphones 30 and 32 are such that the first microphone 30 is mounted proximate to the viewer, Two microphones 32 are connected to the viewer recognition unit 12 by lines or other communication media that allow the microphones 32 to be mounted in close proximity to the receiver 14.

Instead of the noise cancellation discussed above, the noise cancellation arrangement 42 uses phase cancellation. The sound travels through the air at about 341.376 m / sec (1,120 ft / second). Sounds that do not originate equidistant from both the first and second microphones 30 and 32, which are preferably separated by 6.72 inches, reach these microphones at different times and cancel or At least partially offset.
Based on the actual frequency of the sound, the spacing between the first and second microphones 30 and 32 will produce different phase angles. The noise cancellation arrangement 42 that performs noise cancellation
These different phase angles are used to cancel the receiver 14 and other background noise from the audio signal generated by the audience members. In this embodiment of the invention, the first and second microphones 30 and 32 are
Preferably, it is a directional microphone mounted on the same side of the two. The viewer recognition unit 12 can take any arrangement with respect to the viewer and the receiver 14 as long as the first and second microphones 30 and 32 are pointed directly at the receiver 14.

In this embodiment of the present invention, the noise cancellation arrangement 42 may have the shape shown in FIG. According to this embodiment, the noise cancellation arrangement 42 includes an amplifier 60 that amplifies the signal from the first microphone 30 and an amplifier 62 that amplifies the signal from the second microphone 32. Since the amplifiers 60 and 62 are substantially the same,
Only one of the amplifiers 60 and 62 is shown in detail in FIG. Amplifiers 60 and 6
The two outputs are added by adder 64 to substantially eliminate background noise. The output from the adder 64 is detected, and the DC
The level is converted to a level and supplied to the timer 68.

If the output of adder 64 is greater than a predetermined threshold, timer 68 is triggered and switch 7 is activated to connect the output of adder 64 to speech recognition unit 44.
Supply output to 0. Speech recognition unit 44 provides a hold signal via terminal 72 to hold timer 68 in a triggered state until a predetermined time after speech recognition unit 44 last recognized speech. .

The voice recognition unit 44 is provided by Sensory Inc. RS supplied by
It may be a C164 speech recognition unit. The speech recognition unit 12 compares the output from the noise cancellation arrangement 42 with a pre-stored reference speech signal (identifier) stored in a memory. These pre-stored reference speech signals (identifiers) may be speech features from the dictated identifier words or the frequency spectrum of the speaker. If a match is found, the members of the viewer are identified and the speech recognition unit 44 provides an appropriate identification record to the controller 46 for each match. The controller 46 is
An identification record is recorded in the memory 48 for later communication to the base unit 18 by the communication arrangement 50. Communication arrangement 50 may use communication link 20 as described above. The controller 46 also stores the time at which the viewer member was identified in the memory 48.
Connected to a clock 52 for storage in

The prompt unit 54 is connected to the controller 46. The controller 46 supplies audio signals corresponding to the audience members to the first and second microphones 30 and 32, or a keypad 56 for manually identifying them.
The prompt unit 54 is controlled to prompt the audience member to use The prompting unit 54 is determined to provide audible and / or visible prompts to the viewer members, and the voice synthesizer and / or the viewer recognition unit 12.
Or it could be a graphic or video display (such as an LCD or LED). Alternatively, the prompting unit 54 may include a receiver 14 configured to provide audible and / or visible prompts to the viewer members in response to the conversation with the viewer recognition unit 12. It can be a speaker and / or a screen.

The viewer recognition unit 12 can be realized as a handheld unit 90 shown in FIG. The handheld unit 90 has a housing 92 that houses the elements of the viewer recognition unit 12 shown in FIG. Housing 92
Supports an LED display 94 that performs the function of the prompt unit 54. However, as described above, the prompt unit 54 may alternatively be a voice synthesizer, in which case the housing 92 incorporates an output speaker that provides audible prompts from the voice synthesizer.
LED display 94 (and / or a voice synthesizer) may perform functions other than prompting. The housing 92 also houses an IR transmitter 96 for establishing the communication link 20 shown in FIG. Alternatively, communication link 20 may be established by ultrasound, RF link, hard wire, or the like. The housing 92 further supports a keypad 98, which in FIG. 3 is the keypad 56, which allows viewer members to enter data. For example, viewer members and visitors are given one by one of the predetermined keys of keypad 98 so that they can be manually identified by themselves if they cannot be recognized by their audio signals. Is assigned. The housing 92 is configured to power the viewer recognition unit 12 and has a rechargeable battery that can be charged overnight or when the viewer recognition unit 12 is not in use. I have. The handheld unit 90 also has a power on / off button 95 for turning the handheld unit 90 on and off, and an LED 97 for indicating that the viewer recognition unit 12 is ready to identify members and visitors. ing.

Alternatively, the viewer recognition unit 12 may be a desktop unit 1 shown in FIG.
00 can also be implemented. The desktop unit 100 has a housing 102 that houses the elements of the viewer recognition unit 12 shown in FIG. Housing 102 supports an LED display (not shown) that performs the function of prompt unit 54. However, as described above, the prompt unit 54 may alternatively be a voice synthesizer, in which case the housing 102 incorporates an output speaker that provides audible prompts from the voice synthesizer. An LED display or voice synthesizer may perform functions other than prompting. The housing 102 also houses an IR or RF transmitter (not shown) for establishing the communication link 20 shown in FIG. Alternatively, communication link 20 may be established by ultrasound, hardwire, or the like. The housing 102 further supports a keypad 104, which is keypad 56 in FIG. 3, which allows the audience members to enter data. For example, the members of the viewer are assigned one by one of the predetermined keys of the keypad 104 so that they can be manually identified by themselves if they cannot be recognized by their audio signals. . The desktop unit 100 is the viewer recognition unit 1
2 and has a rechargeable battery that can be charged overnight or when not in use. In addition, the desktop unit 100 has a power-on indicator light source 106 and a sound detection indicator light source 108.

During the training mode, the controller 46 of the viewer recognition unit 12 executes the program 120 shown by the flowchart in FIG. The program 120 is executed during a training mode to store identifier words or audio features associated with the members of the viewer. During training, the viewer members are assumed to be located at a certain distance from the viewer recognition unit 12.
The viewer member starts training by pressing a key or button on the keypad 56 of the viewer recognition unit 12.

When the training mode is started so that the program 120 is executed,
Viewer recognition unit 12 determines at block 122 whether a predetermined amount of time has elapsed since the last viewer member activity. If the viewer recognition unit 12 determines that there are no viewer members for a predetermined amount of time, the program 120 assumes that the training mode has ended and the program 120 ends. If the predetermined amount of time has not elapsed since the last viewer member activity, the viewer recognition unit 12 proceeds to block 124 where
Keep waiting for a wakeup command to be dictated. The wake-up command should be the only word or combination of words not normally present in the sound of the receiver 14. The viewer member dictates the wake-up command in a normal voice towards the first and second microphones 30 and 32. When the viewer recognition unit 12 detects the wake-up command, the viewer recognition unit 12 returns to block 126
At, the viewer member is prompted to dictate the word of the identifier selected by the viewer to identify him or her. The word of such an identifier is the name of the viewer.

The viewer recognition unit 12 then determines at block 128 whether the viewer member has dictated the word of the identifier. If not, the viewer recognition unit 12 prompts the viewer member to dictate again the word of the identifier at block 126. If the viewer recognition unit 12 determines at block 128 that the viewer member has dictated the word of the identifier, then at block 130 the viewer recognition unit 12 appropriately processes the word of the identifier and The word of the identifier is stored in the memory 48. Flow returns to block 122, where the training mode is terminated or another viewer member enters the word of the corresponding identifier.

During the identification mode, the controller 46 of the viewer recognition unit 12 executes the program 140 shown as a flowchart in FIG. Program 140
Is executed during the identification mode to identify the members of the viewer. The identification mode is started by pressing a key or button on the keypad 56 of the viewer recognition unit 12.

When the identification mode is started so that the program 140 is executed, the viewer recognition unit 12 proceeds to block 142 where the viewer member determines whether the appropriate key on the keypad 56 of the viewer recognition unit 12 Or, by pressing the button, it is determined whether it is desired to end the identification mode. If the viewer recognition unit 12
Decides that the audience member wants to exit identification mode,
The program 140 ends. If the viewer recognition unit 12 determines that the viewer member does not wish to exit identification mode, the viewer recognition unit 12 waits at block 144 for a wake-up command to be dictated. The viewer member dictates the wake-up command in a normal voice towards the first and second microphones 30 and 32. When the viewer recognition unit 12 detects a wake-up command, the viewer recognition unit 12 prompts the viewer member to dictate the word of his or her identifier at block 146.

The viewer recognition unit 12 then determines at block 148 whether the viewer member has dictated the word of the identifier. If not, the viewer recognition unit 12 prompts the viewer member to dictate again the word of the identifier at block 146. If the viewer recognition unit 12 determines at block 148 that the viewer member has dictated the word of the identifier, the viewer recognition unit 1
2 instructs, at block 150, the speech recognition unit 44 to compare the identifier word with a word vocabulary of identifier words stored in the memory 48.

If the viewer recognition unit 12 detects in block 152 that the speech recognition unit 44 has detected a match between the dictated word of the identifier and one of the identifier words stored in the memory 48. Once determined, the viewer recognition unit 12 then proceeds to block 154 for immediate or subsequent communication to the base unit 18,
The time that matched the dictated identifier word is recorded as an identification record on memory 48. (Alternatively, viewer recognition unit 12 determines in block 154 the dictated identifier word and time. , May be transmitted directly to the base unit 18). If the viewer recognition unit 12 determines, at block 152, that the speech recognition unit 44 does not match the dictated identifier word with one of the identifier words stored in the memory 48, then the viewer Person recognition unit 12,
At block 156, a retry or prompt to use another form of identification (eg, entering the word of the identifier via use of keypad 56). Alternatively or additionally, if the controller 46 determines that the speech recognition unit 44 cannot detect a match at block 152, the controller 46
May send a message to the base unit 18 at block 156 indicating that identification should be performed from the passive human meter 26. After the viewer recognition unit 12 stores in block 154 the time at which the word of the dictated identifier matched the word as an identification record in the memory 48, or the viewer recognition unit 12 proceeds to block 156 After prompting the audience member to retry or use another form of identification, or to prompt the base unit 18 to send an appropriate message to look for identification from the passive human meter 26, flow returns to block 142. .

An advantage of using wake-up commands during normal speech recognition is that the comparison process is simple. That is, during normal viewer recognition, each dictated identifier word is compared to all reference identifier words stored in memory 48. However, the wake-up command need only be compared to one reference in memory 48.

Block 126 of FIG. 7 and block 146 of FIG. 8 may also be configured to prompt each viewer member to dictate a cue in addition to the word of the member's identifier. The advantage of combining the cue with the identifier word is that when the identifier word of a person is included in the acoustic output of the receiver 14, the viewer recognition unit 12 incorrectly identifies the person as a viewer member; But to minimize the possibility that the person is not a viewer. For example, when the word of the identifier is a name, and the signal is common to all viewer members, such as "enter" or "exit". Thus, when a viewer member enters (or exits) the area being monitored, he or she dictates his or her name as a word of the identifier with a signal. For example,
The member is "enter John" or "enter Mary". Alternatively, the signal may be unique to each viewer member. Either configuration reduces the likelihood that words audibly uttered by the receiver 14 will be misinterpreted as identifier words dictated by the audience member.

[0062] Certain modifications of the present invention have been discussed above. Other modifications will occur to those skilled in the art of the invention. For example, the above description has been made in connection with the measurement of viewers of television programs. However, the invention is also useful in measuring viewers of radio programs and other media programs.

In the above description, the program sensor 16 determines whether the program and / or
Or it is described as channel detection. Alternatively, the viewer recognition unit 12 may be configured to accept manual input of identification by a viewer member through use of the keypad 56.

Further, a remote control function can be incorporated in the viewer recognition unit 12.

Further, instead of relying on two microphones, such as the first and second microphones 30 and 32, the viewer recognition unit 12 may rely on a single close range directional microphone. If so, a noise cancellation unit is not needed.

In addition, the use of a wake-up command may be omitted.

Also, as described above, the passive human meter 26 may be of a type that passively identifies viewer members within the area being monitored in accordance with the teachings of US Pat. No. 5,550,928. However, passive human meter 26 may be any of the known passive devices capable of passively identifying a member of a viewer. For example, passive human meter 26 may use ultrasound and / or infrared transducers or an array of such transducers to capture an image of the viewer area, and to recognize members in the viewer. Alternatively, a processing algorithm may be used.

In addition, passive human meter 26 passively identifies audience members in the monitored area, as described above. If the passive human meter 26 cannot passively identify the members of the viewer, it will count the members of the viewer, and will replace the manual identification of the members with the number of manual identifications of the members. And the count is equal. Alternatively, if the passive human meter 26 cannot passively identify all members of the viewer, it will count the members of the viewer, and Can be configured to facilitate manual identification of those members who cannot. Further alternatively, a counter may be used in place of the passive human meter 26. The counter may be an ultrasonic and / or infrared transducer, or an array of such transducers, or other device capable of detecting the number of viewers and counting the number of detected persons or It may have a group of devices. Next, the audience measurement system 10 may be configured to prompt manual identification until the number of manually identified persons equals the count.

Furthermore, the viewer recognition unit 12, the program sensor 16, and the passive human meter 26
Are shown as separate elements in FIG. Alternatively, viewer recognition unit 12 may include program sensor 16 and / or passive human meter 26.

Accordingly, the description of the present invention is illustrative only and is for the purpose of teaching those skilled in the art the best mode of practicing the present invention. The details may be substantially modified without departing from the spirit of the invention, and the exclusive use of all modifications within the scope of the appended claims is assured.

[Brief description of the drawings]

These and other features and advantages of the present invention will become more apparent from a detailed consideration of the invention when taken in conjunction with the following drawings.

FIG. 1 illustrates a system of the present invention used by a television receiver to identify a member of a viewer of a tuned program.

FIG. 2 is a block diagram of an embodiment of a system having an Audience Identification Unit (ARU) according to the present invention.

FIG. 3 is a more detailed block diagram of the viewer identification unit shown in FIG. 2;

FIG. 4 is a diagram of a noise canceling unit used in the viewer identification unit shown in FIG. 2;

FIG. 5 is a hand-held version of the viewer identification unit shown in FIG. 3;

FIG. 6 is a tabletop type of the viewer identification unit shown in FIG. 3;

FIG. 7 is a flowchart showing a program executed by the viewer identification unit shown in FIG. 3 during a training period.

FIG. 8 is a flowchart showing a program executed by the viewer identification unit shown in FIG. 3 during voice identification of a viewer member.

[Procedural Amendment] Submission of translation of Article 34 Amendment of the Patent Cooperation Treaty

[Submission date] January 10, 2000 (2000.1.10)

[Procedure amendment 1]

[Document name to be amended] Statement

[Correction target item name] Claims

[Correction method] Change

[Correction contents]

[Claims]

[Procedure amendment 2]

[Document name to be amended] Statement

[Correction target item name] 0015

[Correction method] Change

[Correction contents]

[0015] The Turner et al. System also includes an utterance output unit coupled to a microprocessor so that members of the audience can identify themselves audibly. Audible utterances from audience members can be detected by the microphone,
Amplified, digitized, and provided to a microprocessor. The microprocessor performs a power density spectral analysis on the signal to derive the voice signature of the particular audience member. These signatures are compared to previously stored reference signatures during system initialization to identify the viewer. The microprocessor also normalizes the digital audio data and removes small amplitude samples to minimize background noise. When the system disclosed in the Turner et al patent is activated, an initial display message is placed on the screen requesting the viewer to dictate their name and surname to provide a reference sign. However, speech recognition by itself does not allow a visible audience measurement system. Further, the disclosed signal processing that minimizes the effects of background noise is not sufficient under many conditions. Kimura in U.S. Pat. No. 5,267,323 discloses a voice operated remote control having a transmitter 10A. The transmitter 10A has first and second microphones M1 and M2. The first microphone M1 functions to input a voice command, and the second microphone M2 functions to capture surrounding sounds and noise. The noise remover 104 removes ambient noise by generating a difference signal representing the difference between the signals from the first and second microphones M1 and M2.

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) H04N 17/00 United States 33741-6413 Florida Pass-A-Grill Beach P. Oh. Box 46413 F-term (reference) 5C061 BB03 BB09 BB13 CC05 5D015 AA02 BB02 EE05 KK02

Claims (72)

[Claims] 1. A system for identifying a viewer member of a receiver, wherein one of the first and second microphones captures a relatively large portion of the audio signal from the viewer member. A first and a second microphone, wherein the other of the second and second microphones is configured to capture a relatively large amount of an acoustic signal from the receiver; and a noise cancellation connected to the first and the second microphones. A unit,
A noise cancellation unit configured to substantially remove the acoustic signal from the audio signal; and a speech recognition unit connected to the noise cancellation unit, wherein the speech recognition unit comprises: A speech recognition unit configured to identify speech signals from members of the viewer. 2. The system of claim 1, further comprising a program sensor configured to identify a program received by the receiver. 3. The system according to claim 2, wherein said program sensor comprises a channel sensor. 4. The system according to claim 2, wherein said program sensor comprises a code reader configured to detect a program identification code in a program to which said receiver is tuned. . 5. The system of claim 2, wherein the program sensor includes a signature extractor configured to extract a uniquely identifiable signature from a program to which the receiver is tuned. System. 6. The system of claim 2, wherein the program sensor has a keypad configured to allow a viewer member to use a key in identifying a program or channel. System. 7. The system of claim 1, further comprising a clock and a memory, wherein the memory stores an identity of a viewer member based on audio signals from the member, and the memory. Is a system that stores the time since the clock when a member is identified. 8. The system according to claim 1, wherein said receiver is a television. 9. The system of claim 1, wherein said receiver is a radio. 10. The system according to claim 1, wherein the receiver is a multimedia receiving device. 11. The system of claim 1, further comprising a keypad configured to receive an identification input from the viewer member. 12. The system of claim 1, further comprising a prompting device configured to prompt a member of the viewer to dictate the audio signal. 13. The system of claim 12, further comprising a sensor that identifies a program or channel being received by the receiver. 14. The system of claim 12, further comprising a passive human meter configured to passively identify the audience member. 15. The system according to claim 14, further comprising a sensor for identifying a program or channel being received by the receiver. 16. The system of claim 1, wherein the audio signal includes an identifier of a member of the viewer. 17. The system of claim 1, wherein the audio signal includes an identifier of a member of the viewer and a cue. 18. The system of claim 1, further comprising a passive human meter configured to passively identify the audience member. 19. The system of claim 1, further comprising a communication unit configured to transmit said member identification and / or program identification to a remote location. 20. The system according to claim 19, further comprising a communication unit configured to transmit the identification of the member to a remote location by an infrared signal. 21. The system according to claim 1, wherein said first and second microphones, said noise canceling unit and said voice recognition unit are housed in a handheld housing. 22. The system according to claim 1, wherein the first and second microphones, the noise canceling unit, and the speech recognition unit are housed in a desktop housing. 23. The system of claim 1, wherein the first and second microphones are oriented in substantially opposite sensing directions. 24. The system of claim 1, wherein the first microphone is mounted closer to a viewer than the second microphone, and wherein the second microphone is closer to the receiver than the first microphone. A system mounted in close proximity. 25. The system according to claim 1, further comprising a counter, a display, and a manual input device, wherein the speech recognition unit identifies the member equal to a count provided by the counter. A prompt on the display to prompt for manual identification through use of the manual input device until the prompt is reached. 26. A system for identifying a member of a viewer of a program to which a receiver is tuned, comprising: a microphone positioned to receive an audio signal from the member; and a microphone connected to the microphone; A system comprising: a voice recognition unit configured to identify a member from a voice signal; and a sensor configured to identify the program or channel. 27. The system of claim 26, further comprising a memory configured to store reference identification information corresponding to the member, wherein the voice recognition unit is configured to store the voice signal and the reference identification information. And a system configured to identify the member based on a comparison with the member. 28. The system of claim 27, wherein the sensor is configured to identify the program based on characteristics of the program. 29. The system of claim 28, wherein the feature is an identification code embedded in the program. 30. The system of claim 28, wherein the feature is a uniquely identifiable signature extracted from the program. 31. The system according to claim 26, wherein said receiver is a television. 32. The system of claim 26, wherein said receiver is a radio. 33. The system according to claim 26, wherein the receiver is a multimedia receiving device. 34. The system of claim 26, further comprising a keypad configured to receive an identification input from the member. 35. The system of claim 34, wherein the input of identification is the identification of the member. 36. The system according to claim 34, wherein the identification input is a program or channel identification. 37. The system of claim 26, further comprising a reminder configured to dictate the audio signal to the member. 38. The system of claim 37, further comprising a passive human meter configured to passively identify said member. 39. The system of claim 26, wherein the audio signal includes an identifier of the member. 40. The system of claim 26, wherein the audio signal includes an identifier of the member and a cue. 41. The system of claim 26, further comprising a passive human meter configured to passively identify said member. 42. The system of claim 26, further comprising a communication unit configured to transmit said member identification to a remote location. 43. The system of claim 42, further comprising a communication unit configured to transmit said member identification to a remote location by an infrared signal. 44. The system of claim 26, wherein the microphone and the voice recognition unit are housed in a handheld housing. 45. The system of claim 26, wherein the microphone and the voice recognition unit are housed in a desktop housing. 46. The system of claim 26, further comprising a counter, a display, and a manual input device, wherein the voice recognition unit identifies the member equal to a count provided by the counter. A prompt on the display to prompt for manual identification through use of the manual input device until the prompt is reached. 47. A system for identifying a member of a viewer of a program to which a receiver is tuned, wherein the system is configured to provide a first identification of the member by passively identifying the member. A passive human meter, a microphone positioned to receive an audio signal from the member, and a microphone connected to the microphone to provide a second identification of the member by identifying the member from the audio signal. And a controller configured to select at least one of the first and second identities to identify the member. 48. The system of claim 47, wherein the controller further comprises a prompting unit configured to dictate the audio signal. 49. The system of claim 48, further comprising a sensor configured to identify the program or channel. 50. The system of claim 47, wherein the controller prompts the member to dictate an audio signal if the passive anthropometric meter is unable to provide the first identification. A system comprising a reminder unit configured to: 51. The system of claim 47, wherein the controller is configured to select the first identification if the speech recognition unit cannot provide the second identification. System. 52. The system of claim 47, wherein said audio signal includes an identifier. 53. The system of claim 47, wherein the audio signal includes an identifier and a cue. 54. The system of claim 47, further comprising a sensor configured to identify the program and channel. 55. The system of claim 47, further comprising a communication unit configured to transmit the member identification to a remote location. 56. The system of claim 55, further comprising a communication unit configured to be able to transmit said member identification to a remote location by infrared signals. 57. The system of claim 47, wherein said microphone and said voice recognition unit are contained in a handheld housing. 58. The system according to claim 47, wherein the microphone and the voice recognition unit are housed in a desktop housing. 59. The system of claim 47, further comprising a counter, a display, and a manual input device, wherein the voice recognition unit identifies the member equal to a count provided by the counter. A prompt on the display to prompt for manual identification through use of the manual input device until the prompt is reached. 60. A viewer recognition unit for identifying a member of a viewer of a program to which a receiver is tuned, comprising: a housing; a microphone supported by the housing; and the microphone supported by the housing. And a speech recognition unit configured to recognize the member from an identifier dictated by the member. 61. The voice recognition unit according to claim 60, further comprising a prompting unit supported by the housing, wherein the prompting unit is configured to prompt the member to take an action. Viewer recognition unit. 62. The voice recognition unit according to claim 61, further comprising a keypad supported by the housing, wherein the keypad comprises:
A viewer recognition unit configured to allow the member to enter data. 63. The speech recognition unit according to claim 62, further comprising a communication unit supported by the housing, wherein the communication unit is configured to transmit data to the base unit. Recognition unit. 64. The speech recognition unit according to claim 63, further comprising a clock and a memory supported by the housing, the memory storing the identification of the member based on the dictated identifier, and The memory is
A viewer recognition unit that stores the time since the clock when the member was identified. 65. The speech recognition unit according to claim 64, wherein the identifier includes a word and a cue. 66. The voice recognition unit according to claim 60, further comprising a keypad supported by the housing, wherein the keypad comprises:
A viewer recognition unit configured to allow the member to enter data. 67. The speech recognition unit of claim 66, further comprising a communication unit supported by the housing, wherein the communication unit is configured to transmit data to the base unit. Recognition unit. 68. The speech recognition unit of claim 60, further comprising a communication unit supported by the housing, wherein the communication unit is configured to transmit data to the base unit. Recognition unit. 69. The speech recognition unit according to claim 60, further comprising a clock and a memory supported by the housing, the memory storing the identification of the member based on the dictated identifier, and The memory is
A viewer recognition unit that stores the time since the clock when the member was identified. 70. The speech recognition unit according to claim 60, wherein said identifier includes a word and a cue. 71. The speech recognition unit according to claim 60, wherein said speech recognition unit is a handheld unit. 72. The speech recognition unit according to claim 60, wherein said speech recognition unit is a desk unit.