JP2005234964A - Fingerprint sensor, biometrics authentication apparatus, method for retrieving fingerprint detection condition, and method for detecting fingerprint - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a fingerprint sensor capable of quickly and efficiently retrieving an optimum fingerprint detection condition. <P>SOLUTION: The fingerprint sensor (1) for converting the rugged information of a fingerprint into binary data of "0" and "1" on the basis of a prescribed threshold is provided with a threshold setting means for setting the threshold so that the rate of "0" to "1" in binary data becomes similar in the detected area of the fingerprint. Since the threshold is set so that the rate of "0" to "1" in binary data becomes similar in the detected area of the fingerprint, the fingerprint can be clearly discriminated. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a fingerprint sensor that reads unevenness information of a fingerprint as a change in capacitance, and more particularly to an improved technique for searching for an optimum fingerprint detection condition.

Conventionally, a fingerprint sensor has formed a detection circuit for reading fingerprint unevenness information on a single crystal silicon substrate (for example, JP-A-11-118415, JP-A-2000-346608, JP-A-2001-56204). JP, 2001-133213, A). Since the fingerprint sensor is inevitably required to have a size of about 20 mm × 20 mm for its use, the fingerprint sensor becomes expensive when it is formed on a single crystal silicon substrate. As a technique for solving this, a fingerprint sensor using a MIS type thin film semiconductor device has been proposed (Japanese Patent Laid-Open No. 2003-254706). By using the MIS thin film semiconductor device, the fingerprint sensor can be formed on a glass substrate or a plastic substrate.
JP-A-11-118415 JP 2000-346608 A JP 2001-56204 A JP 2001-133213 A Japanese Patent Laid-Open No. 2003-254706

  However, in this type of fingerprint sensor, various external factors such as temperature, humidity, skin condition, and pressure from the finger change, so that the optimum value such as the reference potential fluctuates every time the fingerprint is detected. . For this reason, it is necessary to reset the fingerprint detection condition so as to be optimal every time the fingerprint is detected. In the configuration of the conventional fingerprint sensor, even when searching for the optimum fingerprint detection condition, information is read from all the capacitance detection circuits arranged in the fingerprint detection unit in the same manner as the fingerprint unevenness information reading process. Yes. For this reason, it takes a great amount of time to search for fingerprint detection conditions.

  Therefore, an object of the present invention is to propose a fingerprint sensor, a biometric authentication device, a fingerprint detection condition search method, and a fingerprint detection method that can efficiently search for an optimal fingerprint detection condition in a short time.

  In order to solve the above-described problem, the fingerprint sensor of the present invention is a fingerprint sensor that converts the unevenness information of a fingerprint into binary data of “0” and “1” based on a predetermined threshold value. Threshold setting means is provided for setting a threshold so that the ratio of “0” and “1” is approximately the same in the fingerprint detection area. By setting the threshold value so that the ratios of “0” and “1” in the binary data are approximately the same in the detected region of the fingerprint, the fingerprint convexity (crest) and concave (valley) in the fingerprint image are relative to each other. As a result, the distinction between the convex part and the concave part becomes clear, so that the fingerprint discrimination accuracy can be improved.

  Here, the threshold value setting means preferably sets the threshold value based on the standard deviation of the ratio of “0” and “1” of the binary data, and particularly sets the threshold value so that the standard deviation value is maximized. Is more preferable. When the threshold is set so that the standard deviation of the ratios of “0” and “1” in the binary data becomes the maximum value, the relative ratios of “0” and “1” in the binary data are almost the same. .

  Further, the threshold value setting means obtains binary data by changing the threshold value in the vicinity of the previously obtained threshold value, so that the ratio of “0” and “1” of the binary data is approximately the same in the fingerprint detection area. It is preferable to obtain the threshold value so that Although the optimum value for the fingerprint detection condition changes every time the fingerprint is detected, it is rare that the optimum threshold value fluctuates greatly in a short time. The search time can be shortened by searching for the optimum threshold value in the vicinity of the obtained threshold value (for example, the threshold value obtained when the fingerprint of the previous frame is detected). This shortens the entire processing time including the preparation period and the fingerprint detection period, and enables high-speed sensing.

  Here, the detection area that defines the detection range of the fingerprint is preferably set to an area between the location where the logical value of the binary data has changed first and the location at which it has changed last. The binary data is a continuous value of “0” or “1” outside the fingerprint detection area. Even if the threshold is set so that the ratios of “0” and “1” including the binary data other than the fingerprint detection area are approximately the same, a considerably large error is included. Therefore, if a location where the logical value of the binary data changes first and a location where the logical value changes last is detected, it can be determined that the area between both locations is the detected region of the fingerprint.

  The biometric authentication device of the present invention includes the fingerprint sensor of the present invention. Here, “biometrics authentication device” refers to a device equipped with a function for performing personal authentication using fingerprint information as biometric information, and various card media such as an IC card, a cash card, a credit card, and an identification card. In addition, it includes various security systems such as a personal authentication device for electronic commerce, a mobile phone equipped with a fingerprint authentication function, an entrance / exit management device, and an authentication device for a computer terminal device.

  In the method for searching for fingerprint detection conditions according to the present invention, a process of substituting a predetermined value for a threshold value serving as a reference for converting fingerprint unevenness information into binary data of “0” and “1”, and a predetermined value are substituted. The ratio of “0” and “1” in the detected region of the fingerprint is the same by repeatedly performing the process of converting the unevenness information of the fingerprint into binary data “0” and “1” based on the threshold value. The threshold value is searched so as to be approximately. This method makes it possible to search for a threshold with high fingerprint discrimination accuracy in a short time.

  The fingerprint detection method of the present invention uses the process of searching for a threshold value using the fingerprint detection condition search method of the present invention and the fingerprint unevenness information as “0”, “1” based on the threshold value obtained by the fingerprint detection condition search method. And the process of converting to binary data. The fingerprint detection accuracy can be improved by optimizing the fingerprint detection condition as a preliminary preparation for each fingerprint detection.

  Hereinafter, a fingerprint sensor according to an embodiment of the present invention will be described with reference to the drawings. FIG. 4 shows a circuit configuration of a capacitance detection circuit 31 that reads unevenness information of a subject's fingerprint as a capacitance change and converts it into a current signal. The capacitance detection circuit 31 includes a selection transistor 32 for selecting the detection circuit 31, a capacitance 33 formed between the fingertip of the subject and the sensor electrode, and a minute capacitance of the capacitance 33. A signal amplifying transistor 34 that amplifies a detection signal that bears unevenness information of the fingerprint based on the capacitance change, a scanning line 36 that transmits a signal for performing opening / closing control of the selection transistor 32, and a data line 37 that transmits the detection signal The low-potential power line 39 that takes the ground potential Vss of the signal amplification transistor 34 and the reference capacitor Cs having a constant capacitance value are configured. When the capacitance value of the capacitance 33 is Cd, the detection capacitance Cd is determined according to the distance between the unevenness of the fingerprint of the subject and the sensor electrode (see FIG. 10).

  In the above configuration, when a logic level H signal is output on the scanning line 36 and the selection transistor 32 is opened, a detection current determined by the gate potential of the signal amplification transistor 34 flows through the data line 37. This detection current includes information on the unevenness of the fingerprint. The gate potential of the signal amplification transistor 34 is determined by the respective capacitance ratios of the parasitic capacitance Ct (not shown) of the signal amplification transistor 34 itself, the reference capacitance Cs, and the detection capacitance Cd. For example, when the fingertip of the subject is brought close to the sensor electrode and the convex portion of the fingerprint is close to the sensor electrode, the detection capacitor Cd becomes sufficiently larger than the parasitic capacitor Ct and the reference capacitor Cs, and the signal amplification transistor 34 The gate potential approaches the ground potential Vss. As a result, the signal amplification transistor 34 is substantially turned off, and an extremely weak detection current flows between the source and drain of the signal amplification transistor 34. On the other hand, when the concave portion of the fingerprint is close to the sensor electrode, the detection capacitance Cd becomes sufficiently smaller than the parasitic capacitance Ct and the reference capacitance Cs, and the gate potential of the signal amplification transistor 34 approaches the potential of the scanning line 36. In a state where the scanning line 36 is active, the potential of the scanning line 36 is the high potential Vdd. As a result, the signal amplification transistor 34 is substantially turned on, and a detection current larger than the above-described weak current flows between the source / drain of the signal amplification transistor 34. Here, since the source terminal of the signal amplification transistor 34 is connected to the low potential power supply line 39, the direction of the detection current flowing through the signal amplification transistor 34 is the direction flowing from the data line 37 to the low potential power supply line 39. In other words, the detection current carrying the unevenness information of the subject's fingerprint is output so as to flow into the capacitance detection circuit 31 from the external circuit.

  FIG. 10 is a cross-sectional structure diagram of the capacitance detection circuit 31 centering on the sensor electrode. As shown in the figure, in the capacitance detection circuit 31, a capacitance 33 is formed between a signal amplification transistor 34 that outputs a detection signal that bears the unevenness information of the fingerprint and the fingertip F of the subject. Sensor electrode 72 is formed. The signal amplification transistor 34 is a transistor including a gate electrode 68, a gate insulating film 67, a polycrystalline silicon layer 63, a source electrode 65, and a drain electrode 66. The capacitance 33 is a variable capacitance whose capacitance value changes according to the concave / convex pattern of the fingerprint. The potential of the fingertip F is set to the reference potential. The sensor electrode 72 is connected to the gate electrode 68 and is configured to transmit a change in the detection capacitance Cd due to the unevenness of the fingerprint to the signal amplification transistor 34 and to sense a change in capacitance by an amplification action of the drain current flowing through the channel. Has been.

  FIG. 1 shows a circuit configuration of the fingerprint sensor 1. As shown in the figure, the fingerprint sensor 1 includes a data line driver 10 for selecting a data line 37, a scanning line driver 20 for selecting a scanning line 36, and the capacitance detection circuit 31 described above in a matrix. And an amplification circuit 40 for amplifying a detection signal output from the capacitance detection circuit 31. Capacitance detection circuits 31 are arranged in a matrix (n rows × m columns) in the fingerprint detection unit 30, and n scanning lines 36 and n low potential power supply lines 39 are wired in the row direction. The m data lines 37 are wired along the column direction. The n × m capacitance detection circuits 30 arranged in the fingerprint detection unit 30 are referred to as one frame. The data line driver 10 includes a shift register 11 that generates a timing signal for selecting the data line 37 by analog dot sequential driving, a buffer 12, and an analog switch 13 for selecting the data line 37. The scanning line driver 20 includes a shift register 21 that generates a timing signal for sequentially selecting the scanning lines 36 and a buffer 22.

  Here, CLKX is a clock signal serving as a reference for selecting the data line 37, CLKBX is an inverted signal thereof, RSTX is a reset signal of the data line driver 10, SPX is a start pulse signal of the data line driver 10, and CLKY is a scanning line. A clock signal serving as a reference timing for selecting 36, CLKBY is an inverted signal thereof, RSTY is a reset signal of the scanning line driver 20, SPY is a start pulse signal of the scanning line driver 20, and OUT is a fingerprint amplified by the amplifier circuit 40. This is a detection signal. When fingerprint detection is performed, the reset signal RSTX is at a high level, and the data line driver 10 and the scanning line driver 20 perform a normal shift register operation. The scanning line driver 20 sequentially transfers input information of the start pulse signal SPY in synchronization with the clock signal CLKY. On the other hand, the data line driver 10 sequentially transfers input information of the start pulse SPX in synchronization with the clock signal CLKX in synchronization with the operation of the scanning line driver 20. Thereby, the capacitance detection circuit 31 arranged in the fingerprint detection unit 30 is selected one by one for each row and each column, and the fingerprint detection operation is performed.

  FIG. 2 shows a detailed circuit configuration of the data line driver 10. As shown in the figure, the shift register 11 includes a clocked inverter 14 that controls reception of a pulse signal input from the previous stage of the shift register, an inverter 15 that inverts the output of the clocked inverter 14, The clocked NAND circuit 16 is configured to invert the output (output to the subsequent stage of the shift register). XSEL {1}, XSEL {2},..., XSEL {m} are buffer outputs for controlling opening / closing of the m analog switches 13. Here, the above-mentioned clocked NAND circuit 16 has a circuit configuration shown in FIG.

  FIG. 3 shows a detailed circuit configuration of the scanning line driver 20. As shown in the figure, the shift register 21 includes a clocked inverter 23 that controls reception of a pulse signal input from the previous stage of the shift register, an inverter 24 that inverts the output of the clocked inverter 23, The clocked NAND circuit 25 is configured to invert the output (output to the subsequent stage of the shift register). YSEL {1}, YSEL {2},..., YSEL {n} are buffer outputs output to the scanning line 36. Here, the clocked NAND circuit 25 described above has a configuration similar to the circuit configuration shown in FIG. 5 (configuration in which CLKX, CLKBX, and RSTX are replaced with CLKY, CLKBY, and RSTY, respectively).

FIG. 6 shows a circuit configuration of the amplifier circuit 40. The amplifier circuit 40 is a circuit for amplifying the detection signal of the capacitance detection circuit 31, and includes a current mirror circuit 41 at the front stage and a current mirror circuit 42 at the rear stage. The current mirror circuit 41 in the previous stage compares the constant reference current I _ref output from the transistor 38 whose gate potential is held at the reference potential VR with the detection current I _dat output from the signal amplification transistor 34. The subsequent current mirror circuit 42 outputs a detection signal OUT obtained by amplifying the difference between the reference current I _ref and the detection current I _dat . Fingerprint information composed of binary data (digital data) can be obtained by comparing the signal level of this detection signal OUT with a predetermined threshold value. Here, the case where the detection signal OUT is equal to or higher than the threshold value is set to the high level, and the case where the detection signal OUT is less than the threshold value is set to the low level. Since the value of the reference current I _ref is determined by the reference potential VR, the difference between the reference current I _ref and the detection current I _dat can be increased or decreased by adjusting the reference potential VR, and the contrast of the fingerprint can be adjusted. In the figure, the CLK signal is twice as fast as the clock signal CLKX input to the shift register 11 and is synchronized with the switching timing of the analog switch 13.

  FIG. 7 shows a timing chart of the scanning line driver 20 when reading the fingerprint information of the first frame. The optimum fingerprint detection condition is searched in the preparation period, and the fingerprint information is actually read in the subsequent fingerprint detection period. In both the preparation period and the fingerprint detection period, the reset signal RSTY of the scanning line driver 20 is at a high level. In the preparatory period before fingerprint detection, the scanning line driver 20 sets the outputs of YSEL {1}, YSEL {2},..., YSEL {n} to H level in synchronization with the clock signal CLKY and the clock inversion signal CLKBY. 36 are sequentially selected. At this time, the reset signal RSTX of the data line driver 10 is at a low level, and the operation of the data line driver 10 is stopped (period A1). Next, the clock signal CLKY and the clock inversion signal CLKBY are stopped at a timing when a specific scanning line 36 (YSEL {n / 2} in FIG. 7) is selected. At this time, only one scanning line 36 (YSEL {n / 2}) is selected. In this state, the reset signal RSTX of the data line driver 10 becomes High level, and the data line driver 10 sequentially selects the data lines 37 in synchronization with the clock signal CLKX and the clock inversion signal CLKBX (period B). As the data lines 37 are sequentially selected, the capacitance detection circuit 31 connected to the specific scanning line 36 is sequentially selected. Then, a current corresponding to the unevenness information of the fingerprint flows from the amplifier circuit 40 toward the capacitance detection circuit 31. While the specific scanning line 36 is selected, the fingerprint detection conditions are set to various conditions, fingerprint information is read, and the optimum fingerprint detection conditions are searched. Here, the reference potential VR is swept from Vss to Vdd as fingerprint detection conditions, the detection signal OUT is obtained for each reference potential VR, and the optimum condition is searched.

A method for searching for the optimum value of the reference potential VR will be described below.
(1) “1” is set when the detection signal OUT is at a high level, and “0” is set when the detection signal OUT is at a low level.
(2) Reference potential VR = V11 (Vss ≦ V11 ≦ Vdd).
(3) A standard deviation is obtained between a point that first becomes “0” and a point that finally becomes “0” on a specific scanning line 36, and this is set as σ11 (reference potential VR = V11). Standard deviation).
(4) The reference potential VR is swept from V12 to V1j and the above (3) is repeated to obtain the reference potential VR1max that maximizes the standard deviations σ11 to σ1j.
(5) VR1max is set to the optimum value of the reference potential VR in the first frame.

  Here, when the reference potential VR = Vss, the transistor 38 is substantially turned off, and the detection signals OUT are all “1”. On the other hand, when the reference potential VR = Vdd, the transistor 38 is substantially on, and all the detection signals OUT are “0”. The optimum reference potential VR for reading the fingerprint information is determined in the range from Vss to Vdd.

  If the optimum reference potential VR is determined by the above-described search method, the reset signal RSTX of the data line driver 10 again becomes the low level, and the selection operation of the data line 37 by the data line driver 10 is stopped (period A2). Then, the input of the clock signal CKLY and the clock inversion signal CLKBY to the scan line driver 20 is resumed, and the scan line driver 20 sequentially selects the scan lines 36 until the last row (YESL {n}), and the preparation period ends. To do. After completion of the preparation period, the reset signal RSTX becomes High level, and the process proceeds to the fingerprint detection period. In the fingerprint detection period, fingerprint information is read using the optimum value VR1max of the reference potential VR obtained in the preparation period preceding this.

FIG. 8 shows a timing chart of the scanning line driver 20 when reading fingerprint information from the second frame. Since the driver operation in the periods A1 and A2 and the driver operation in the fingerprint detection period in the preparation period are the same as the driver operation in the first frame described above, detailed description thereof is omitted. In period B of the preparation period, optimal fingerprint detection conditions are searched for by performing the following processing.
(1) “1” is set when the detection signal OUT is at a high level, and “0” is set when the detection signal OUT is at a low level.
(2) The optimum value VR1max of the first frame is set to V21, and V21 is set to the reference potential VR.
(3) A standard deviation is obtained between a point that first becomes “0” and a point that finally becomes “0” on a specific scanning line 36, and this is set as σ21 (reference potential VR = V21). Standard deviation).
(4) The reference potential VR is swept from V21 to V2k (k <j) and the above (3) is repeated to obtain the reference potential VR2max that maximizes the standard deviations σ21 to σ2k.
(5) VR2max is set to the optimum value of the reference potential VR in the second frame.

  In the third and subsequent frames, the optimum value of the reference potential VR is searched in the same manner as in the second frame.

  FIG. 9 shows a fingerprint image obtained from the detection signal OUT converted into binary data. When the crest of the fingerprint is close to the sensor electrode 72, the logical level of the detection signal OUT is low, and when the valley of the fingerprint is close to the sensor electrode 72, the logical level of the detection signal OUT is high. When the low level of the detection signal OUT is displayed in black and the high level is displayed in white, as shown in FIG. 9, the peak of the fingerprint is displayed in black and the valley is displayed in white. When the reference potential VR is close to Vss, the fingerprint image is displayed in white as a whole, and the standard deviation becomes a value close to 0. On the other hand, when the reference potential VR is close to Vdd, the fingerprint image is displayed in black as a whole, and the standard deviation is close to 0. Thus, the contrast of the fingerprint image varies greatly depending on the value of the reference potential VR. When the ratio of white and black is approximately the same, the contrast of the fingerprint image becomes clear and the fingerprint discrimination accuracy can be improved. The standard deviation at this time is a value close to 0.5 (maximum value). Therefore, in order to search for the optimum reference potential VR, reference is made so that white and black of the fingerprint image have the same ratio, that is, the standard deviation of the ratio of “0” and “1” is maximized. The potential VR may be set. Here, the data line driver 10, the scanning line driver 20, and the amplifier circuit 40 function as a threshold setting unit that sets a threshold value serving as a reference for converting fingerprint unevenness information into binary data “0” and “1”. . In the above-described example, the standard deviation is exemplified as a measure for making the ratios of “0” and “1” in the binary data comparable, but the present invention is not limited to this, and various methods can be used. .

  Next, an application example of the fingerprint sensor 1 will be described. FIG. 11 shows a schematic configuration diagram of a smart card 81, which includes the above-described fingerprint sensor 1, an IC chip 82 on which a CPU, a memory element, and the like are mounted, and a display device 83 such as a liquid crystal display. In the IC chip 82, fingerprint information of the cardholder is registered as biometric information. FIG. 12 shows the authentication procedure of the smart card 81. When the card user brings the fingertip into contact with the fingerprint sensor 1 and fingerprint information is input to the smart card 81 (step S1), the fingerprint information is checked against previously registered fingerprint information (step S2). Here, if the fingerprints match (step S2; YES), a personal identification number is issued (step S3). Next, a password is entered by the cardholder (step S4). It is checked whether or not the password issued in step S3 and the password entered in step S4 match (step S5). If they match (step S5; YES), the card is used. Is permitted (step S6).

  In this way, a smart card with high security can be provided by authenticating the person using fingerprint information in addition to the personal identification number. A smart card with a biometrics authentication function can be used for cash cards, credit cards, identification cards, etc. The fingerprint sensor of the present embodiment can be applied to any biometric authentication device for performing personal authentication. For example, as a security system for managing entry / exit into the room, the fingerprint sensor of this embodiment is attached to the door, and the fingerprint information of the occupant input to the fingerprint sensor is compared with the fingerprint information registered in advance, If the two match, the entry is permitted, while if the two do not match, the entry is not permitted and the system can be applied to a security company or the like as necessary. The fingerprint sensor of the present embodiment can also be effectively applied as a biometric authentication device for identity verification in electronic commerce through an open network such as the Internet. Further, the present invention can be widely applied to a user authentication device for a computer terminal device and a management device for a copying machine user of a copying machine. Further, it can be applied to a mobile communication terminal such as a mobile phone equipped with a fingerprint authentication function. The capacitance detection device of the present invention is not limited to fingerprint detection, and can be applied to a device that reads the surface shape of a test object having minute irregularities as a change in capacitance.

  According to the present embodiment, by setting the threshold value so that the ratio of “0” and “1” of the binary data is approximately the same in the detected region of the fingerprint, the convex portion (mountain) of the fingerprint in the fingerprint image. As a result of the relative ratio of the concave portions (valleys) being approximately the same, the distinction between the convex portions and the concave portions becomes clear, so that the fingerprint discrimination accuracy can be increased. In addition, the optimum value of the fingerprint detection condition changes every time the fingerprint is detected, but the optimum threshold value rarely fluctuates greatly in a short time. The search time can be shortened by searching for an optimum threshold value in the vicinity of the threshold value obtained at times (for example, the threshold value obtained at the time of fingerprint detection in the previous frame). As a result, it is possible to flexibly cope with the deterioration with time of the fingerprint detection conditions caused by external factors, and it is possible to perform stable and stable fingerprint detection.

It is a circuit block diagram of the fingerprint sensor of this embodiment. It is a circuit block diagram of a data line driver. It is a circuit block diagram of a scanning line driver. It is a circuit block diagram of an electrostatic capacitance detection circuit. It is a circuit block diagram of clocked NAND. It is a circuit block diagram of an amplifier circuit. 6 is a timing chart of the scanning line driver (first frame). 6 is a timing chart of the scanning line driver (after the second frame). It is explanatory drawing of a fingerprint image. It is sectional drawing of an electrostatic capacitance detection circuit. It is explanatory drawing of a smart card. It is a flowchart describing an authentication procedure.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Fingerprint sensor 10 ... Data line driver 20 ... Scan line driver 30 ... Fingerprint detection part 31 ... Capacitance detection circuit 36 ... Scan line 37 ... Data line 40 ... Amplifier circuit

Claims (8)

A fingerprint sensor that converts unevenness information of a fingerprint into binary data of “0” and “1” based on a predetermined threshold,
A fingerprint sensor comprising threshold setting means for setting the threshold so that the ratios of “0” and “1” of the binary data are substantially the same in the detected region of the fingerprint. The fingerprint sensor according to claim 1,
The threshold value setting means sets the threshold value based on a standard deviation of a ratio of “0” and “1” of the binary data. The fingerprint sensor according to claim 2,
The fingerprint sensor, wherein the threshold value setting means sets the threshold value so that the standard deviation value is maximized. The fingerprint sensor according to any one of claims 1 to 3,
The threshold value setting means obtains binary data by changing the threshold value in the vicinity of the previously obtained threshold value, so that the ratio of “0” and “1” of the binary data is approximately the same in the detected region of the fingerprint. A fingerprint sensor that finds a threshold value. The fingerprint sensor according to any one of claims 1 to 4,
The to-be-detected area is a fingerprint sensor that is an area between a location where the logical value of binary data has changed first and a location where it has changed last.   A biometric authentication device comprising the fingerprint sensor according to any one of claims 1 to 5. A process of substituting a predetermined value for a threshold value serving as a reference for converting the unevenness information of the fingerprint into binary data of “0” and “1”;
Converting the unevenness information of the fingerprint into binary data of “0” and “1” based on the threshold value into which the predetermined value is substituted;
A method for searching for a fingerprint detection condition, in which the threshold value is searched so that the ratio of “0” and “1” in the detected region of the fingerprint becomes approximately the same by repeatedly executing. The process of searching for the threshold value by the fingerprint detection condition search method according to claim 7;
A process of converting the unevenness information of the fingerprint into binary data of “0” and “1” based on the threshold value obtained by the search method of the fingerprint detection condition;
A fingerprint detection method including:

Images