GB2508039A

GB2508039A - A fingerprint sensor having a number of separately operable sections

Info

Publication number: GB2508039A
Application number: GB1304639.6A
Authority: GB
Inventors: Zeljko Stojkovic; Bojan Kosic
Original assignee: Zwipe AS
Current assignee: Zwipe AS
Priority date: 2012-09-14
Filing date: 2013-03-14
Publication date: 2014-05-21
Anticipated expiration: 2033-03-14
Also published as: GB201304639D0; GB2508039B; GB201216433D0

Abstract

A fingerprint sensor apparatus comprises: a controller (14); and a sensor area 8 formed of a plurality of separately operable sensor sections 16, where each sensor section captures a representation of an area of a fingerprint. The controller is arranged to activate each of the plurality of sensor sections at separate times to obtain a representation of the fingerprint area at each section, thereby permitting a representation of the fingerprint over the entire sensor area to be generated. The sensor sections could be capacitive sensors. Each sensor section may contain an array of pixels (18). The aims of the invention include reducing power usage compared to a single full-area sensor. The sensor may find application in a smart card (2) or an ID token.

Description

FINGERPRINT SENSOR

The present invention relates to a fingerprint sensor and in particular to a fingerprint sensor for use in an ID token such as an electronic card.

A fingerprint sensor is an electronic device used to capture a representation of the fingerprint ridge pattern. Often the sensor is designed to capture a representation of just a part of the fingerprint pattern since it is possible to correctly identify a person with a high degree of confidence from even a partial fingerprint.

Known fingerprint sensors include optical sensors and capacitive sensors.

An optical sensor is basically a specialised form of digital camera and a digital image of the print is captured using visible light. Capacitive sensors use the electrically conductive dermal layer of the skin as one plate of a parallel plate capacitor with the sensor forming the other plate, typically in the form of an array of pixels. The non-conductive epidermal layer of the skin acts as the dielectric. The capacitance differs for ridges and valleys of the fingerprint due to the air gap between the dermal layer and the sensor in the valleys. In a passive capacitance sensor each pixel measures the local capacitance and this allows a map of the pattern of ridges and valleys to be built up. An active capacitance sensor applies a voltage to the skin in order to charge the capacitor formed by the skin and the sensor. When the voltage is discharged the electric field follows the patterns of ridges and pixels of the sensor can therefore map the fingerprint pattern.

The image data (capacitive or optical) from the sensor is processed to generate a template or map of features of the fingerprint, which can be based on the locations and orientations of basic fingerprint patterns (arches, loops and whorls) and/or minutiae. Minutiae is a term used to describe fingerprint features including ridge bifurcation points, ridge end points and so on. The template generated from the captured data is stored and/or used for matching by comparison with previously stored known fingerprint data for authentication purposes.

Matching algorithms are used to compare the fingerprint templates. These can be pattern based algorithms or image based algorithms. Pattern based algorithms are more efficient in terms of processor load and can be used with templates extracted from any type of sensor, not just optical image type data.

Fingerprint sensors can be area sensors that capture an image of an area of the fingerprint when the finger is held stationary on the sensor. These sensors tend to have a relatively high power consumption due to the large number of pixels required. An example of a finger print area sensor is the FPC1O11 F area sensor manufactured by Fingerprint Cards AB of Sweden.

Fingerprint swipe sensors are also known. These sensors require the finger to be slid over the sensor while the sensor takes a series of images or representations of the fingerprint. A processor then stitches the images together to form an image of an area of the fingerprint. An example of a fingerprint swipe sensor is the FPC1O8OA swipe sensor, also manufactured by Fingerprint Cards AB of Sweden. The swipe sensor has a smaller area and uses a smaller number of pixels than an area sensor. Hence it has a lower power consumption and it is also cheaper to manufacture.

One application for fingerprint sensors is to verify the identity of the user of electronic cards such as smart cards. A smart card is a pocket-sized card with an embedded integrated circuit along with volatile and non-volatile memory and microprocessor components. Smart cards provide a way of electronically authenticating the bearer of the card and carrying a secure message from the card to a card reader. For example, if a "non-smart' credit card were to be lost or stolen) an unauthorised user would be able to use the credit card until the credit card is cancelled. Conversely, a smart" credit card may include many more levels of security that would prevent such use by an unauthorised user. When a smart card incorporates a fingerprint sensor this provides a high degree of confidence about the identification of the user of the card. Similar identification techniques can be used with other ID tokens, such as ID tokens used for access to online banking or other computer networks requiring access authentication.

When fingerprint sensors are used in ID token applications like smart cards there are constraints on the power requirements. Typically ID tokens must either harvest energy from an outside source, for example by interaction between an antenna of a contactless smart card and an electromagnetic field, or they must have a battery onboard. Either way the power and voltage available is usually relatively small.

As noted above a swipe sensor uses only a small amount of power and hence is suited for use as a fingerprint sensor for an ID token. However, swipe sensors have disadvantages compared to area sensors. The user needs to have some training to ensure that they use the swipe sensor correctly, since the direction and speed of movement of the finger will affect the quality of the fingerprint image that can be generated. The swipe action means that two hands are required to hold the ID token, which can be cumbersome and uncomfortable for many uses of ID tokens, for example smart cards for access to buildings. Also, the processing load for creating the fingerprint image can be high, since it is necessary to stitch together captured image slices that overlap and may be at slightly different angles.

Viewed from a first aspect, the present invention provides a fingerprint sensor apparatus comprising: a sensor area formed of a plurality of separately operable sensor sections, each sensor section being for capturing a representation of an area of a fingerprint; and a controller; wherein the controller is arranged to activate each of the plurality of sensor sections at separate times to obtain a representation of the fingerprint area at each section, thereby permitting a representation of the fingerprint over the entire sensor area to be generated.

This fingerprint sensor apparatus can advantageously have power consumption in the range of that of a typical swipe type sensor, but the ease of operation and accuracy of a conventional area sensor. This sensor offers more comfort for the user than a swipe sensor and allows a simple single-handed operation while also maintaining a total power consumption considerably lower than that of a conventional fingerprint area sensor of equivalent size. Instead of using a single swipe sensor to capture a series of representations of a moving finger a series of adjacent sensor sections are activated to obtain a series of representations of a static finger. As only one section is activated at a time then the power consumption can be the same as for a single swipe sensor. Also, since the representations will not overlap and as they come from adjacent sections of a larger sensor array then only minimal processing is required to combine the representations to generate a larger representation of the fingerprint over the entire area of the sensor. Advantageously, each sensor section may comprise an array of pixels.

In the current context activation of the sensor section comprises supplying power to the sensor section so that it can capture a representation of a fingerprint placed over the sensor section. The non-activated sensor sections are sections that are not supplied with power. The representation may be an image of the fingerprint or a data structure representing properties of the fingerprint, such as the location of ridges and valleys.

The controller may comprise one or more processors or computer devices, memory, communication device(s) such as RFID transceiver(s) and other elements typically found in controllers for fingerprint sensors and/or in ID tokens such as electronic cards. The controller for the fingerprint sensor area may be a controller also having other functions, for example in relation to controlling an electronic card or other device into which the fingerprint sensor apparatus is integrated. The exact form of the controller is not of great importance provided it is capable of performing the required function(s). The controller may include parts integrated into a circuit or circuit board that also includes the sensor area and/or the controller may include parts on a separate circuit/circuit board.

Preferably the controller is arranged to collect the representations from each sensor section. The plurality of representations may be stored, at least temporarily, in a memory of the controller. In a preferred embodiment the controller is arranged to carry out processing to combine the plurality of representations into a representation of the area of the fingerprint covering the whole of the sensor area.

As noted above this processing is simplified compared to a swipe sensor since there is no overlap of the mages/representations from the separate sensor sections.

The controller may be arranged to convert the representation into a fingerprint template, preferably based on the identification and mapping of minutiae, The controller may then be arranged to compare this fingerprint template with a fingerprint template stored in a memory to determine if the fingerprint matches that of an authorised user of the sensor and/or of a device into which the fingerprint sensor apparatus is incorporated, for example an electronic card. The controller may hence be arranged to run a fingerprint matching algorithm such as a pattern matching algorithm.

The sensor apparatus may be arranged to commence activation of the sensor sections and obtain a representation of a fingerprint automatically when it is determined by the controller that a finger is pressed against the sensor area.

The sensor sections may be any sub-part of the sensor area. Preferably the size of the sensor sections is set based on a desired maximum power consumption during activation of the sensor section. The sensor sections may be sized so that the maximum power consumption of one sensor section, when activated, is less than 10 mW, preferably less than 5 mW, more preferably less than 3 mW. For example, the size of the sensor sections may be based on the sensor array of the known FPC1O8OA sensor, which uses a supply voltage of 1.8 V and draws a maximum total current of 1.5 mA, giving a maximum power consumption of 2.7 mW. By setting the size of the sensor sections based on a maximum power consumption the power usage for the fingerprint sensor apparatus can be managed. As noted above the reduced power consumption compared to standard fingerprint area sensors is an important advantage of the invention.

In a preferred embodiment the sensor sections are located adjacent one another without significant gaps, thereby providing the ability to capture a representation of the fingerprint at the sensor area without significant gaps. The sensor sections are preferably placed adjacent one another with maximum gaps between active sensor parts of less than 0.1 mm, preferably less than 0.05 mm. In some preferred embodiments the spacing of the sensor sections is such that the spacing between adjacent pixels of adjacent sensor sections is about the same as the spacing between the pixels within the sensor sections. Preferably the spacing between adjacent pixels of adjacent sensor sections is identical to the spacing between the pixels within the sensor sections. This arrangement may be achieved by use of a one-chip sensor in which a sensor area comprising an array of pixels is functionally divided by control circuitry into sensor sections that can be activated separately, each sensor section comprising a subset of the pixels of the sensor array.

Preferably the sensor area size is set based on a minimum area required for accurate fingerprint recognition. In preferred embodiments the sensor area is larger than 75 mm2, preferably larger than 100 mm2. For example, the sensor area may be about 125 mm2. The exact minimum area required will depend on the matching algorithm used and more advanced algorithms may be able to match a fingerprint based on a smaller size of partial fingerprint. Since the sensor is a fingerprint sensor then the sensor area need not be any larger than a typical area required for fingerprint recognition, for example the sensor area may be less than 200 mm2.

The number of sensor sections may vary depending on the size of the entire sensor area. Also, as noted above, the size of the sensor sections, which will determine the total number required to cover the sensor area, is preferably set based on a maximum power consumption. The sensor apparatus preferably comprises at least thirty sensor sections, more preferably at least forty, for example about tifty sensor sections may be used.

In a preferred embodiment the sensor area is a rectangular strip more than mm in length, for example a strip of about 20 mm in length. The strip may have a width as required to provide the minimum area, for example a width of 5 mm or above. A rectangular strip of these dimensions has been found to provide a fingerprint image suitable for fingerprint matching whilst also minimising the size of the sensor and hence reducing its cost. The use of a rectangular strip with this type of aspect ratio also allows the fingerprint area sensor to generate a representation similar to the representations generated by conventional swipe sensors. This means that the controller and matching algorithms can make use of proven fingerprint identification techniques developed for swipe sensors.

The sensor sections preferably each comprise an array of pixels with at least two rows and at least two columns in the array. The sensor sections may be elongate sections running across the width of the sensor area. For example, where the sensor area is a rectangle the sensor sections may be slices across the width (short side) of the rectangle. This is a convenient way to create a rectangle with dimensions as discussed above, and also advantageously permits the sensor sections to be based on the sensor arrays of a conventional swipe sensor. Thus, in a preferred embodiment the individual sensor sections comprise sensor arrays similar to those found in conventional swipe sensors, for example a pixel array of 128 columns and eight rows as in the FPC1O8OA sensor mentioned above. Once again this facilitates the use of existing control techniques and matching algorithms, since the output of a plurality of sensor sections formed as slices across the width of a rectangle will be similar to the output of similar sensor sections used as a swipe sensor, The sensor sections may form slices of 1 mm or less extending across the width (short side) of the rectangle, preferably slices of 0.5 mm or less.

It is preferred for the sensor sections to be activated by the controller in a sequence corresponding to the location of the sensor sections in the sensor area.

For example where the sensor sections are elongate sections across the width of the sensor area then they may be activated in sequence starting at one end of the sensor area and progressing along the length of the sensor area. With this arrangement the output of the sensor sections is a sequence of representations that will be adjacent to the preceding/following representation(s) when combined to generate a fingerprint image for the whole sensor area. This simplifies processing of the fingerprint data.

The use of separately operable sensor sections and the reduction in processing loads due to the absence of overlap in the captured representations allows activation of the sensor sections to occur at a frame rate that is higher than the frame rate for conventional swipe sensors. Frame rate in the context of the current fingerprint sensor apparatus is the rate at which the sensor sections capture representations and hence represents the rate at which the separate sensor sections are activated. Preferably the sensor sections are activated at a frame rate of 500 sections per second or above, more preferably 1000 sections per second or above. With a typical sensor area made up of perhaps fifty sensor sections a rate of 1000 sections per second results in a total scanning time of just 50 ps.

The sensor sections preferably comprise an array of pixels, each pixel being for producing a signal based on the properties of an adjacent pad of the fingerprint.

For example the pixels may generate a voltage or current that varies depending on if the adjacent pad of the fingerprint is a ridge or a valley. The sensor sections may be optical sensors, but preferably they are capacitive sensors. The capacitive sensing of fingerprints may be based on known sensor types, as discussed above.

The number of pixels of the sensor sections is preferably set to give a minimum required resolution, for example at least 250 dpi, preferably at least 500 dpi.

In preferred embodiments the fingerprint sensor apparatus is for an ID token. An ID token in this context should include any security device given to authorized users, for example ID tokens for access to online banking or other secure networks, or an electronic card, such as a smart card. An example of an ID token used for network access is the SecureIDTM Card manufactured by RSA Security, Inc., of Bedford, MA, United States of America. The controller may be arranged to perform fingerprint based authentication of a user and to allow access to the function(s) of the ID token only for authorised users. The invention extends to an ID token including the fingerprint sensor apparatus. One preferred embodiment is an electronic card including the fingerprint sensor apparatus.

In a preferred embodiment, the electronic card comprises a card body and a circuit embedded within the card body, wherein the circuit comprises the controller and the sensor area is connected to the circuit.

The circuit may include a processor and a memory, which may be a part of the controller. The memory is preferably arranged to store fingerprint information relating to authorised user(s) of the card and the processor is preferably arranged to compare the stored information with a representation of the fingerprint obtained by the fingerprint sensor apparatus.

The electronic card may be any one of: an access card; a credit card; a debit card; a pre-pay card; a loyalty card; an identity card or a cryptographic card.

The electronic card is preferably arranged to be inoperable if the fingerprint sensor apparatus does not provide an indication of an authorised user.

Viewed from a second aspect, the invention provides a fingerprint sensing method using a fingerprint sensor apparatus comprising a sensor area formed of a plurality of separately operable sensor sections each for capturing a representation of an area of a fingerprint, the method comprising: activating each of the plurality of sensor sections at separate times to obtain a representation of the fingerprint area at each section, thereby permitting a representation of the fingerprint over the entire sensor area to be generated.

The sensor sections preferably each comprise an array of pixels with at least two rows and at least two columns in the array.

In preferred embodiments the method comprises use of a fingerprint sensor apparatus having features as described above in connection with the first aspect.

The method may comprise the use of the fingerprint sensor apparatus in an ID token for authentication of the user of the ID token and the method may include allowing access to the function(s) of the ID token only for authorised users. The ID token may be an electronic card.

The method may comprise collecting the representations from each sensor section. The plurality of representations may be stored in a memory. The method preferably includes a step of processing the representations to combine them into a representation of the area of the fingerprint covering the sensor area. This representation may be converted into a fingerprint template, preferably based on the identification and mapping of minutiae, and the fingerprint template may be compared with a fingerprint template stored in a memory of the controller to determine if the fingerprint matches that of an authorised user of the sensor and/or of a device into which the fingerprint sensor apparatus is incorporated, for example an electronic card. The method may hence comprise the use of a fingerprint matching algorithm such as a pattern matching algorithm.

It is preferred for the sensor sections to be activated in a sequence corresponding to the location of the sensor sections in the sensor area. For example where the sensor sections are elongate sections across the width of the sensor area then they may be activated in sequence starting at one end of the sensor area and progressing along the length of the sensor area.

The invention also extends to a method of manufacture of a fingerprint sensor apparatus comprising providing an apparatus with features as described above in connection with the first aspect.

The invention further extends to the use of the fingerprint sensor apparatus of the first aspect and the preferred embodiments thereof for user authentication, for example to authenticate the user of an ID token.

Certain preferred embodiments of the present invention will now be described in greater detail by way of example only and with reference to the accompanying schematic drawings, in which: Figure 1 is a partially cut-away side view of an example of a smart card with a fingerprint sensor; Figure 2 shows a partially cut-away plan view of the smart card of Figure 1; Figure 3 shows the pixel matrix of a sensor section of the fingerprint sensor of the invention; and Figure 4 illustrates an embodiment of a fingerprint sensor in accordance with the invention.

In the preferred embodiments a fingerprint area sensor 8 is provided for an electronic card 2 and in particular for a smart card 2. Whilst the following discussion relates particularly to fingerprints it will be understood that the fingerprint sensor can be used in the identification of any print that can be represented by minutiae points including, for example, prints from fingers, thumbs, palms, toes and soles.

The fingerprint area sensor 8 can be incorporated into any suitable smart card as a biometric sensor for additional security. An exemplary smart card 2 is shown schematically in Figure 1 and Figure 2. These drawings are not to scale and certain features have been emphasised. Particularly, the thicknesses of the card body 4 and of the circuit 6 have been enlarged so as to more clearly illustrate the features of the smart card. Also, the size of the fingerprint area sensor 8 is not necessarily in scale with the size of the smart card 2.

Smart cards are generally of a similar size to a conventional credit card and have a similar look and feel. Conventional credit cards are manufactured in accordance with international standard ID-i of ISO/IEC 7810, that is to say having dimensions of 3 inches by 2 1/8 inches (approx. 86 mm by 54 mm) and a thickness of 6 mu (approx. 0.75 mm). In some embodiments, the smart card may be thicker than a conventional credit card in order to accommodate a circuit and biometric sensor.

In contact smart cards, often known as chip cards, metal contacts are brought to the surface of the card in a predetermined pattern to form a contact pad, which is accessible from the outside of the card. These contacts are connected to a microprocessor inside of the card. Smart cards of the first type are commonly used by being slid into a recess of a card reader such that spring loaded contacts in the card reader make contact with the contact pad on the card in order to read the contents of the microprocessor.

In non-contact smart cards, often known as proximity cards, the contents of the microprocessor are transmitted to the reader using non-contact communication technology. One example of such is radio-frequency identification (REID), where an antenna is formed within the body of the card and radio-frequency electromagnetic fields are used by a reader to read of the contents of the microprocessor. Encryption may be employed in smart cards of this type to ensure secure transmission of messages between the card and the reader.

The example smart card 2 of Figures 1 and 2 is a non-contact smart card 2 that comprises a card body 4 and a circuit 6 enclosed within the card body. The circuit 6 is in the form of a printed circuit board, which is preferably made from poly amide or FR-4 grade glass-reinforced epoxy laminate. A fingerprint area sensor 8 is attached to the circuit 6 via a hole formed in the card body 4.

The circuit 6 is laminated between at least two layers of plastic. The layers of plastic would typically be made of PVC; however, other plastics may be used.

Examples of other suitable plastics include polyester, acrylonitrile-butadiene-styrene (ABS), and any other suitable plastic. Additionally, plasticisers or dyes may be added to the plastic to achieve a desired look and feel.

An antenna 10 is connected to the circuit 6. The antenna 10 is used to communicate with a card reader, which is external to card 2, The antenna 10 may be formed by etching a suitable pattern onto a copper cladding of the printed circuit board.

The fingerprint area sensor 8 is joined to the circuitS via contacts 12. The circuitS also includes a number of additional components 14. These can include a processor and a memory. The memory is arranged to store information associated with the smart card 2. For example, this may include the identity of a bearer of the smart card 2, account information of the bearer of the smart card 2, and so on. The processor is arranged to control operation of the smart card. Particularly, subject to verification of the bearer of the smart card 2, for example by use of a password or PIN and/or via data from the fingerprint sensor 8, the processor is arranged to communicate the data stored on the memory to a card reader.

The additional components 14 may, in some embodiments, include a battery configured to power the memory and processor. Alternatively, or in addition to the battery, the card may be arranged to be powered via a contact pad external to the smart card 2 or to draw power from the antenna 10 when it is energised by a card reader.

Figure 3 shows the features of an individual sensor section 16 of the fingerprint area sensor 8 of the preferred embodiment. The sensor section 16 is composed of an array of sensor pixels 18 including at least two rows Rand at least two columns C. The sensor section 16 shown in Figure 3 is similar in size to typical fingerprint swipe sensors. As discussed below this means that elements of known control systems can be used. The sensor section 16 in this example is made up of an array of pixels 18 with eight rows Rand 128 columns C. The array has a width of 6.4 mm and a height of 0.4 mm, hence providing a resolution of 508 dpi. It is an active capacitive type sensor. The sensor 16 generates an 8 bit grey scale value as a representation of the fingerprint and hence a single framelimage from the sensor 16 is 8 kb large.

Figure 4 shows the arrangement of a plurality of sensor sections 16 forming the fingerprint area sensor 8. The fingerprint area sensor 8 consists of several dozens of the 8 x 128 pixel arrays of the sensor sections 16 as illustrated in Figure 4. The fingerprint area sensor 8 is hence made up of a plurality of sections of 128 columns and eight rows lines up together to provide an array with 128 columns and a greatly increased number of rows. In practice the whole area would be built up as one large array of pixels, in this example 128 columns and 400 rows, with a power supply and control system arranged to activate individual eight row sectors of the larger array separately, thereby individually activating the sensor sections 16.

The fingerprint area sensor 8 can be manufactured by making use of conventional ASIC manufacturing techniques to provide a sensor area comprising an array of pixels to form the plurality of sensor sections 16. Instead of a conventional power supply and control system as used with known area sensors the fingerprint area sensor 8 of the invention is provided with a new power supply and control system to separately activate sensor sections as described above. This may be implemented in hardware with software as required. The power supply and control system of one embodiment consists of a timing circuit that shifts the control register and hence enables sequential power supply for each sensor section in turn.

The preferred embodiment covers an area 20 mm long (with a width of 6.4 mm) and hence with the 0.4 mm size of the pixel array of Figure 3 the fingerprint area sensor 8 of this example comprises fifty sensor sections 16. Fingerprint identification is carried out with the area sensor 8 based on the location and mapping of minutiae.

With suitable fingerprint recognition algorithms a 20 mm by 6.4 mm section of fingerprint contains enough minutiae to perform authentication.

A suitable basic fingerprint recognition algorithm is provided with the TMS320C5515 Fingerprint Development Kit product sold by Texas Instruments Inc. of Texas, United States of America. A preferred fingerprint recognition algorithm is described in the applicants co-pending UK patent application No. 1219749.7, entitled "Fingerprint matching algorithm". The sensor may also make use of the fingerprint enrolment algorithm described in the applicant's co-pending UK patent application No. 1219750.5, entitled "Fingerprint enrolment algorithm".

In the prior art swipe sensor, as discussed above, an image of the fingerprint is obtained by swiping the finger over a single sensor to obtain a sequence of images. With the prior art device, when the finger is moved across the sensor at an appropriate rate for the frame rate of the sensor (i.e. frames per second captured by the sensor) then it is possible to construct an image of a slice of the fingerprint by assembling a number of frames/images captured by the sensor.

In the present invention, in order to obtain an image of the 20 mm long area of the fingerprint without any increase in peak power consumption compared to a single sensor array 16 the sensor sections 16 (i.e. individual sets of 8 x 128 sensor cells) of the larger sensor area 8 are turned on sequentially and supply the resultant images sequentially to a controller. Each sensor section 16 in the larger array of the fingerprint area sensor 8 captures a single horizontal slice of the fingerprint image. With the finger stationary on the fingerprint area sensor 8 the sequence of sensors 16 will scan parts of the fingerprint image that can be combined to create a complete image of the fingerprint over the entire area of the area sensor 8. The power consumption is the same as for a swipe sensor with an equivalent pixel array, and considerably less than a conventional area sensor. Advantageously, there is no overlapping of the adjacent slices, which greatly simplifies the software process of slice stitching.

The sequential activation of each sensor can be achieved by a timing circuit as described above that shifts the control register and hence enables sequential power supply for each sensor section in turn. The output from the sensor is similar to the output from a swipe sensor with an array of pixels equivalent to the array of pixels in the sensor sections. Hence, from the host controller's point of view the process of acquiring the fingerprint images is basically unchanged from the process for a single sensor 16 used in a conventional swipe mode (e.g. the use of a single FPC1O8OA swipe sensor). The controller can use a Read and Capture command that is functionally the same as in the case of the swipe sensor. Thus, aside from the elements of the power supply and control system required to sequentially activate the sensor sections, the host controller can essentially be identical to known controllers for swipe sensors. The same memory capacity would be needed for the fingerprint area sensor 8 as for a swipe sensor intended to produce an equivalent size of image. The host controller can read image slices in real time, but with the fingerprint area sensor 8 described herein there is no additional processing for image stitching required at the controller since there is no overlap between the slices captured by the sensors. This means that the image reconstruction process is simply a matter of combining all the individual images by placing the adjacent one another without any complicated image stitching process.

The reduced processing means that a higher frame rate is possible without increasing peak power or processing load. For a conventional swipe sensor a frame rate of 250 fps frame rate is typical. If this were used with the array of sensors in the fingerprint area sensor 8 then the whole process of fingerprint scanning would take 0.2 s. Increasing the frame rate decreases the time required to take a scan of the fingerprint. The exact frame rate can be set based on other features of the smart card, such as the available processing speed. In a preferred embodiment a frame rate of 1000 fps is used, which gives a scanning time of only 50 ps when the array comprises fifty sensors as in Figure 4. Since, for the example sensor of Figure 3, each frame represents 8 kb of data then the serial peripheral interface bus (SPI) clock speed required would be 8 MHz, while the minimum digital signal processor (DSP) clock speed would be 40 MHz.

The smart card 2 of Figures 1 and 2 can be produced by any suitable method, for example the method discussed in US 6,586,078 B2 or a hot lamination method. A typical hot lamination method comprises the following steps: forming a core by providing first and second layers of plastic and positioning the circuit 6 between the first and second layers of plastic to thus form a card core; placing the core in a laminator; applying a heat cycle to the core in the laminator to liquefying or partially liquefying the layers of plastic, the heat cycle operating at a temperature of between 135°C and 250°C; increasing a laminator ram pressure in combination with the heat to a pressure of approximately 6.5 MPa; applying a cooling cycle to the core in the laminator with an associated increase in ram pressure of approximately 25% until the core has cooled to approximately 5°C to 20°C; and removing the core from the laminator.

Conventional processing techniques, that would be well known to the person skilled in the art, may then be applied to the core to complete the card body 4. Such processing techniques may include inking, the formation of an overlaminate film, or the like.

The fingerprint area sensor 8 can be fitted to the card body using a method as described below, although other techniques can also be used.

In the preferred method the smart card 2 is first formed with the laminated body 4 enclosing the circuit 6. The circuit 6 includes contacts 12 that are designed to connect with corresponding contacts on the fingerprint area sensor 8, but immediately after the card body is formed these contacts 12 are encased within the laminated body 4.

A cavity is formed in the card body 4 to expose the contacts 12 in the circuit 6. The cavity is formed on an upper surface of smart card body 4 and is sized substantially in conformity with the shape of the fingerprint area sensor 8, such that the fingerprint area sensor 8 will just fit within the cavity. The cavity is milled into the surface of the card body 4. This may be done using a precision end mill or, more preferably, a laser mill. The depth of the milling is set so that the base of the cavity is at the level of the circuit 6 within the card body 4, such that the contacts 12 are exposed.

A conductive epoxy is applied to the surface of the exposed contacts 12 in order to provide the necessary electrical connection between the sensor 8 and the circuit 6. A suitable conductive epoxy is type SEC1 222 epoxy, manufactured by Resinlab, LLC of Wisconsin USA, which cures at room temperatures (approx.

25°C).

Alternatively, a conductive epoxy having a strongly anisotropic characteristic may be used. This is beneficial when the contacts on the fingerprint area sensor 8 are very close together because it provides the required conductivity between the fingerprint area sensor 8 and the contacts 12 in the circuit 6, whilst ensuring that even if the conductive epoxy flows between adjacent contacts 12, it will not form any appreciable conductive path between them.

Interior walls of the cavity are coated with an adhesive epoxy prior to the fingerprint area sensor 8 being inserted. The adhesive epoxy seals the fingerprint area sensor 8 in place.

The fingerprint area sensor 8 is then is then aligned with the cavity arid pushed into the cavity, such that the contacts on the fingerprint area sensor 8 and the contacts 12 in the circuit 6 are brought into electrical contact through the conductive epoxy.

The conductive epoxy and adhesive epoxy preferably cure without heating.

However, alternatively, one or both of the conductive epoxy and adhesive epoxy may require heat curing where the curing temperature of the conductive epoxy and/or adhesive epoxy is below a safe temperature of the fingerprint area sensor 8, for example below 60°C, which is a typical maximum operating temperature for capacitive sensors of the type used in the preferred embodiment. Higher temperatures may be possible for short time periods and/or for different sensor types.

It will be appreciated that the invention is not limited to the preferred embodiment described above and that various alternatives and variations will be within the scope of the claims. For example the fingerprint area sensor could be split into sections of different shape or size than the horizontal sections of Figure 4, such as sixty-four sections of two columns and 400 rows in place of sections having 128 columns and eight rows. The arrangement of Figure 4 is selected for convenience based on an existing array and existing controller set-up for capturing data from that array size, but it is not essential to use this arrangement. The fingerprint area sensor could also be split into sections in a checkerboard fashion.