AU2003244829A1

AU2003244829A1 - An automated teller machine system having function of digit photograph, record and recognition

Info

Publication number: AU2003244829A1
Application number: AU2003244829A
Authority: AU
Inventors: Chungmin Lee
Original assignee: WAYK CO Ltd
Current assignee: WAYK CO Ltd
Priority date: 2002-02-05
Filing date: 2003-02-08
Publication date: 2003-09-04
Also published as: JP2005526309A; CA2475315A1; WO2003069569A1

Description

VERIFICATION OF TRANSLATION I,. .......... Tim Le.e... (insert translator's name) of 8th Fl., 245 DunHua S. Road Sec.1, Taipei 106, .......................................

.

I........... ...... .... .. ... .... ........ ......... Taiwan, R.O.C. .................... ...... ** ....................

* ..... .................. .... ..... .. (translator's address) declare as follows: 1. That I am well acquainted with both the English and Chinese languages, and 2. That the attached document is a true and correct translation made by me to the best of my knowledge and belief of: (a) The specification of International Bureau pamphlet numbered WO 03/069569 International Application No. PCT/CNO3/00126 .Aug.v.2.. UP A........... (Date) (Signature of Translator) (No witness required) Automated Teller Machine System with Intelligent Digital Photography, Recording and Image Recognition BACKGROUND OF THE INVENTION 5 The present invention relates to banking automation systems, and more particularly to the security of ATM systems. Banking automation and unmanned banking are the most important subjects in promoting the productivity and competence of the banking, and are also the important subjects in raising the efficacy of the financial 10 system. To perform the automation and unmanned operations the banking has to face security problems. Usually happen many such cases such as using fake bank cards, unauthorized transactions with bank cards obtained illegally, destructing ATMs to steal bills, etc. Such problems make consumers place lower trust in banks, thereby impeding the 15 development of unmanned banking to provide 24-hour services for more competence and profits. SUMMARY OF THE INVENTION The ATM system of the present invention employs at least one digital camera, controls the photographing at various resolutions and different 20 sampling frequencies according to various operation modes, and recognizes the facial features of a bank card holder in transaction. The system provides the advanced and complicated bio-character or facial features recognition, or simply detects if the card holder masks its face or wears any helmet to determine whether to proceed with next step 25 transaction. The system changes the sampling frequencies such as low, middle or high frequency to improve the security of the system and -1 G:\Tim(TL)\SPE90537-WAYKPCT.doc transaction. In hardware aspect, in addition to a general automated teller machine, the ATM system of the present invention comprises an image input selection device, an A/D converter and synchrous processing device, a 5 filter, a video signal decoder and image scaling device, a data formatter, a data compressor, a DMA controller, an image data buffer memory and an interface circuit, wherein the image input selection device is controlled by a controller in the ATM or an external controller to perform the selection. The data compressor is also controlled by the controller in the ATM or the 10 external controller to determine whether to perform the compression. The image data of the facial features shot by different cameras at different positions and different sampling frequencies and resolutions are sent to the input selection device for selection under control. The selected image data are then processed via the A/D converter and synchrous 15 processing device, the filter, and the decoder and image scaling device to become elementary digital data. Next, the data formatter converts the format of the elementary digital data into a format acceptable to the controller, and the buffer memory stores the formatted digital data to be further transmitted via the DMA controller and the interface circuit to the 20 controller to activate the recognition process of the facial features. Additionally the formatted digital data are compressed before they are stored in the storage device in the ATM or in the external control/storage apparatus. The external control/storage apparatus comprises at least one 25 interface port able to communicate with at least one correspondent ATM, -2 G:\Tim(TL)\SPE90537-WAYKPCT.doc and the communication type between the external control/storage apparatus and each ATM can be wired or wireless. The external control/storage apparatus can be contained in the ATM or in a separated, solid and safe case or place. 5 The present invention also provides a method of intelligent digital photographing, recording and image recognition for the ATM system to improve the security of transaction. The method divides general ATM transactions into the following three operation modes: (1) transaction-free mode; (2) transaction-starting mode; and (3) bill-delivering/taking mode. 10 In the transaction-free mode the camera towards the card holder is set at the low sampling frequency to save the space of the storage device, and a regular equipment and security test of the system is performed to make sure of a proper and secure transaction to run. In the transaction-starting mode the camera towards the card holder is set at the middle sampling 15 frequency, and the image data from at least another camera are retrieved in an uncompressed manner to start the facial features recognition process to determine whether the recognition result is accepted. If yes, the system will proceed with the regular transaction and service. If not, the system will proceed with an irregular program to instruct the card holder by 20 sound, light or any mean to remove the barrier interfering with the facial recognition. The system allows the card holder to retry some predetermined times. If the recognition results still fail, the system will terminate the transaction eventually. In the bill-delivering/taking mode the camera positioned at top is set at the high resolution and sampling 25 frequency to provide its image data as evidence of transaction. Optionally -3 G:\Tim(TL)\SPE90537-WAYKPCT.doc the system can add a function or process of recognizing the amount and genuineness of the delivered bills. The system can access the image data stored in the storage device in a real-time manner according to the date, time or relevant account being 5 recorded during transactions. The system can change relevant preset threshold values for different service regions, features of customers, or time. Also the system can change relevant standards of facial recognition for different service regions, races, or time. 10 The ATM system includes an operation system for the intelligent photographing, recording and image recognition. In addition to general ATM functions the operation system can judge whether the functions of the intelligent photographing, recording and image recognition is normal. The operation system has the following three operation modes: (1) 15 transaction-free mode; (2) transaction-starting mode; and (3) bill-delivering/taking mode. In different modes the system activates cameras at different positions to shoot at different resolutions and sampling frequencies to save the space of the storage device. The captured image data of the facial features or bio-character of the card 20 holder can be used to perform a complicated recognition process by comparing them with those in the database to screen a particular person or to perform a simple process to identify if the card holder masks its face or wears any helmet to interfere with the recognition. The three operation modes are further described below: 25 (1) Transaction-free mode: the camera towards the card holder is -4 G:\Tim(TL)\SPE90537-WAYKPCT.doc activated and set at the low sampling frequency, and a regular equipment and security test of the system is performed. (2) Transaction-starting mode: the camera towards the card holder is set at the middle sampling frequency for transaction security, and the 5 image data from at least another camera are retrieved in an uncompressed manner to start the facial features recognition process to determine whether the recognition result is accepted. If yes, the system will proceed with the regular transaction and service. If not, the system will proceed with an irregular program to instruct the card holder by sound, light or 10 any mean to remove the barrier interfering with the facial recognition. (3) Bill-delivering/taking mode: the camera positioned at the upper side is set at the high resolution and sampling frequency to provide its image data as evidence of transaction. Optionally the system can add a function or process of recognizing the amount and genuineness of the 15 delivered bills. The operation system also has an image data access module allowing the image data stored in the storage device to be accessed in a real-time manner according to the date, time or relevant account being recorded during transactions. 20 BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 a shows the first embodiment of the system configuration of the present invention. Figure lb shows the second embodiment of the system configuration of the present invention. 25 Figure 2 shows the schematic block diagram of the hardware -5 G:\Tim(TL)\SPE90537-WAYKPCT.doc configuration of the present invention. Figure 3 shows the flow chart of the main operation modes of the operation system of the present invention. Figure 4 shows the schematic diagram of the image data access 5 module of the present invention. Figure 5 shows the flow chart of an embodiment of the continuous shooting procedure of the present invention. Figure 6 shows the flow chart of an embodiment of the snap shooting procedure of the present invention. 10 Figure 7 shows the flow chart of an embodiment of the image recognition procedure of the present invention. DETAILED DESCRIPTION OF THE INVENTION In Figure 1 a showing the first embodiment of the system configuration of the present invention, the system 100 comprises an 15 automated teller machine (ATM) 110, the circuit module of which has a built-in controller 111 including a storage device 112, as well as three digital cameras 120, 130 and 140, e.g. CCD cameras, placed at different positions and directions. Camera 120 towards the front of the ATM, i.e. towards the bank card holder in transaction, is essential for the 20 recognition of the facial features of the card holder. Camera 130 is positioned on the upper side of the ATM and towards the lower side of the ATM. Camera 140 is positioned towards the back side of the ATM. System 100 manipulates these three cameras to shoot at various resolutions and sampling frequencies according to different operation 25 modes, and based on captured image data, performs the recognition of the -6 G:\Tim(TL)\SPE90537-WAYKPCT.doc facial features of the card holder. System 100 provides an advanced and complicated bio-character or facial features recognition, or simply detects if the card holder masks its face or wears any helmet to determine whether to proceed with next step transaction. System 100 changes the 5 sampling frequencies such as low, middle or high frequency to improve the security of the system and transaction. The image data captured from the cameras are stored in the storage device 112. In Figure l b showing the second embodiment of the system configuration of the present invention, the controller 111 in Figure 1 a is 10 separated from the circuit module of the ATM 110 to form an external controller including the storage device 112 of Figure la. The external controller has at least one communication interface port, e.g. RS232, to communicate with at least one corresponding ATM in a wire or wireless type. The ATM system with the external controller operates in a way 15 similar to the ATM system with the built-in controller. The external controller can be contained in the ATM or in a separated, solid and safe case or place for security and convenient maintenance. In Figure 2 the hardware configuration of the present invention comprises an image input selection device 210, an A/D converter and 20 synchrous processor 220, a filter 230, a video signal decoder and image scaling device 240, a data formatter 250, a data compressor 260, an image data buffer memory 270, a DMA controller 280, and an interface circuit 290. The image input selection device 210 is controlled by the controller 111 or the external controller to perform the selection. The 25 data compressor 260 is also controlled by the controller 111 or the -7 G:\Tim(TL)\SPE90537-WAYKPCT.doc external controller to determine whether to perform the compression. The image data of the facial features of the card holder from the cameras of Figure 1 shot at different sampling frequencies and resolutions are sent to the image input selection device 210 for selection under control. The 5 selected image data are then processed via the A/D converter and synchrous processor 220, the filter 230, and the video signal decoder and image scaling device 240 to become elementary digital data. Next, the data formatter 250 converts the format of the elementary digital data into a format acceptable to the controller 111, and the buffer memory 270 10 stores the formatted digital data to be further transmitted via the DMA controller 280 and the interface circuit 290 to the controller 111 to activate the recognition process of the facial features. Additionally the formatted digital data are compressed by the data compressor 260 before they are stored in the storage device 112 or in the storage device of the 15 external controller. The ATM system comprises an operating system for the intelligent digital photographing, recording, and image recognition, which provides various transaction services according to the results of the facial features recognition. The operation system allows the image data stored in the 20 storage device to be accessed in a real-time manner according to the date, time or relevant account data being recorded during the services. Please refer to Figure 3. In addition to ordinary ATM functions, the operation system has the following three operation modes: (1) Transaction-free mode: the camera towards the card holder is 25 activated and set at the low sampling frequency (step 302), and a regular -8 G:\Tim(TL)\SPE90537-WAYKPCT.doc equipment and security test of the system is performed (step 304). If the test result is abnormal, the system will enter a preset process (step 308), such as suspending services, or skip the abnormal status to continue services but at the same time call maintenance. 5 (2) Transaction-starting mode: If the test result is normal, then judge if there is a request for transaction (step 306). If not, go back to step 302. If yes, start the transaction (step 310). In this step the camera towards the card holder is set at the middle sampling frequency for transaction security, and the image data from at least another camera are retrieved in 10 an uncompressed manner to start the facial features recognition process. The recognition process will determine whether the recognition result is accepted (312). If yes, the system will proceed with the regular transaction service (318). If not, the system will proceed with an irregular program to instruct the card holder by sound, light or any mean 15 to remove the barrier interfering with the facial recognition, and to increase a retry-count value (step 314). Then, check if the value meets the preset value (step 324). If not, go back to step 310. If yes, return the bank card to the card holder and terminate the transaction (step 316). (3) Bill-delivering/taking mode: the camera positioned at the upper 20 side is set at the high resolution and sampling frequency to provide its image data as evidence of transaction (step 320). Optionally the system can add a function or process of recognizing the amount and genuineness of the delivered bills. The regular equipment and security test in step 304 comprises the test 25 of the functions of the cameras, the left space of the storage device, and -9 G:\Tim(TL)\SPE90537-WAYKPCT.doc the functions of all the modules. The embodiment provides three correspondent processes, that is, the continuous shooting process when performing ordinary transactions, the snap shooting process when performing the facial recognition, and the 5 image recognition process. Referring to Figure 3, the continuous shooting process will occur in the step 302 with the low sampling frequency shooting, in the step 310 with the middle sampling frequency shooting, and in the step 320 with the high sampling frequency shooting. Referring to Figure 5, the continuous shooting process is illustrated in 10 more detail. First, receive commands from main application program (AP) and check if there is any new command being sent (steps 502 and 504). If not, go to step 520 to wait until the finish of time counting. If yes, close the old data file, recount the amount of pictures shot per second, and check if the cameras work properly (steps 506 and 508). If not, go 15 to the error process (step 510) to record the error message in the historical log file and reset the parameter to recover the main AP and send out the error message. If yes, set the time counting value according to the amount of pictures shot per second and open the appointed data file (step 512). Then, check whether the file opening is successful (step 514). If 20 yes, write relevant parameters into the command register area of the controller (step 516). If not, enter the error process (step 510). Following step 516, check whether the writing of relevant parameters into the command register area successful (step 518). If not, enter the error process (step 510). If yes, enter step 520, waiting until the finish of the 25 time counting. Then, write the compressed AVI image file from the -10 G:\Tim(TL)\SPE90537-WAYKPCT.doc DMA memory device into the storage device such as a hard disc (step 522). Then, check whether the writing of the image file is successful (step 524). If yes, go back to step 502 and wait next transaction. If not, enter the error process (step 510). 5 Another design, the snap shooting process, is shown in Figure 6, which is performed in the process of facial recognition in real time. First, receive commands from main AP and check if there is any new command being sent (steps 602 and 604). If not, go to the waiting loop to wait until a new command. If yes, start the subprogram of facial 10 recognition by first checking if the cameras work properly (steps 606 and 608). If not, enter the error process (step 610). If yes, write relevant parameters into the command register area of the controller (step 612). Then, check whether the writing is successful (step 614). If not, enter the error process (step 610). If yes, write the uncompressed AVI image 15 file from the DMA memory device into the storage device such as a hard disc (step 616). Then, check whether the writing is successful (step 618). If not, enter the error process (610). If yes, call the image recognition process 700 and retrieve the result of facial recognition (step 620), and then send the result of face recognition to main AP (step 622). 20 The image recognition process 700, referring to Figure 7, first will apply the differential process to the image data captured from the camera(s) to get rid of the drawbacks of the image data caused by the factors such as the rotation of the head of the card holder, the contrast of light from environment, etc. (step 710). Next, compare refined image 25 data to the standard templates of human face, which are vital parameters -ll G:\Tim(TL)\SPE90537-WAYKPCT.doc abstracted from general human (step 720). Then, transmit the result of comparison to the original calling program (step 730). Figure 4 the software system of the present invention further includes the image data access module, which can search and retrieve the desired 5 image data from the database in the storage device by the date, time and relevant account (steps 402 and 404). The format of the name for the image file includes fields for date, time and account-number, and its length is at least 30 bytes. The present invention has the following commands and formats for 10 the photographing control: 1) For the continuous shooting at the low, middle and high sampling frequencies for transaction security, the command and format are: C SHOT (AA, BB, CC, DD, filename) wherein AA, BB, CC, and DD are relevant numeric parameters of 15 the sampling frequency or resolution sent to camera 1, camera 2, camera 3, and camera 4, respectively, and if the camera is not selected, then a "00" parameter is sent. The filename is 30 bytes long with relevant date, time and account number or an assigned name included. For example: 20 C SHOT (0.2, 0.2, 0.1, 0.5, 020107XYZ (30 bytes)) which indicates the sampling frequency for camera 1 and camera 2 as one time per 0.2 second, for camera 3 as one time per 0.1 second, and for camera 4 as one time per 0.5 second, and the filename as 020107XYZ for saving image data. Besides the 25 returned error codes may be: camera fail, disk fail, disk full or -12 G:\Tim(TL)\SPE90537-WAYKPCT.doc image retrieving module (card) fail. 2) For the snap shooting when performing the facial recognition, the command and format are: S SHOT (AA, BB, CC, DD, filename) 5 The command is to interrupt said C SHOT command and send the system a facial comparison index value ranging from 00 to 99. The system can preset or adjust a threshold value to determine whether to continue the transaction. Although the foregoing embodiments have described the present 10 invention and its advantages, they are not to be construed as limitation. Various modifications and combinations are apparent to any person skilled in this art in light of the above teaching. It is thus intended to form the scope of the present invention with the appended claims. -13 G:\Tim(TL)\SPE90537-WAYKPCT.doc

Claims

1. An ATM (automated teller machine) system with intelligent digital photographing, recording and image recognition comprising: an automated teller machine, equipped with at least one digital 5 camera, and connected with a storage and control device for storing the image data captured from the digital camera; and an operating system for the intelligent digital photographing, recording and image recognition, operable to perform various ATM services based on the image data. 10

2. The ATM system according to claim 1, wherein the operating system includes the following operation modes: transaction-free mode: the camera towards the front of the ATM being activated and set at a low sampling frequency; transaction-starting mode: at least one camera being activated when a 15 bank card holder inserts the bank card, and the image data from said at least one camera being retrieved to start a facial recognition process to determine whether to proceed with the transaction; and bill-delivering/taking mode: the camera positioned at the upper side being activated and set at a high sampling frequency. 20

3. The ATM system according to claims 1, wherein the storage and control device is contained within the ATM.

4. The ATM system according to claim 1, wherein the storage and control device is placed externally to the circuit module of the ATM.

5. The ATM system according to claim 4, wherein the external storage 25 and control device is contained within the ATM or placed in a separate -14 G:\Tim(TL)\SPE90537-WAYKPCT.doc case.

6. The ATM system according to claim 1, wherein the communication between the ATM and the storage and control device is wired or wireless.

7. The ATM system according to claim 1, further comprising an image 5 data access module to access the image data stored in the storage device according to the date, time or relevant account being recorded during transactions.

8. The ATM system according to claim 2, wherein the facial recognition process is to recognize the face features of the bank card holder, and to 10 recognize whether the bank card holder wears any mask, helmet, or other object which interferes with the recognition.

9. The ATM system according to claim 2, wherein in the bill-delivering/taking mode, the amount and genuineness of the delivered bills are recognized according to the image data obtained from the camera 15 set at the high sampling frequency.

10. The ATM system according to claim 1, wherein the operating system is operable to access the stored image data in a real-time manner according to the date, time or relevant account being recorded during transactions. 20

11. An automated teller machine, in addition to the devices of a conventional automated teller machine, the automated teller machine comprising: at least one digital camera; an image input selection device; 25 an A/D converter and synchrous processing device; -15 G:\Tim(TL)\SPE90537-WAYKPCT.doc a filter, a video signal decoder and image scaling device; a data formatter; a data compressor; 5 a DMA controller; an image data buffer memory; and an interface circuit, wherein the input selection device selectively inputs at least one kind of image data from said at least one digital camera; the selected 10 image data are then processed via the A/D converter and synchrous processing device, the filter, and the decoder and image scaling device to become elementary digital data; the data formatter converts the format of the elementary digital data into a format acceptable to the main program of the system; the formatted digital data are 15 compressed by the data compressor and then stored in the buffer memory to be further transmitted via the DMA controller and the interface circuit to the main program of the system.

12. An automated teller machine, the automated teller machine communicating with an external storage and control device, and in 20 addition to the devices of a conventional automated teller machine, the automated teller machine comprising at least one digital camera, wherein said external storage and control device comprises: an image input selection device; an A/D converter and synchrous processing device; 25 a filter, -16 G:\Tim(TL)\SPE90537-WAYKPCT.doc a video signal decoder and image scaling device; a data formatter; a data compressor; a DMA controller; 5 an image data buffer memory; an interface circuit, and a data storage device, wherein the image data from said at least one digital camera are sent to said external storage and control device; the input selection device 10 selectively inputs at least one kind of the image data; the selected image data are then processed via the A/D converter and synchrous processing device, the filter, and the decoder and image scaling device to become elementary digital data; the data formatter converts the format of the elementary digital data into a format acceptable to 15 the main program of the system; the formatted digital data are compressed by the data compressor and then stored in the buffer memory to be further transmitted via the DMA controller and the interface circuit to the main program of the system.

13. The automated teller machine according to claim 12, wherein the 20 external storage and control device comprises at least one interface port operable to communicate with at least one automated teller machine.

14. The automated teller machine according to claim 12, wherein the external storage and control device is contained within any one of relevant automated teller machines with which the external storage and 25 control device communicates, or placed in a separate case. -17 G:\Tim(TL)\SPE90537-WAYKPCT.doc

15. The automated teller machine according to claim 12, wherein the communication between the external storage and control device and the automated teller machine is wired or wireless.

16. The automated teller machine according to claim 11 or 12, wherein at 5 least one digital camera is positioned towards the front, the lower side, or the back side of the automated teller machine.

17. An operating system for the intelligent digital photographing, recording and image recognition as claim 1, in addition to the operating system and modules of a general ATM, the operating system comprising 10 an intelligent digital photographing, recording and image recognition module, characterized in that the module comprises the following modes: (1) transaction-free mode: the camera towards the front of the ATM being activated and set at the low sampling frequency, and a regular equipment and security test of the system being performed; 15 (2) transaction-starting mode: the image data from at least one camera being retrieved to start a facial recognition process, if the recognition result is acceptable, proceeding with the regular transaction and service, if the recognition result is not acceptable, proceeding with an irregular program to instruct the card holder by sound, light or any mean 20 to remove the barrier interfering with the facial recognition, if the removal fails, returning the card; and (3) bill-delivering/taking mode: the camera positioned at the upper side being activated and set at the high sampling frequency to provide its image data as evidence of transaction. 25

18. The operating system according to claim 17, further comprising an -18 G:\Tim(TL)\SPE90537-WAYKPCT.doc image data access module to access the image data stored in the storage device according to the date, time or relevant account being recorded during transactions.

19. The operating system according to claim 17, wherein the facial 5 recognition process utilizes a differential process to get rid of the drawbacks of the image data caused by the factors such as the rotation of the head of the card holder, the contrast of light from environment, etc.

20. The operating system according to claim 17, wherein the recognition result forms an index value ranging from 0 to 99, and the operating 10 system presets a threshold value to determine whether the recognition result is acceptable.

21. The operating system according to claim 20, wherein if the index value formed by the recognition result does not meet the threshold value, the operating system increases a retry-count value, and the procedure 15 repeats until the retry-count value reaches a preset value.

22. The operating system according to claim 17, wherein in the bill-delivering/taking mode, the operating system provides a module to recognize the amount and genuineness of the delivered bills.

23. The operating system according to claim 17, wherein the facial 20 recognition process presets the facial comparison templates for various service regions or races.

24. An method of intelligent photographing, recording and image recognition for an ATM, characterized in that the method controls sampling frequencies for the photographing by each digital camera, and 25 presets a threshold value for the image recognition, the method -19 G:\Tim(TL)\SPE90537-WAYKPCT.doc comprising the following steps: (1) in transaction-free mode, activating the camera towards the front of the ATM, setting the photographing at the low sampling frequency, and performing a regular equipment and security test of the system; 5 (2) in transaction-starting mode, setting the photographing at the middle sampling frequency, retrieving the uncompressed image data of a bank card holder to start a facial recognition process to decide to proceed with regular or irregular transaction services; (3) in bill-delivering/taking mode, setting the photographing at the 10 high sampling frequency; and (4) accessing the image data at any time according to the date, time or relevant account.

25. The method according to claim 24, wherein the facial recognition process applies a differential process to promote the resolution of the 15 facial recognition.

26. The method according to claim 24, wherein the recognition result forms an index value, and an acceptable threshold value for the index value is preset.

27. The method according to claim 24, wherein in the transaction-starting 20 mode, if the recognition result does not meet the threshold value, proceeding with the irregular service, alerting the card holder by sound, light or video signal.

28. The method according to claim 27, wherein a retry-count value is increased, and the procedure repeats until the retry-count value reaches a 25 preset value. -20 G:\Tim(TL)\SPE90537-WAYKPCT.doc

29. The method according to claim 24, wherein in the bill-delivering/taking mode, the amount and genuineness of the delivered bills are recognized according to the image data obtained from the camera set at the high sampling frequency. 5

30. The method according to claim 24, wherein the preset threshold value is adjusted according to different service regions, features of customers, or time.

31. The method according to claim 24, wherein the standards of facial recognition is adjusted according to different service regions, races, or 10 time. -21 G:\Tim(TL)\SPE90537-WAYKPCT.doc