"System, method and computer program product for bioidentification and traceability of livestock by analysis of biometrical features of livestock" * * * TEXT OF THE DESCRIPTION Technical Field The present invention relates to techniques for the bioidentification and traceability of livestock by analysis of biometrical features of livestock, in particular by analysing the retinal vasculature. The invention has been developed with particular focus on its possible application to the bioidentification of bovine livestock and to the traceability of their meat even after their slaughter. For greater simplicity of illustration, in the1 remainder of the present description shall refer almost
-constantly • to said possible field of application. It
■will be appreciated, however, that the scope of the. invention is quite general and hence not limited to said specific context of application. By way of general foreword to the description of the prior art, of the problems set as the basis of the invention and of the solution proposed herein, it seems useful to summarise some essential characteristics of the technical context of the invention. Background Art Methods for identifying animals are very ancient and include rudimentary techniques, such as iron branding, or affixing medals, rings or other identifying marks on the animal. More recently, the need has emerged, especially due to the desire to certify the origin of meat cuts, to have available more secure identifying systems. Therefore several concepts have been introduced: bioidentity, relating to the set of methods implemented
to obtain the secure identification of an animal, as well as traceability, relating instead to the set of methods that allow to maintain traceability on the products derived from said animals, starting from the farm or zootechnical structure, through to the counter where the meats are sold. In particular, in the past few years systems that connect biometrical information obtained on the animal to its identity and to that of its meats. In particular, for this purpose a known method is to acquire the retinal image of the animal, which has a unique vascular print and to associate said vascular print to the animal's identifying data. However, this solution is poorly suitable for traceability, since 'this type of information obviously cannot be obtained on meat cuts. Moreover, bioidentification and traceability operations assume, in most case, legal relevance, requiring co-ordination with certification authorities and/or oversight agencies (Anti-Sophistication Squads, Local Health Care Firms, Ministry of Health and Agriculture, etc.). Description of the invention The object of the present invention is to provide a solution for the bioidentification of animals by the analysis of biometrical information, in particular retinal images, such as to overcome the intrinsic drawbacks of the solutions of the prior art whereto reference is made above. According to the present invention, said object is achieved thanks to what is specifically set out in the claims that follow. In particular, it will be appreciated that the present invention can be related to a method, a system, and also as a computer program product directly
loadable into the memory of a digital computer and capable of implementing the steps of the method of the invention when the computer program product is run on a digital computer. The bioidentification solution described herein allows to evaluate in integrated form the identifying information of the animal, checking them in a certified manner. Description of the Drawings The invention shall now be described, purely by way of non limiting example, with reference to the accompanying drawings, in which: Figure 1 shows a general diagram of the architecture of a bioidentification system according to the invention; - Figure 2 shows a block- diagram1 of a procedure ' suitable for being implemented. ι ;■ i.n. the present ••- invention. .■ ;. !.•.._•._ - Figures 3A, 3B, 3C, 3D, 3E, 3F- 3G, 3H, 31, 3L show an example of a second procedure suitable for being implemented in the present invention, Figure 4 shows a diagram of a detail of the system of Figure 1. Detailed Description of Illustrative Embodiments Figure 1 shows a general diagram of a system implementing the proposed method. Said proposed method substantially comprises three main subsystems: a subsystem 10 for the acquisition of retinal images and the creation of an identifying key; a subsystem 20 for data acquisition on ear tag and translation into text format by optical character recognition; a computer processing subsystem 30 for the certification of the acquired data.
Said subsystems comprise hardware modules driven by appropriate control software. The subsystem 10 for the acquisition of retinal images mainly performs the following functions: acquisition of an image of the retinal fundus of the animal ; identification of the features of said retinal fundus of the animal ; generation of an identifying key based on said features of the retinal fundus of the animal. Said subsystem for the acquisition of retinal images, schematically shown in Figure 1 and designated as a whole by the reference number 10, thus comprises an optical system for acquiring images 11, which in turn comprises a lens 18, to shoot a retinal image I of a bovine eye and a matrix 19 of. CCD (Charge Coupled. Devices) for the acquisition of' the ■ image captured by < said lens 18. Said matrix of CCD devices preferably has a resolution starting from 640x480 pixels, in colour. It is clearly possible to use a different sensor device from the CCD matrix, such as a matrix of silicon CMOS sensors. An image grabbing electronic device 12, preferably obtained with an FPGA circuit, receives the image acquired by the matrix 19 of CCD devices and performs a pre-treatment of said image, which is also sent to a monitor 13 to be displayed. Said image grabbing electronic device 12 operates receiving from the matrix 19 of CCD devices the image in the form of data sequences, or streams, and inserts them within its own memory, a so-called frame buffer. The retinal image I, pre-treated by the image grabbing electronic device 12, is then provided to a processing module 14 for the execution of a procedure 140 for the creation of an identifying key K. The subsystem for reading the ear tag 20 similarly
transmits its own images to said processing module 14 through said image grabbing device 12. To the processing module 14 is also associated a data management system 15, to manage a database DB 5 contained in memory means M, in order to store the identifying key K, or a file of biometric data C containing said identifying K, and to execute operations for comparison with said key K. The proposed system is then completed by a module 0 for communication with external systems 16, which communicates in a known manner with remote computer systems, via cable or in wireless manner. In particular, as shown in Figure 1, the module for '' communication with external systems 16 communicates 5 wireless by means of a mobile phone network 31 with a loiometric information server 32. Said biometric ' information server 32 comprises a general database DBG
~- r '- of biometric information about animals of different stock farms, which, as shall be illustrated in greater 0 ' detail below, corresponds to the database of a certification authority, which holds all the biometric data pertaining to the stock forms of a given area. The processing module 14 is therefore able both to load the acquired biometric information, which may have been5 temporarily stored in the local database, into said biometric information server 32, and consult and download information from a general reference database DBG of said server 32, and in this case it may also store said downloaded information in the local data 0 base DB. According to a preferred embodiment, the optical image acquisition system 11 comprises a group of lens assembly able to focus the retinal image I, which is positioned 2 cm behind the lens of the eye of a stock5 animal, as well as a soft diffusion lighting system by
means of a suitable lens positioned between the light source and the animal's eye. The retinal image I is acquired focusing the image behind the eye lens and creating, with the lighting system, a uniform veil on the eye region of interest, so that part of the light reaches the retinal fundus. Light intensity is adjusted manually and/or automatically by the operator and by the processing module 14 which also controls the graphic and dynamic evaluation of the image. Figure 2 shows a more detailed block diagram of a graphic treatment procedure executed by the processing module 14, which, as mentioned, receives the retinal image I, pre-treated by the electronic image grabbing device 12. A block 141 indicates a chromatic analysis procedure performed by said processing system 14 as a function of the lighting adjustment. Subsequently, a chromatic pre-treatment analysis of the image 142 is performed. A spatial pre-treatment of the image 143 is then performed. The image is then displayed on the monitor 13. The procedures described with reference to Figure 2 can be described in algorithm form, e.g. with VDHL language, and introduced in an FPGA (Field Programmable Gate-Array) circuit by means of a synthesis program. In this way the FPGA circuit embodies the processing system 14 and optionally also the image grabbing device 12. Within said FPGA circuit are preferably included also the circuits able to analyse the image and to adjust the lighting diffuser in the system 11. Since, as shown in the figure, the processing module 14 operates in association with a monitor 13, said FPGA circuit preferably also comprises the functionalities for managing said monitor 13.
Alternatively, the functionalities described above can be implemented using the standard electronic components of a computer, in particular of a palmtop computer, e.g. using the memory resources, the microprocessor and the LCD screen with which said palmtop computer is normally equipped. The processing module 14, as stated, provides for the creation of the identifying key K on the basis of the acquired retinal image I. This is accomplished with a procedure for the creation of an identifying key K of particularly suitable format for executing the search of a key amidst thousands of different keys, in addition to executing a comparison test. For this purpose, said procedure therefore creates a numeric key of suitable format to be broken down in. search procedures. Moreover, since the image I, when it is acquired, can be taken from different positions and angles, the, procedure for creating the identifying key K is able to minimise the effect of different dimensions and positions of the image. The proposed procedure for creating the identifying key K provides in general for making use of quantitative information obtained on the retina, such as: number of existing venous vessels dimensions of said vessels relationships between them The proposed procedure provides for selecting, among said information, information which cannot change from one shot to another, selecting an appropriate portion of the image. The portion of image to be treated is usually distributed for about 80% on the horizontal axis and for 20% on the vertical axis.
The procedure for creating the identifying key K here comprises the steps of : - converting the acquired image I into grey tones; - creating an image 21 of double size and white background; - identifying an end E of a major venous segment SVM; - rotating the image 21 relative to said end E; - cutting the image 21 around the major venous segment SVM; calculating a number NS of branches SVS of secondary venous segments . Figure 3A thus shows the operation of creating an image 21 of double" size. Subsequently, one the ends E of the major venous segment SVM is searched, as shown in Figure 3B. Once it is identified, the end E of said major venous segment SVM is used as a reference point around which the image 21 is rotated in order to align the image according to a horizontal axis X. The image 21 is then rotated around the end E by an angle A equal to the one formed by an axis S passing through the major venous segment SVM and the horizontal axis X. The image rotated along the horizontal axis is shown in Figure 3C, which also illustrates a rectangular template R which is applied to cut the image 21 around the major venous vessel SVM. Figure 3D shows the cut-out of the image 21. Subsequently, as shown in Figure 3E, where for the sake of more convenient representation the entire image 21 is shown, on the image are identified a first segment SI and a last segment SN, among the branches SVS which develop from the vessel that constitutes the major venous vessel SVM, and a distance DIN between said first segment SI and last segment SN is
calculated. The distance DIN is normalised to the dimension of the image 21, i.e. it is expressed as a fraction, e.g. of the length of the acquired image. Usually, said length of the image is constant from acquisition to acquisition. In the case of the example, the measured distance DIN is assumed to be equal to 160. The distance DIN is used as a first part Kl of the identifying key. Hence, Kl=[160]. Subsequently, a number of segments present respectively in the upper part, designated by the reference NSU, and in the lower part, designated by NSD, of the image 21 with respect to the horizontal axis X, is calculated. This information constitutes a second part K2 of the key K, thus equal to [NSU, NSD], i.e. in the example shown in Figure 3F [4,5] . The identifying key K until this point of the method is thus composed by the first part Kl , and by the second part K2 , i.e. [NSU, NSD] [DIN] , in the example [4, 5] [160] . At this point, the area of analysis of the image is narrowed to a rectangle R which contains all the branches SVS counted previously, i.e. a rectangle R whose side has a length of DIN. In this rectangle Rl are therefore calculated thicknesses TI and T2 of two ends El and E2 which are on the part of major venous segment SVM included in said rectangle Rl . Said thickness TI and T2 constitute a third part K3 of the identifying key K. In the illustrated example, said thicknesses TI and T2 have respective values of 24 and 29, in terms of virtual points, so that K3 is [TI, T2] , i.e. the values [24,29].
Subsequently, as shown in Figure 3EH, the major venous segment SVM is followed, scanning the branches SVS which depart, both in the lower part and in the upper part of the image 21 with respect to the horizontal axis X, coinciding with the main axis of the major venous segment SVM. For every identified branch SVS, the corresponding width or thickness T is calculated, then starting from the centre of said branch SVS, which corresponds to a measured point P, a segment SP is created, projected, as shown in Figure 31, on the lower or upper edge of the rectangle R. The coding of a fourth part K4 of the identifying key K is calculated by processing different values, based on the positions of said projected segments SP. With reference to Figure 3L, ^every measured point P on the image is defined in the following way: [DP, T, DS] , • , ■ ' where DP is a distance from the previous point, T the thickness of the base, DS a distance from the subsequent point. As mentioned above, their reference positional values are calculated as virtual points given by calculations in proportion to the size of the image . If there are multiple branches SVS, the fourth part K4 , to avoid redundancies, is expressed as a sequence where the distance from the subsequent point DS becomes the distance from the previous point DP of the next point P. For the first and last point P measured in order, the values for the distance of the previous point DP and to the distance from the subsequent point are set to zero. With respect to the measurements provided in the example shown in Figure 3L, the following values can be
had, respectively for the upper part and the lower part : [0,7,34,3,22,2,46,2,0] [0,8,30,5,36,1,46,1,18,2,0] The values of these sequences are then added in an upper sum SMU and a lower sum SMD. [0,7,34,3,22,2,46,0] = 114 [0,8,30,5,36,1,46,1,18,2,0] = 147 The sums SMU and SMD constitute the elements of the fourth part K4 of the identifying key. Therefore, the identifying key K is defined as K2K1K3K4, i.e., [NSU, NSD] [DIN] [TI, T2] [SMU, SMD]. In the described example, the numeric value of the identifying key is thus [4,5] [160] [24,29] [114,147]. The different parts Kl, K2 , K3 and K4 of the key K can be used to address the search in the database DB, e.g. using the second part K2 to perform a preselection on the retinas that have 4 top segments and 5 bottom segments, then selecting those whose distance DIN between the first and the last segment is equal to
160. Hence, the identifying key K is subdivided into sub-keys whereto are associated respective measures of different retinal features, to allow search operations in the databases on subsets of files C according to said sub-keys. This can have particular importance considering that the reference database DBG can also contain many millions of biometric data files C. It should be noted that in the procedure described above, the quantification of the information aimed at identifying the features to be treated to obtain the identifying key K essentially comprises the coordinates of the venous segments and their thicknesses. The thickness of a retinal venous segment, however, can change as a result of several factors, e.g.
diabetes or hypertension. To overcome this problem, in an alternative embodiment of the procedure for creating the identifying key K, just described above, the second 5 part K2 can be maintained related to the number of segments respectively in the upper part, NSU, and in the lower part, NSD. For the remaining three parts of the key, the following measurements are respectively used: -10 the distances between the first and the last branch SVS present respectively in the upper part and in the lower part of the image 21; the distances between the adjacent measured points P in the upper part of the image 21; -15 the distances between the measured points P in the lower part of the image 21. In this way the value assumed by the identifying "key K-does not depend on the thickness of the major venous segment SVM or on the thickness of the branches 20 SVS . Once generated, with the procedure described above, the identifying key K from the retinal image, said image can be used to create the aforementioned biometric data file C associated to the same animal 25 whereto the identifying key K is associated, or for an identity search function among existing biometric data files C. Said biometric data file C associated to the animal is stored in the local database DB associated to the processing module 14, in the memory M which can be 30 a magnetic mass memory or a Flash memory. If the processing module 14 is obtained from a palmtop computer, clearly the management functions of the processing module 14 and of the management system 15 shall be obtained from the operating system, e.g.
35 Windows CE, whereas in case of a processing structure
dedicated by logic on FPGA they may be implemented at the BIOS level. The module for communication with external systems
16, for communicating with remote computer systems, via cable or in wireless manner, in particular can comprise an USB port or an Ethernet port . To the processing means 14 can optionally be associated a GPS module, which allows, upon the creation of the identifying key K, to enter the corresponding values of position of the stock farm in which the biometric information was taken. Said GPS module can be provided with an appropriate antenna, obtained on a printed circuit board. The biometric data file C in the data base DB preferably comprises the following data: code of the animal ; code of the stock farm; description; identifying key K creation date; identifying key K; positional information (GPS) ; notes ; transfer date. According to the proposed method, when the data of a new animal are to be acquired, the data of the stock farm whereto said animal is to be associated are selected or entered. Then the optical image acquisition means 11 are positioned about 30 centimetres from the animal's eye. As mentioned previously, within the procedure for the creation of the identifying key K, at this point a step of centering the retinal image is followed by the acquisition of the image I by means of a command activated through the processing means 14 or directly on the optical means 11.
The position data may also be acquired automatically and simultaneously through the GPS module. On the monitor 13 is then displayed the biometric data file C already containing the data acquired automatically such as the identifying key K and the position data, which can be complemented with the remaining information. Subsequently, the file C is stored in the database DB under the control of the management module 15 implementing an appropriate management program. The program for managing the database DB shall also be equipped with functions for modifying the data entered in the file C, and, especially with a test function, able to allow the search of a specific stock animal on the basis of an image I acquired through the subsystem 10 and to the identifying key K that derives therefrom. This can be accomplished accessing the local database DB on the basis of the identifying key K acquired at a previous time and performing the search, or acquiring the retinal image of the animal at that moment, obtaining the identifying key K and performing a search among the identifying keys K already stored in the database DB. It is clear that said operations, as shall be better specified below, can be performed alternatively on the general database DBG of the server 32, either in real time whilst the retinal image is being acquired, or at a later time, storing the information temporarily in the local database DB. In the proposed system, to the processing module 14 can optionally be associated a programming system for smart card as well as a corresponding smart card reading and writing device 33, in order to write data from the file C of an animal into the smart card itself. In this way, the stock animal is provided with
a transportable electronic document, which can be associated to the animal itself for subsequent checks. As mentioned above, a relevant aspect of the proposed method and system provides for using the subsystem for reading the ear tag 20, to read an ear tag U, shown in Figure 4. Said ear tag U comprises the code of the animal, expressed with an alphanumeric string UI and an optically readable bar code U2. The subsystem 20 therefore comprises a supplementary lens 18' able to acquire the image of the ear tag U as well as a corresponding matrix 19' of CCD devices, connected to the image grabbing electronic device 12. The image of the ear tag U is provided to a character recognition program 240 loaded in the processing module 14, returning the value of the data 'recorded on the ear tag U,- which is inserted in the file C of the animal . The ear tag reading subsystem 20 thus substantially comprises the supplementary lens 18' and the corresponding matrix 19' of CCD devices, whilst it shares with the subsystem 10 resources such as the image grabbing device 12 and the processing module 14, to run a character recognition program and update the database DB. Said supplementary lens 18' can be obtained from the lens of a normal television camera with focal distance 35 to 75 mm and associated to a corresponding matrix 19' of CCDs with high resolution, e.g. 1320x1024 pixels. To the supplementary lens 18' can be associated an infrared lighting device, of such power as to allow shooting at a distance of about 3 metres. Preferably, the character recognition program is based on a system with neural networks, alternatively implemented via software in the palmtop or by means of
neural networks which can be implemented by means of software algorithms or hardware processors in the FPGA embodiment . Thus, the procedure for acquiring the data of a new stock animal provides in this case, in a first step of aiming the supplementary lens 18' on the ear tag U, for acquiring its image and decoding the value recorded in the alphanumeric string UI or in the bar code U2. In a second step, the optical means 11 are aimed at the animal's eye, the retinal image is acquired and the identifying key K is created. The expected cases in which a check of the identity of stock animals is performed are the following : - birth and acquisition of new biometric data; ■identification checks during the animal's lifetime; : - identification check and consequent granting or denial of the authorisation to slaughter. In all these cases, in general, an operator visits a stock farm or a slaughter house to acquire the biometric information and consequently a transaction will take place, i.e. an exchange of data corresponding to the creation or modification of one or more files C. While the image acquisition optical system 11 may already be available in the site where said information is to be obtained, e.g. in the case of a slaughter house a so-called biometric gate may be available, i.e. a television camera positioned in a corridors where animals to be slaughtered transit and stop, the operator may be provided with the palmtop computer incorporating the processing module 14, which, in particular is associated to the certification subsystem 30 with legal value by means of digital signature, i.e. the association to each transaction of a digital
certificate DC containing information identifying the operator and the originating origin. Said certification subsystem 30 can be implemented, with respect to the digital signature, for instance by means of a smart card system, e.g. exploiting the smart card reading and writing system 33 to store the aforementioned files C. The processing module 14, e.g. of the palmtop computer, is able, through the communication module 16, to communicate in wireless fashion through the mobile telephony network 31 with the biometric information server 32 and the general database GDB of biometric information pertaining to animals from different stock farms, contained therein. The processing module 14 can therefore both load the information acquired on said biometric information. server 32, and consult and download information rom said server 32. ■ Said biometric information server 32 shall also be provided with procedures ■ and•■ functionalities linked to the digital signature, homogeneous- to the procedures available to the operators. Therefore, both the operators and the server 32 may, as stated, make transactions only with a digital signature having legal value, constituting a closed group of users of the bioidentification system. At least the two following cases are envisaged:
- upon the operator's request, the server 32 enables the acquisition on the palmtop of the data needed by the operator to check a stock farm, associating a digital certificate DC,- the operator acquires the biometric data at the stock farm and then verifies the data thus obtained, issuing a corresponding digital certificate DC, connecting to a server 32. Of course, this can also be performed via remote wireless connection from the stock farm itself.
The two-way passage of data between processing module 14 and server 32 is performed in certified mode, issuing digital certificates DC, so that a certification authority, whereto the server 32 is associated, certifies the exchange of data with the different users of the system. If a non-conformity in the bioidentity is detected, the certification authority will be able automatically to open a procedure to record the occurrence of said non-conformity and initiate the traditional oversight activities and search for responsibilities and causes. More in detail, the certification subsystem 30 executes a certification procedure by mutually associating the fundamental fields in the files C, i.e. the ear tag U and the retinal identifying key K. Said association step is performed entering an animal authentication code UK in the' file C. The animal authentication code UK is generated according to the other two codes acquired on the animal, the ear tag code U and - the identifying key K and therefore it constitutes the liaison between said codes U and K obtained by the optical image acquisition system 11. The animal authentication code UK operates similarly to a checksum code with respect to the other two identifying codes U and K. The animal authentication code UK finds application in the issue of the aforementioned digital certificate DC, in particular according to the standard ITU-T- X.509, on the occasion of each transaction completed both with respect to bioidentification operations, and more in general relative to subsequent operations involving the meat cuts, to implement certified traceability functions.
In an embodiment of the certification procedure, to reach the compilation of the data that allow the creation of a digital certificate DC with X509 standard, said procedure comprises one or more of the following three steps: - a step for identifying the certifying operator performed when the system is turned on; a step of identifying the certifying structure performed during the loading of the files C on the general database DBG in the server 32; - a step of identifying the completed transaction. For example, the operator that accesses the acquisition subsystems 10 and 20, interfaces its own palmtop computer representing the processing module 14, provided with the smart card reader 33, through which the identity of the operator the performs the measurements is entered in the system. The operator " acquires the identifying key K and .the ear code U of a stock animal; the authentication code UK is calculated by means of a calculation procedure with symmetric keys, similar to the one used for PGP keys for the digital signature, for the issue of a DC certificate that accompanies the transmission of the corresponding biometric data C to the server 23 and the reference database DBG. Anyway said authentication code UK is used for the creation of the DC certificate of every completed operation used to manage the traceability, also subsequently, of the meats corresponding to the stock animal. Therefore, the reference database DBG according to the described procedures essentially contains: - the ear code U; - the identifying key K; - the authentication code UK;
Within said reference database DBG or in another database or server connected to the certifying authority, it will thus be possible to define a database of the traceability movements, containing in particular the information on the certified operations, in which, in addition to the aforementioned codes, for each operation is present the corresponding X.509 digital certificate DC. Thus, the signature and the digital certification is applied in all transactions between the server 32 of the certification authority and the processing devices, in particular palmtop computers, provided to operators. A relevant aspect of the proposed method is constituted, moreover, by the association of the biometric reading of the retinal image to the reading of the - ear tag: this advantageously enables to have the certainty that the animal present in the stock farm is the one recorded traditionally according to national regulations. The biometric data shall therefore be associated to the traditional ones prescribed, by the regulations . The conjugation of alphanumeric or bar codes obtained from the ear tag, which can be transferred by means of stamps on the meat cuts obtained from the slaughter of the corresponding stock animal, with the identifying key obtained from retinal analysis advantageously enables to facilitate every traceability procedure, introducing an authentication code, linked to the ear tag and to the key derived from the biometric information. Said authentication code is used to issue digital certificates relating to each traceability procedure. Note that this links with the ear tag is particularly important, because it currently is the only information legally recognised for purposes of the traceability of the meats.
Of course, without altering the principles of the invention, the construction details and the embodiments may be widely varied from what is described and illustrated herein, without thereby departing from the scope of the invention. In this perspective, it should once again be recalled that, although for the sake of simplicity of illustration the present description has made nearly constant reference to the possible application of the invention to a bovine livestock farming context, the scope of the invention is wholly general and hence not limited to said specific application context. ; In particular, it is possible to use the system in a simplified version to acquire the ear tag code along ' and to employ said ear tag code directly as a key for the creation of the digital certificate to be associated to the transactions. Such a simplified system would allow to exploit the proposed system in those cases where the use of biometric information is ' not proposed.