Embedded information identification method and system

Abstract

The present disclosure generally relates to a method for identifying predefined components embedded in target object image data. According to the present disclosure, implementation is achieved by applying a machine learning-based scheme arranged to identify noise components from an image showing a target object. The disclosure also relates to a corresponding computer system and computer program product.

Description

Technical Field

[0001] The present disclosure generally relates to a method for identifying predefined components embedded in image data of a target object. The present disclosure is achieved by applying a machine learning-based scheme arranged to identify noise components from an image showing the target object. The present disclosure also relates to a corresponding computer system and a computer program product. Background Art

[0002] For a long time, unauthorized copying of physical objects has been a persistent problem, such as copyrighted paintings, photographs, and analog audio tapes, as well as packaging bags, fabrics, etc. Watermarking technology has been proposed as a security means to distinguish between the original and the copy. Over the years, a variety of different technologies have been developed, but usually rely on embedding an identification code in the physical object.

[0003] Preferably, the identification code is imperceptible to a human observer, or at least does not damage the visual appearance and look of the physical object. However, if the identification code is "hidden too thoroughly", its identification process may also become complex, making the process of verifying the authenticity of the physical object unreliable. On the other hand, if the identification code is too obvious and easy to identify, it may also be "too easy" to be successfully copied. The best balance is achieved when the watermark strength of the package is increased to just below the human eye visible threshold.

[0004] US Patent Application US2020311505 proposes a solution to the above problem. US2020311505 specifically proposes to form a stylized version of a digital watermark signal based on an input digital watermark signal image and one or more source images. The stylized version of the watermark signal is composed of pixel blocks from the target image, forming a collage or mosaic effect. For a human observer, the stylized image is more similar to the source image than to the watermark signal; however, for a compatible digital watermark decoder, the stylized image is recognized as a signal carrier carrying an encoded multi-symbol payload.

[0005] However, although the solution proposed by US2020311505 has a positive effect on watermarking technology, it still relies on recognizable modified content, which can be recognized and, if necessary, copied as long as a suitable type of decoder is used. Therefore, there is still room for further improvement in watermarking technology to enhance the security of the watermark and further increase the difficulty for unauthorized parties to identify the watermark. Summary of the Invention

[0006] According to one aspect of the present disclosure, there is provided a method for identifying a predefined component embedded in target object image data, wherein a computer system includes a processing unit, and the method includes the following steps: the processing unit receives the image data; through the processing unit, applying a machine learning-based scheme to the image data to identify noise components from the image data, wherein the noise components are embedded in the target object; through the processing unit, based on recognizable features in the noise components, extracting the predefined component from the identified noise components; and through comparison with pre-stored data in a storage element arranged to communicate with the processing unit, identifying the predefined component by the processing unit.

[0007] The present disclosure is based on the recognition that by embedding a watermark (defined above as a predefined embedded component) as a noise component in an image of a target object and generally not being visually perceived by the naked eye of a person observing the target object, the security of the watermark can be further improved. When the watermark is embedded as noise, the user will not perceive it as a change in the target object image but only as a normal image. However, the previously proposed existing decoders will not be able to identify relevant information from the image data because the predefined component will be well embedded.

[0008] According to the present disclosure, for the above-mentioned existing decoder problem, it is processed by applying a machine learning-based scheme to the image data. The machine learning-based scheme is preferably trained for different relevant noise-based predefined components corresponding to the noise components embedded in the currently processed target object image data, but not necessarily exactly the same. Once the training is completed, the machine learning-based scheme will be able to identify at least one specific type of noise component. Then, this noise component can be modified within a predefined range to form a large number of different recognizable noise components for use with the target object.

[0009] The term "noise component" should be broadly interpreted within the scope of the present disclosure. For example, the noise component can include a predefined pattern that may be modified according to a predefined scheme, a predefined frequency component that may be modified within a predefined range, and the like.

[0010] According to the definition of the present disclosure, applying a machine learning-based scheme to correctly identify the noise components in the image data. As described above, it is generally desirable to ensure that the machine learning-based scheme has been "trained" to be able to quickly identify different noise components of a specific type, for example, by training the machine learning scheme using multiple different image data including known noise components.

[0011] However, the training does not have to be done for each processing unit, but can be done in advance in a general way when developing a machine learning-based solution. In addition, it should be understood that a machine learning-based solution can be implemented using a combination of one or more different machine learning algorithms, including neural networks in deep learning, including artificial neural networks (ANNs), such as but not limited to convolutional neural networks (CNNs), feedforward neural networks (FNNs), etc. A machine learning-based solution can also be arranged to include a machine learning pipeline, where the machine learning pipeline can be further arranged to include multiple autoencoder components.

[0012] In some embodiments, the step of extracting a predefined component from the identified noise component may include a process of identifying recognizable features from the identified noise component. The noise component can thus be regarded as, for example, a "fingerprint", and the process of extracting a predefined component from the identified noise component can be regarded as a feature extraction process, in which specific types of features are identified in the noise component. Of course, other schemes for extracting a predefined component from the identified noise component are also possible and fall within the scope of the present disclosure.

[0013] It should be understood that in some embodiments, the steps of identifying a noise component and (from the identified noise component) extracting a predefined component can be performed by a processing unit as a single step.

[0014] When performing the step of identifying a predefined component by comparing with pre-stored data, it is generally desirable to arrange the pre-stored data in some form of storage element, such as included in a database. Thus, the processing unit can be arranged to perform a comparison operation by trying to find the best match (or candidates) between the predefined component and other components stored in the database. Of course, the database can also be arranged at a location remote from the computer system and / or the processing unit, thereby allowing the database to be shared among a large number of different processing units.

[0015] The present disclosure can be applied to different fields. Thus, in some embodiments, the solution according to the present disclosure can be used, for example, for digital watermarking of images, i.e., the target object is defined as an image, preferably a digital image. Such images can be, for example, artistic photographic works, etc., where a preselected noise component is combined with the image to form a "noisy image".

[0016] However, the solution according to the present disclosure is also highly useful in the case where the target object is a physical object. Such a physical object can be any physical object suitable for allowing a predefined component to be embedded, i.e., the predefined component is embedded in the physical object to some extent.

[0017] Thus, according to the present disclosure, the method can be further adjusted to include the step of embedding a predefined component into such a physical target object. The process of embedding the predefined component can, for example, include embedding a preselected noise component into the target object, such as by applying the preselected noise component to the surface of the target object.

[0018] For example, the surface can be processed to include the noise component. It should be understood that the selection of the noise component should ensure that the device used to capture the image data can process the noise component according to the solution of the present disclosure. For example, the resolution of the image capture device must be selected such that the image data of the target object allows the noise component to be recognized.

[0019] However, according to the present disclosure, different solutions may be needed to embed the predefined component into the physical target object based on the type of the target object. For example, a target object mainly composed of fabric may require a different solution for embedding the predefined component compared to a target object mainly composed of metal.

[0020] When the target object is a physical target object, the image data of the target object is captured / obtained and processed according to the solution defined in the present disclosure to identify the predefined component. Then, the identification of the predefined component can be used, for example, to verify the authenticity of the target object.

[0021] Preferably, confirming the authenticity of the target object can also include determining whether the identified predefined component has a defined relationship with the target object. That is, for example, if the target object is a watch and the identified predefined component is defined as being related to a shoe, it can be determined that the identified predefined component is incorrect and the authenticity of the target object does not meet the standard. Thus, the image data of the target object can be used to determine the type of the object, such as determining whether the target object is a watch or a shoe. Of course, the granularity can be further increased such that the computer system can determine the type or brand of a watch, a shoe, a bag, etc. In some embodiments, the processing unit is thus adapted to include a module for such determination, wherein the module can also implement a machine learning solution that has been trained for different types of objects. Other unique features of the target object can also be used in the process of determining the authenticity of the target object.

[0022] According to another aspect of the present disclosure, there is provided a computer system adapted to identify predefined components embedded in image data of a target object. The computer system includes a processing unit, where the processing unit is adapted to receive the image data, identify noise components from the image data by applying a machine learning-based scheme to the image data, where the noise components are embedded in the target object; extract the predefined components from the identified noise components based on recognizable features in the noise components; and identify the predefined components by comparing them with pre-stored data in a storage element communicating with the processing unit. This aspect of the present disclosure provides advantages similar to those of the foregoing aspects of the present disclosure.

[0023] Preferably, the computer system further includes an object capture device for capturing the image data of the target object. However, the object capture device is not limited to including an image capture device (i.e., a camera). Instead, depending on the type of the target object and / or the structure of the predefined components, other types of object capture devices are possible. For example, if the predefined components are engraved on the surface of a physical target object, devices such as lidar, radar, laser scanners, etc. can be used. Of course, other existing and future sensor systems are possible and fall within the scope of the present disclosure. Of course, multiple sensors can also be combined with the object capture device, such as an image capture device and lidar.

[0024] In some embodiments, it may be desirable to implement the machine learning-based scheme as a supervised machine learning process. Since the machine learning process is supervised, the operator / user can "correct" decisions made by the machine learning scheme that are considered incorrect. However, it should be understood that, contrary to the above, the machine learning-based scheme can also be implemented as an unsupervised machine learning process, thus allowing for fully autonomous implementation in identification and determination. Of course, in line with the content of the present disclosure, a hybrid use of supervised and unsupervised is also allowed, for example, depending on the implementation status of the image processing scheme, such as allowing the machine learning-based scheme to be initially supervised and then switched to unsupervised, or vice versa.

[0025] As described above, the database and the computer system can be remotely arranged. The computer system can also be a so-called cloud-based computing system in some embodiments, where the processing unit is part of a so-called cloud server. Therefore, the computing power provided by the present disclosure can be distributed among multiple processing units or servers, and the location of the processing unit / server does not have to be explicitly specified. The advantage of adopting a cloud-based solution also lies in the ability to achieve inherent redundancy and the ability to apply more complex machine learning processes compared to using a single embedded processing unit.

[0026] However, it should also be understood that the computer system can be provided as part of an electronic user device. Such an electronic user device can be, for example, a mobile phone or a similar device. Thus, one or more object capture devices of the mobile phone can be used to obtain, for example, image data of a physical target object, and the processing unit in the mobile phone can be used to identify predefined components embedded in the target object. Therefore, the mobile phone can be used to perform watermark processing "instantly", that is, to identify the watermark (i.e., the predefined component) embedded in the target object. The user can, for example, take a photo of the target object and obtain (almost) instant confirmation of whether the target object is a replica or an "original" object. Thus, the user can use the solution of the present disclosure to verify the target object, thereby minimizing the risk of fraudsters providing illegal replicas.

[0027] According to another aspect of the present disclosure, there is provided a computer program product including a non-transitory computer-readable medium having stored thereon computer program means for controlling a computer system adapted to identify predefined components in image data of a target object, the computer system including a processing unit, wherein the computer program product includes code for receiving the image data by the processing unit, code for the processing unit to identify noise components from the image data by applying a machine learning-based scheme to the image data, wherein the noise components are embedded in the target object; code for the processing unit to extract predefined components from the identified noise components based on recognizable features in the noise components; and code for the processing unit to identify the predefined components by comparing with pre-stored data in a storage element communicating with the processing unit. This aspect also provides advantages similar to those of the foregoing aspects of the present disclosure.

[0028] According to the present disclosure, the software executed by the processing unit can be stored on a computer-readable medium, which can be any type of storage device, including a removable non-volatile random access memory, a hard disk drive, a floppy disk, a CD-ROM, a DVD-ROM, a USB storage, an SD memory card, a solid state drive, other non-volatile flash storage media, or similar computer-readable media known in the art.

[0029] In summary, the present disclosure generally relates to an innovative concept for identifying predefined components embedded in image data of a target object, wherein the computer system includes a processing unit, and the method includes the following steps: the processing unit receives the image data; the processing unit applies a machine learning-based scheme to the image data to identify noise components from the image data; the processing unit extracts predefined components from the identified noise components; and the processing unit compares the predefined components with pre-stored data in a storage element communicating with the processing unit to identify the predefined components.

[0030] According to the present disclosure, by applying a machine learning-based scheme to identify noise components from image data representative of a target object, predefined components are extracted from the identified noise components. The predefined components are then compared with other objects to determine the identity of the predefined components.

[0031] Other features and advantages of the present disclosure will become more apparent upon reading the appended claims and the following description. Those skilled in the art should recognize that, without departing from the scope of the present disclosure, different features of the present disclosure can be combined to form other embodiments in addition to those described below. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Various aspects of the present disclosure, including its specific features and advantages, will be more readily understood from the following detailed description and the accompanying drawings, in which: Figure 1 Conceptually shows a computer system according to a currently preferred embodiment of the present disclosure; Figure 2A –2C shows different physical target objects, which include embedded predefined components conforming to the present disclosure; Figure 3 Is a flowchart showing the steps of performing a method according to a currently preferred embodiment, and Figure 4A and 4B Schematically shows a possible implementation of a machine learning-based scheme used in conjunction with the present disclosure. DETAILED DESCRIPTION

[0033] The present disclosure will be described more fully with reference to the accompanying drawings, which show currently preferred embodiments of the present disclosure. However, the present disclosure may be implemented in many different forms and should not be construed as limited to the embodiments described herein; rather, these embodiments are provided for thoroughness and completeness and are intended to fully convey the scope of the present disclosure to those skilled in the art. Like reference numerals denote like elements. The following examples are used to illustrate the present disclosure and are not intended to limit the scope of the present disclosure.

[0034] Now turning to the drawings, specifically Figure 1 , in which a computer system 100 is conceptually shown, adapted to identify predefined components embedded in image data of a target object 102. In Figure 1 , the computer system 100 is shown as a smart phone. It should also be noted that, as an alternative, the computer system can be any other type of portable electronic device, such as a laptop computer, a tablet computer, or any other type of currently or future similarly configured device. The target object 102 is shown in Figure 1 as a watch, where the predefined components have been embedded in the region 103 of the watch 102, particularly on the dial (or face) of the watch 102.

[0035] The smart phone 100 includes a display unit 104 having a touch screen interface, and at least one image capture device (such as a camera) 106. Preferably, and as will be apparent to those skilled in the art, the smart phone 100 also includes a first antenna for WLAN / Wi-Fi communication, a second antenna for telecommunication, a microphone, a speaker, and a phone control unit. Of course, other hardware elements may also be included in the smart phone.

[0036] The smart phone 100 further includes a processing unit 108, which is arranged to communicate with the display unit 104 and the camera 106. By way of reference, the processing unit may be embodied as a general-purpose processor, a graphics processing unit, an application-specific processor, a circuit including processing components, a set of distributed processing components, a set of distributed computers configured to perform processing, a field-programmable gate array (FPGA), etc. The processor may be or include any number of hardware components for performing data, signal, and / or image processing or executing computer code stored in a memory. It is also possible and within the scope of the present disclosure to implement using a system-on-chip (SOC). The memory may be one or more devices for storing data and / or computer code to complete or facilitate the various methods described in the present disclosure. The memory may include volatile memory or non-volatile memory. The memory may include database components, object code components, script components, or any other type of information structure to support the various activities described in the present disclosure. According to an exemplary embodiment, any distributed or local storage device may be used in the systems and methods of the present disclosure. According to an exemplary embodiment, the memory communicates with the processor (e.g., via a circuit or any other wired, wireless, or network connection) and includes computer code for executing one or more of the processes described herein.

[0037] As described above, in certain embodiments, the present disclosure may utilize the smart phone 100 to assist a user / operator in determining whether a target object 102 is indeed an authentic original or an illegal copy. According to the present disclosure, an authentic original is defined as a target object 102 that includes correctly matched predefined components, i.e., the predefined components are associated with the target 100 to some extent. For example, this relationship can be determined by requesting from a database which predefined components are embedded together with which type of target object 102.

[0038] Now turning to Figures 2A to 2C , three different target objects 202, 204, and 206 according to the present disclosure are shown.

[0039] The target object 202 is in Figure 2AShown is an automotive electrical connector, in which predefined components are embedded in region 203 of automotive electrical connector 202, specifically in the plastic housing of automotive electrical connector 202.

[0040] The target object 204 is in Figure 2B Shown is a brake disc, in which predefined components are embedded in region 205 of brake disc 204, specifically on the surface of brake disc 204.

[0041] The target object 206 is in Figure 2C Shown is a bag, in which predefined components are embedded in region 207 of bag 206, specifically in the brand seal attached to bag 206.

[0042] Of course, other target objects, whether physical or digital, are possible and fall within the scope of this disclosure.

[0043] In some embodiments, a surface conditioning device is used to embed the noise components in the target objects 102, 202, 204, 206. Such surface conditioning devices can include, for example, an etching machine. In some embodiments, the etching machine can be operated to engrave the noise components in a two-dimensional manner. However, the noise components can also be or instead be three-dimensional noise components embedded in the surface of the target object.

[0044] Further reference is made to Figure 3 , the computer system 100 (which generally also includes the smartphone 100) is adapted to perform a plurality of functions during operation to identify predefined components.

[0045] The steps performed by the computer system 100 include: receiving, at processing unit 108, S1, image data. For example, as Figure 1 shown, the image data can be acquired by camera 106. Then, S2, the processing unit 108 will identify the noise components from the image data by applying a machine learning-based scheme to the image data. This process will be described in further detail in Figure 4A and 4B .

[0046] Once the noise components are identified from the image data, the processing unit will extract the predefined components from the noise components, S3. As described above, extracting the predefined components from the noise components can be achieved by providing the noise components as input to a feature extraction algorithm, where the feature extraction algorithm is configured to identify specific features included in the noise components.

[0047] Then, according to the solution of the present disclosure, predefined components are identified through a comparison process, S4. During the comparison process, the predefined components extracted from the noise components are compared with other pre-stored comparable components. The comparison process can, for example, include comparing the predefined components with multiple comparable components stored in a database. The processing unit 108 can, for example, be adapted to determine the matching level between the predefined components and each of the multiple comparable components stored in the database. Then, the best-matching component in the database is used to identify the predefined component.

[0048] As described above, the type of the target object can also be determined (from the image of the target object). The database storing multiple different comparable components can also include the identifier of the target object. Therefore, the identifier of the target object can be compared with the actual target object (within the image). If there is a match, it can be determined that the target object is a genuine original.

[0049] Now turning to Figure 4A and Figure 4B , a schematic diagram showing a possible implementation of a machine learning-based solution used in conjunction with the present disclosure is presented. As described above, the machine learning-based solution is used according to the present disclosure to identify noise components from image data and may be used to determine the type of the target object. The following discussion will focus on how to identify noise components from image data.

[0050] Figure 4A A possible method of implementing the deep neural network 400 is shown, which is adapted to identify noise components from image data.

[0051] The block diagram includes an input layer 402 for receiving the input data of the deep neural network. The input data includes the mathematical representation of the image data and data related to the expected structure of the noise components. Information related to the type of the target object in which the noise components are embedded can also be included.

[0052] The image data, the data related to the expected structure of the noise components, and the data related to the type of the target object can be provided as a data matrix or a graph. The image data may include a series of images showing the target object in some embodiments. The input layer includes nodes 404 associated with each input.

[0053] In block 406, the deep neural network 400 may also include one or more convolutional layers. The deep neural network based on a recurrent layer takes not only the current data of the input layer 402 as input but also processes the previously processed data. In other words, the recurrent layer is beneficial for capturing the historical information of the input data.

[0054] The nodes 404 of the input layer 402 communicate with the nodes 408 of layer 406 via connections 410. The connections 410 and the weights of the connections are determined during the training phase, such as supervised or unsupervised training.

[0055] The identified noise components are provided as output in a mathematically represented form in the output layer 412. It should be noted that the number of connections and nodes in each layer may vary, Figure 4A and are provided only as examples. Thus, in some deep neural network designs, more layers may be used than Figure 4A shown herein.

[0056] In an exemplary embodiment of the present disclosure, a machine learning-based scheme applies a convolutional neural network to at least a portion of object recognition. In a convolutional neural network, as is known to those skilled in the art themselves, convolutions of the input layer are used to calculate the output. The local connections are formed such that each part of the input layer is connected to the nodes of the output layer. Filters are applied to each layer, and the parameters of the filters are learned during the training phase of the neural network.

[0057] The deep neural network can be trained based on supervised learning of general noise components and their associated target objects, where a supervisor (user) will assist in the training. Alternatively, the deep neural network can also be trained in an unsupervised learning manner based on similar information.

[0058] In addition, the control functions of the present disclosure can be implemented by an existing computer processor, or by a dedicated computer processor designed for a suitable system, which can be used for this purpose or other purposes, or by a hardware system. Embodiments within the scope of the present disclosure include program products that include a machine-readable medium for carrying or having machine-executable instructions or data structures stored thereon. Such a machine-readable medium can be any available medium accessible by a general-purpose computer, a dedicated computer, or other machines having a processor. For example, these machine-readable media can include RAM, ROM, EPROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, solid-state drives or other flash-based non-volatile storage devices, or any other medium that can be used to carry or store the required program code in the form of machine-executable instructions or data structures and can be accessed by a general-purpose computer, a dedicated computer, or other machines having a processor. When information is transmitted or provided to a machine via a network or other communication connection (whether wired, wireless, or a combination of wired and wireless), the machine treats that connection as a machine-readable medium. Thus, any such connection is properly referred to as a machine-readable medium. Combinations of the above media are also included within the scope of machine-readable media. Machine-executable instructions include, for example, instructions and data that cause a general-purpose computer, a dedicated computer, or a dedicated processing machine to perform a certain function or a set of functions.

[0059] Although the illustration may show an order, the order of steps may be different from that shown. Additionally, two or more steps may be performed simultaneously or partially in parallel. Such variations will depend on the software and hardware systems chosen and the designer's selections. All such variations are within the scope of this disclosure. Similarly, the software implementation may be achieved by standard programming techniques, using rule-based logic and other logic to accomplish the various connection steps, processing steps, comparison steps, and decision steps. Additionally, although this disclosure has been described with reference to its specific exemplary embodiments, many different alterations, modifications, etc. will become apparent to those skilled in the art.

[0060] Furthermore, through the study of the drawings, this disclosure, and the appended claims, those skilled in the art can understand and implement variations of the disclosed embodiments when practicing the claimed disclosure. Additionally, in the claims, the word "comprising" does not exclude other elements or steps, and the singular indefinite article does not exclude the plural form.

Claims

1. A method for using a computer system to identify predefined components embedded in image data of a target object, characterized in that, The computer system includes a processing unit, and the method includes the following steps: Receiving image data at the processing unit, Applying, by the processing unit, a machine learning-based scheme to the image data to identify noise components from the image data, wherein the noise components are embedded in a target object, Extracting, by the processing unit, predefined components from the identified noise components based on recognizable features in the noise components, and Comparing, by the processing unit, the predefined components with pre-stored data in a storage element communicating with the processing unit to identify the predefined components.

2. The method according to claim 1, characterized in that The machine learning-based scheme includes a machine learning pipeline.

3. The method according to claim 2, wherein The machine learning pipeline includes a plurality of autoencoder components.

4. The method according to any one of the preceding claims, characterized in that, The machine learning-based scheme is trained using a plurality of different image data, each of which includes known noise components.

5. The method according to any one of the preceding claims, characterized in that, The storage element is a database storing a plurality of different target objects.

6. The method according to any one of the preceding claims, characterized in that, The target objects are selected from the group including images and physical products.

7. The method according to any one of the preceding claims, characterized in that, It further includes the following steps: Embedding, by the processing unit, a preselected noise component into the target object.

8. The method according to claim 7, characterized in that, Selecting a scheme for embedding the preselected noise component into the target object based on the type of the target object.

9. The method according to claim 8, wherein When the target object is a physical component, the method further includes the following steps: Applying the preselected noise component to the surface of the target object.

10. The method according to claim 8, wherein When the target object is an image, the method further includes the following steps: Combining the original image with the preselected noise component to form a noisy image.

11. A computer system adapted to recognize predefined components embedded in target object image data, the computer system including a processing unit, characterized in that, The processing unit is adapted to: Receive image data, Identify noise components from the image data by applying a machine learning-based scheme to the image data, wherein the noise components are embedded in a target object, Extract predefined components from the identified noise components based on recognizable features in the noise components, and Identify the predefined components by comparing with pre-stored data in a storage element communicating with the processing unit.

12. The computer system according to claim 11, wherein It further includes an object capture device for capturing image data of the target object.

13. The computer system according to any one of claims 11 and 12, further comprising a storage element, characterized in that, The storage element is a database filled with a plurality of different target objects.

14. The computer system according to any one of claims 11 to 13, characterized in that, The machine learning-based scheme includes a machine learning pipeline.

15. The computer system according to any one of claims 11 to 14, characterized in that, The target object is a physical component, and the processing unit is further adapted to: Form a control signal to control a surface adjustment device to apply the preselected noise component to the surface of the target object.

16. An electronic user device includes the computer system according to any one of the preceding claims.

17. A computer program product, comprising a non-transitory computer-readable medium having stored thereon a computer program means for controlling a computer system adapted to identify predefined components embedded in image data of a target object, the computer system including a processing unit, characterized in that, The computer program product includes: Code for receiving image data at the processing unit, Code for applying, by the processing unit, a machine learning-based scheme to the image data to identify noise components from the image data, wherein the noise components are embedded in a target object, Code for extracting, by the processing unit, predefined components from the identified noise components based on recognizable features in the noise components, and Code for comparing, by the processing unit, the predefined components with pre-stored data in a storage element communicating with the processing unit to identify the predefined components.

Images