Checking compliance of a packaged product

Abstract

Example approaches for performing a compliance check and generating a compliance report for a packaged product are described. In an example, an image of a packaged product is obtained for determining its compliance with respect to certain regulatory guidelines. Thereafter, certain descriptive information is extracted from the package to be processed using a compliance checking model to ascertain whether the packaged product is compliant or not. Based on the processing, the compliance checking model generates a compliance report indicating the extent of compliance and some recommendations as well. In an example, the model is trained using a dataset of product package images, each associated with a reference product category and relevant regulatory guidelines. The training process involves extracting package description information from training images and correlating it with product categories and regulatory requirements. The trained model can then assess new packaged products for compliance.

Description

CHECKING COMPLIANCE OF A PACKAGED PRODUCTBACKGROUND

[0001] Finished products or goods are articles that have been manufactured using manufacturing process. Such finished products may be subsequently packaged before they are rolled out for sale or distribution in markets through various distribution channels. Examples of such finished products include, but are not limited to, food items, cosmetic items, clothes, furniture, smartphones, etc. Regional regulations may require that the packaging of such finished products carry certain descriptive information in relation to either the product itself, or information about the manufacturer. If such descriptive information is not in compliance with the regulatory guidelines, then the product is considered to be as non-compliant and the manufacturer I packer in question is liable to a penaltyBRIEF DESCRIPTION OF FIGURES

[0002] Systems and / or methods, in accordance with examples of the present subject matter are now described and with reference to the accompanying figures, in which:

[0003] FIG. 1 illustrates a training system for training a compliance checking model, as per an example;

[0004] FIG. 2 illustrates a compliance checking system for generating a compliance report for an input packaged product, as per one example;

[0005] FIG. 3 illustrates a method for training a compliance checking model, as per an example;

[0006] FIG. 4 illustrates a method for generating a compliance report for an input packaged product based on a trained compliance checking model, as per an example;

[0007] FIG. 5A-5B illustrates exemplary views for generating and obtaining input package images, as per an example;

[0008] FIG. 6A illustrates an exemplary view of product categories displayed by a compliance checking system, as per an example;

[0009] FIG. 6B illustrates another exemplary view of product subcategories displayed by a compliance checking system for receiving first set of user inputs from the user, as per an example;

[0010] FIG. 7 illustrates an exemplary view of package detail fields (panel details) displayed by the compliance checking system for receiving second set of user inputs from the user, as per an example;

[0011] FIG. 8A-8B illustrates exemplary views of questions displayed by the compliance checking system for receiving third set of user inputs answering the displayed questions, as per an example;

[0012] FIG. 9 illustrates an exemplary view of a compliance report, as per an example; and

[0013] FIG. 10. Illustrates an exemplary view of certain general recommendations made by the compliance checking system, as per an example.DETAILED DESCRIPTION

[0014] As described above, a finished good or product is a good that has moved through the production process and is ready for distribution in commercial markets. Such products may be packaged before they are distributed. Besides providing an environmental barrier for protecting the finished product, the packaging may also bear information not only pertaining to the finished product, but also information providing usage instructions, safety related information, or may bear various branding or marketing artifacts pertaining to the manufacturer or the product in question. The manner in which such information may be depicted as to conform to certain regulations. If this descriptive information is not in compliance with certain regulatory guidelines, then the product is considered to be as non- compliant

[0015] These regulations are regional and may differ across jurisdictions. Within India, the applicable legislation is the Legal Metrology Act of 2009. The Act prescribes various obligations and requirements as tohow the descriptive information in relation to the product ought to be provided on the packaging itself. For example, the Act prescribes certain declarations or statements that are to be mandatorily mentioned on the packages of the product. In addition to such information, the Act (or such similar regulations) may prescribe the area or other parameters that are to be complied with, for depicting such information.

[0016] Such regulations may change over a period of time. In instances, owing to changing regulations, packaging of product already ready for distribution may have to be undertaken. There may be other such instances which may require checking whether the packaging requirements are conformed with. Conventionally, compliance of guidelines laid down in different regulations is done manually by physically analysing and / or measuring portions of the product or package of the product or physically inspecting the position, visibility, or size of area over which the relevant information is to be provided. Such manual compliance check pose a number of challenges in relation to its efficiency and correctness. For instance, manual compliance checks may take a long time for checking the packaging of individual product categories. Such activities are prone to human error wherein which some information or aspect may be inadvertently missed. Therefore, there is a need for such a technique which enable automatic check of compliance of regulatory guidelines on packaged products and generates a report indicating results of compliance check.

[0017] Approaches for performing a compliance check and generating a compliance report of a packaged product, are described. The packaged product may be a product which is being checked for compliance under the applicable regulatory guidelines. In an example, an image of the package may be captured and analyzed. During analysis, the package description information is extracted from the package image. In an example, the package description information includes an attribute value corresponding to a regulatory compliance attribute of the packaged product. Examples of such regulatory compliance attribute include, but are not limited to,presence, size, style, colour, and location, or combination thereof, of various recommended descriptive information on the package of packaged product. The recommended descriptive information is one of details of manufacturer, packer, importer, consumer complaints, relevant dates, sale price in different units, quantity of package in different units, dimension of the product, dimension of the text written on the label, and type of product, or combination thereof, on the package of the packaged product.

[0018] The package description information is checked and analyzed for determining compliance. To do so, a machine learning based compliance checking model is used. In an example, the compliance checking model is trained based on a training product package image, corresponding category of the product, and applicable regulatory guidelines. In one example, the training may be based on a set of training product package images including images corresponding to different portions or sides of the product or packaging. Examples of such portions include, but are not limited to, front side, back side, manufacturing details cropped image, and many more.

[0019] For training, the training product package images are initially processed to extract training package description, for each image. In another example, the training package description corresponding to training product package images may be annotated with each image and thereafter extracted or may be detected through image recognition means. The training package description includes an attribute value corresponding to a regulatory compliance attribute of the packaged product. It may be noted that, the attribute value corresponding to the regulatory compliance attribute may include values ranging from compliant as well as non-compliant attribute values corresponding to different regulatory compliance attributes. For example, dimensions of the text on the label which conforms to the specification prescribed under applicable regulations would be considered a compliant value. On the other hand, dimensions of the text on the label which does not conform with the prescribed specifications, would be considered as a non-compliant attribute value. Whether the attribute valueis compliant or not is determined based on the values or ranges mentioned in the relevant regulatory guidelines.

[0020] Examples of regulatory compliance attribute include, but are not limited to, presence, size, style, colour, and location, or combination thereof, of various recommended descriptive information on the package of packaged product. The recommended descriptive information is one of details of manufacturer, packer, importer, consumer complaints, relevant dates, sale price in different units, quantity of package in different units, dimension of the product, dimension of the text written on the label, and type of product, or combination thereof, on the package of the packaged product.

[0021] During the training process, the compliance checking model correlates the training information, i.e., the attribute value corresponding to compliance attribute, the product category, and the specific regulation. The compliance checking model based on the information correlates the above described elements and the necessary requirements for any packaged product of any category to be in compliance with certain regulatory guidelines. It may be noted that, the examples of package description information are exemplary and may change based on the need of the corresponding regulatory guidelines. For example, the above-described examples of package description information are for checking compliance with respect to Legal Metrology Act, 2009. Other Examples of regulatory guidelines include, but are not limited to, Packaging and Labelling Regulations under the Food Safety and Standards Authority of India (FSSAI), Drugs and Cosmetics Act, 1940, Bureau of Indian Standards (BIS) Act, 2016, Hazardous Substances (Classification, Packaging and Labelling) Rules, 2011 , and The Insecticides Act, 1968, or combination thereof. It may be noted that although the present explanation has been provided in the context of regulations applicable for India, such approaches may be extended for any jurisdiction or a set of regulations, without deviating from the scope of the present subject matter.

[0022] Once trained, the compliance checking model is used for assessing labels or packaging of one or more finished products to ascertain whether the packaging of the finished product under consideration complies with one or more applicable regulations. To this end, an input image of the packaging of a packaged product may be initially captured. The input image may be processed to determine input package description The package description may indicate the physical and visual attributes of the product packaging that are relevant to regulatory compliance. This includes details such as the presence, size, style, color, and location of required information on the package, such as manufacturer details, expiration details, ingredient lists, and product measurements.

[0023] Once the package description is determined, the same may be analyzed using the trained compliance checking model to determine a compliance evaluation results. The compliance evaluation results, in an example, may indicate whether the packaging under evaluation conforms with the applicable regulations, or whether it deviates from the conditions stipulated under the regulations.

[0024] In an example, a compliance report may be generated depicting the compliance evaluation results. The compliance report may, in instances of non-conformance, provide various recommendations pertaining to the packaged product being assessed. Such recommendations may provide any corrective changes that may be affected to the packaging of the finished product under consideration.

[0025] As will be explained further, the present approaches enable identification of the recommendations or requirements which are met by the package of the packaged product, and which are not met by the package based on the processing of package images using the compliance checking model. The compliance checking model is trained based on a plurality of training package images which are captured for various products available in the market. Therefore, based on the observations and learnings which the model had from the vast variety of training images the compliancechecking model provide details regarding compliance or non-compliance of certain regulatory guidelines. Therefore, there is no need to manually check the product for compliance with respect to certain regulatory guidelines and the accuracy of compliance check is also enhanced.

[0026] FIG. 1 illustrates a training system 102 comprising a processor or memory (not shown), for training a compliance checking model. In an example, the training system 102 (referred to as system 102) may be communicatively coupled to a repository 104 through a network 106. The repository 104 may further include a training information 108. The training information 108 may include a plurality of training product package images corresponding to various training packaged products. In another example, these plurality of training images may be arranged in different set with each set including images of each side of the packaged product

[0027] In one example, along with plurality of images, the training information 108 may further include training package description information, a reference product category, and applicable or relevant regulatory guidelines corresponding to the training images. In an example, the training package description information includes an attribute value corresponding to a regulatory compliance attribute of the training packaged product. It may be noted that, the attribute value corresponding to the regulatory compliance attribute may include values ranging from compliant as well as non-compliant attribute values corresponding to different regulatory compliance attributes. For example, dimensions of the text on a label which conforms to the specification prescribed under applicable regulations would be considered a compliant value. On the other hand, dimensions of the text on the label which does not conform with the prescribed specifications, would be considered as a non-compliant attribute value. Whether the attribute value is compliant or not is determined based on the details mentioned in the relevant regulatory guidelines.

[0028] Examples of such regulatory compliance attribute include, but are not limited to, presence, size, style, colour, and location, or combinationthereof, of various recommended descriptive information on the package of packaged product. The recommended descriptive information is one of details of manufacturer, packer, importer, consumer complaints, relevant dates, sale price in different units, quantity of package in different units, dimension of the product, dimension of the text written on the label, and type of product, or combination thereof, on the package of the packaged product.

[0029] It may be noted that, the training information 108, although depicted as being obtained from a single repository, such as repository 104, may also be obtained from multiple other sources without deviating from the scope of the present subject matter. In such cases, each of such multiple repositories may be interconnected through a network, such as the network 106.

[0030] The network 106 may be a private network or a public network and may be implemented as a wired network, a wireless network, or a combination of a wired and wireless network. The network 106 may also include a collection of individual networks, interconnected with each other and functioning as a single large network, such as the Internet. Examples of such individual networks include, but are not limited to, Global System for Mobile Communication (GSM) network, Universal Mobile Telecommunications System (UMTS) network, Personal Communications Service (PCS) network, Time Division Multiple Access (TDMA) network, Code Division Multiple Access (CDMA) network, Next Generation Network (NGN), Public Switched Telephone Network (PSTN), Long Term Evolution (LTE), and Integrated Services Digital Network (ISDN).

[0031] The system 102 may further include instructions 1 10 and a training engine 1 12. In an example, the instructions 1 10 are fetched from the memory and executed by the processor included within the system 102. The training engine 1 12 may be implemented as a combination of hardware and programming, for example, programmable instructions to implement a variety of functionalities. In examples described herein, such combinations of hardware and programming may be implemented in several differentways. For example, the programming for the training engine 1 12 may be executable instructions, such as instructions 110. Such instructions may be stored on a non-transitory machine-readable storage medium which may be coupled either directly with the system 102 or indirectly (for example, through networked means). In an example, the training engine 112 may include a processing resource, for example, either a single processor or a combination of multiple processors, to execute such instructions. In the present examples, the non-transitory machine-readable storage medium may store instructions, such as instructions 110, that when executed by the processing resource, implement training engine 1 12. In other examples, the training engine 112 may be implemented as electronic circuitry.

[0032] The instructions 110 when executed by the processing resource, cause the training engine 1 12 to train a compliance checking model 114. The instructions 1 10 may be executed by the processing resource for training the compliance checking model 114 based on the training information 108. The system 102 may further include a training product package image(s) 116, a training package description information 1 18, a reference product category 120, and a relevant regulatory guideline(s) 122. In an example, the system 102 may obtain training information 108 corresponding to a single reference product from the repository 104, and the information pertaining to that is stored as training product package image(s) 1 16, training package description information 1 18, reference product category 120, and relevant regulatory guideline(s) 122 in the system 102.

[0033] For training, the package description information, such as training package description information 1 18, of the training product package images may be used to train the compliance checking model 1 14. The compliance checking model 1 14 may then be used to provide compliance evaluation results corresponding to a plurality of recommendations pertaining to a category of an input packaged productbased on the package description information of the input packaged product and relevant regulatory guidelines.

[0034] In operation, the training engine 112 of the system 102 obtain the training information 108 from the repository 104 which may include training product package image(s) 1 16, training package description information 1 18, reference product category 120, and relevant regulatory guideline(s) 122 in the system 102. The training information 108 may include a plurality training product package images with corresponding reference product category 120 and the relevant regulatory guideline(s) 122 annotated.

[0035] In another example, the system 102 obtain the training information 108 from the repository 104 which may only include training product package image(s) 1 16, corresponding reference product category 120, and relevant regulatory guideline(s) 122. In such a case, once the training information 108 is obtained, the training product package image(s) 1 16 are further processed to extract training package description information 1 18 for training images. In an example, the training package description information 1 18 may be annotated with each image and thereafter extracted or may be detected through image recognition means. The training package description information 1 18 includes an attribute value corresponding to a regulatory compliance attribute of the training packaged product. It may be noted that, the attribute value corresponding to the regulatory compliance attribute may include values ranging from compliant as well as non-compliant attribute values corresponding to different regulatory compliance attributes. For example, dimensions of the text on the label which conforms to the specification prescribed under applicable regulations would be considered a compliant value. On the other hand, dimensions of the text on the label which does not conform with the prescribed specifications, would be considered as a non-compliant attribute value. Whether the attribute value is compliant or not is determined based on the details mentioned in the relevant regulatory guidelines.

[0036] Examples of such regulatory compliance attribute include, but are not limited to, presence, size, style, colour, and location, or combination thereof, of various recommended descriptive information on the package of packaged product. The recommended descriptive information is one of details of manufacturer, packer, importer, consumer complaints, relevant dates, sale price in different units, quantity of package in different units, dimension of the product, dimension of the text written on the label, and type of product, or combination thereof, on the package of the packaged product. It may be noted, although, the training information 108 is described as having plurality of training images, however, it may include various sets of training images with each set corresponding to single training packaged product. In such a case, the training of the model for the product is done based on all the images of each set.

[0037] Thereafter, the training engine 1 12 process the training package description information 1 18, reference product category 120, and relevant regulatory guideline(s) 120 to make an interrelation between these elements and trains the compliance checking model 114 based on this interrelation. During the training process, the compliance checking model 1 14 is to associate the package description information and the product categories with the requirements specified in the relevant regulatory guideline(s) 120. In an example, it may identify patterns and relationship between these elements, learning to recognize what constitutes compliance for different product categories under various regulations. When trained, the compliance checking model develops an understanding of the regulatory requirements, distinguishing between essential and non-essential requirements for compliance across various product categories and their respective regulatory guidelines. It may be noted that, the examples of training package description information 1 18 are exemplary and may change based on the corresponding regulatory guidelines. For example, the abovedescribed examples of package description information are for checking compliance with respect to Legal Metrology Act, 2009. Other Examples ofregulatory guidelines include, but are not limited to, Packaging and Labelling Regulations under the Food Safety and Standards Authority of India (FSSAI), Drugs and Cosmetics Act, 1940, Bureau of Indian Standards (BIS) Act, 2016, Hazardous Substances (Classification, Packaging and Labelling) Rules, 2011 , and The Insecticides Act, 1968, or combination thereof.

[0038] Once trained, the compliance checking model 1 14 may be used for generating a compliance report indicating compliance results corresponding to a plurality of requirements pertaining to a category of an input packaged product. For example, package description information pertaining to the input packaged product be processed using the compliance checking model 1 14. In such a case, based on the compliance checking model 1 14, it is determined whether the package description information depicting information regarding details printed on the package of the input packaged product are in compliance with the regulatory guidelines. If non-compliance, details which are not compliant with the regulator guidelines and corresponding corrective recommendations for rectifying irregularities are mentioned in a compliance report. The manner in which the package description information of an input packaged product is processed to generate a compliance report based on the trained compliance checking model 114 is further described in conjunction with FIG. 2.

[0039] FIG. 2 illustrates a computing environment 200 including a compliance checking system 202 (referred to as system 202), for performing a compliance check and generating a compliance report indicating details regarding compliance of an input packaged product with respect to certain regulatory guidelines. To do so, an image or images of input packaged product are required to be processed based on a compliance checking model, such as compliance checking model 1 14. In an example, the input package image is captured by the user in real-time. In another example, the system 202 may be communicatively coupled to a repository 204 through a network 206 (similar to network 106). The repository 204 may include a setof input package images corresponding to input packaged product. The set of images include images of each side of the input packaged product and even cropped images as well for certain portions of the product, e.g., manufacturing date, sale price, and expiry date label.

[0040] Similar to the system 102, system 202 may further include instructions 208 and an analysis engine 210. In an example, the instructions 208 are fetched from a memory and executed by a processor included within the system 202. The analysis engine 210 may be implemented as a combination of hardware and programming, for example, programmable instructions to implement a variety of functionalities. In examples described herein, such combinations of hardware and programming may be implemented in several different ways.

[0041] For example, the programming for the analysis engine 210 may be executable instructions, such as instructions 208. Such instructions 208 may be stored on a non-transitory machine-readable storage medium which may be coupled either directly with the system 202 or indirectly (for example, through networked means). In an example, the analysis engine 210 may include a processing resource, for example, either a single processor or a combination of multiple processors, to execute such instructions. In the present examples, the non-transitory machine-readable storage medium may store instructions, such as instructions 208, that when executed by the processing resource, implement analysis engine 210. In other examples, the analysis engine 210 may be implemented as electronic circuitry.

[0042] The system 202 may include a compliance checking model, such as the compliance checking model 1 14. The system 202 may further include an input package image(s) 212, an input package description information 214, and a compliance report 216. In an example, the compliance checking model 1 14 is a trained model based on thousands of training images and corresponding training package description information with respect to relevant regulatory guidelines.

[0043] In operation, initially, an input package image or a set of input package images corresponding to an input packaged product is obtained and stored as input package image(s) 212 in the system 202. In an example, the input packaged product is the product which is being checked for compliance with respect to certain regulatory guidelines. Such input package image(s) 212 may be either stored in a repository, such as repository 204 or may be captured by a camera device in real time. In one example, a user interface (such as depicted in FIG. 5A-5B) is displayed before the user to obtain the input package image(s) 212 corresponding to different sides. On seeing the user interface, the user accordingly selects the input package image(s) 212 from the system 202 itself or captures with the camera device in real-time.

[0044] In another example, the input package image(s) 212 may be obtained with a reference object placed alongside the input packaged product in the image. Examples of such reference object include, but are not limited to, a payment card, a ruler, a paper currency note having standard dimensions, compact discs and many more For example, while clicking a picture of the input packaged product, the user has to place the reference object alongside the input packaged product and then take the picture. As depicted in FIG. 5B, an image 502 of package of an input packaged product 504 includes a reference object 506 placed alongside with the input packaged product 504.

[0045] Thereafter, the analysis engine 210 analyse the input package image(s) 212 that includes both the input packaged product and the reference object to determine the real-world dimensions of the input packaged product. In an example, consider the reference object to be as a payment card having standardized dimensions of 85.60 x 53.98 mm. In such a case, the analysis engine 210 first identifies and measures the payment card in the input package image(s) 212, establishing a pixel-to-millimeter ratio. It then uses this ratio to calculate the actual dimensions of the input packaged product. For example, if the card appears 100 pixels wide in theimage, and its known width is 85.60 mm, the analysis engine 210 may determine that each pixel represents 0.856 mm in the real world. Using this scale, the analysis engine 212 may then measure the input packaged product in the input package image(s) 212 and convert those pixel measurements to real-world dimensions.

[0046] Once obtaining of input images and subsequent dimension measurement is performed, the analysis engine 210 renders a first set of question indicating various product categories and corresponding subcategories for the user to select one of them for the input packaged product (e.g., as depicted in FIG. 6A-6B). In response, the user provides a first set of user inputs to select the product category and corresponding subcategory for the input packaged product from the rendered first set of questions. The analysis engine 210 receives the first set of user inputs indicating category and sub-category of the input packaged product. Examples of categories include, but are not limited to, cosmetics, drugs, incense sticks, aerosol products, electronics, fabrics, food, fuels, oils, sheets, liquid chemicals, etc. For one of the categories, e.g., cosmetics, sub-categories include perfumes / deos, soaps, shampoos, serums, creams, lotions, and other cosmetic products. The manner in which the descriptive information is mentioned on various products will depend category or subcategory of the product.

[0047] Once selected, the analysis engine 210 renders a second set of questions via another user interface to obtain information from the user indicating package detail fields. These package detail fields may include, but are not limited to, dimensions of the package, dimension of the text written on the label, surface area of a principal image among the images comprised in the set of input images, and a generic name of the input packaged product. It may be noted that, details such as hight, width, and surface area of the principal image of the input packaged product is used to calculate size of text written on the product’s principal image. In an example, the principal image includes most of the details corresponding to the inputpackaged product. For example, details such as details of manufacturer, packer, importer, consumer complaints, relevant dates, sale price in different units, quantity of package in different units, dimension of the product, dimension of the text written on the label, and type of the product.

[0048] In an example, one or more second set of questions rendered to the user are pre-filled based on the dimensions determined using the reference object. For example, in FIG. 7 a user interface is depicted with value 702 corresponding to the height of the principle image of the product is pre-filled using the dimensions determined by the analysis engine 210 using standard dimension of the reference object in the input image as reference.

[0049] In response, the user provides a second set of user inputs corresponding to the second set of questions, providing values for the package detail fields. In case of pre-filled fields, the user provides either a confirmation or modification to the pre-filled value in the form of second set of user inputs.

[0050] Upon receiving the second set of user inputs, the analysis engine 210 renders a third set of questions on the user interface to obtain information from the user. These questions may include, but are not limited to, whether the packaged product is for retail sale or wholesale, uses of the packaged product, range of net quantity of pre-packaged commodity, surface on which label information is molded, and country of manufacturing. In response, the user provides a third set of user inputs corresponding to the third set of questions. It may be noted that the first, second, and third sets of user inputs provided by the user help in narrowing down the processing of the input package images.

[0051] Once the third set of user inputs are received, the analysis engine 210 extracts input package description information 214 from the input package image(s) 212 using image recognition techniques using first, second and third set of user inputs as a basis. It may be noted that, since presence, style, location, color of various descriptive information mentionedon the package is dependent on category, dimension, and other product related details, the analysis engine 210 utilizes these inputs to accurately identify the descriptive information. In an example, this information is extracted by analysing the input package image(s) 212 on pixel level. As described above as well, the input package description information 214 includes an attribute value corresponding to a regulatory compliance attribute of the input packaged product. Examples of such regulatory compliance attribute include, but are not limited to, presence, size, style, colour, and location, or combination thereof, of various recommended descriptive information on the package of packaged product. The recommended descriptive information is one of details of manufacturer, packer, importer, consumer complaints, relevant dates, sale price in different units, quantity of package in different units, dimension of the product, dimension of the text written on the label, and type of product, or combination thereof, on the package of the packaged product.

[0052] Once extracted, the analysis engine 210 processes the input package description information 214 using a trained compliance checking model, such as compliance checking model 1 14, to obtain compliance evaluation results. In an example, the compliance evaluation result indicates the degree of adherence of the input packaged product to specific regulatory guidelines or requirements. It may include assessments of individual compliance attributes, such as the presence, size, style, color, and location of required information on the product packaging. These results serve as the basis for determining whether the product meets various regulatory recommendations or requirements pertaining to its category.

[0053] The analysis engine 210 then analyzes these compliance evaluation results to determine the compliance status for a plurality of recommendations pertaining to the product category of the input packaged product. Based on this analysis, the analysis engine 210 generates a compliance report, such as compliance report 216, indicating whether each recommendation is met by the input packaged product.

[0054] FIG. 3 illustrates example method 300 for training a compliance checking model, in accordance with examples of the present subject matter. The order in which the above-mentioned method is described is not intended to be construed as a limitation, and some of the described method blocks may be combined in a different order to implement the method, or alternative method.

[0055] Furthermore, the above-mentioned method may be implemented in a suitable hardware, computer-readable instructions, or combination thereof. The steps of such method may be performed by either a system under the instruction of machine executable instructions stored on a non- transitory computer readable medium or by dedicated hardware circuits, microcontrollers, or logic circuits. For example, the method may be performed by a training system, such as system 102. In an implementation, the method may be performed under an “as a service” delivery model, where the system 102, operated by a provider, receives programmable code. Herein, some examples are also intended to cover non-transitory computer readable medium, for example, digital data storage media, which are computer readable and encode computer-executable instructions, where said instructions perform some or all the steps of the above- mentioned method.

[0056] In an example, the method 300 may be implemented by the system 102 for training the compliance checking model 1 14 based on a training information, such as training information 108. At block 302, a training information is obtained from a repository over a network. For example, the training engine 1 12 of the system 102 obtain the training information 108 from the repository 104 which may be further stored as training product package image(s) 1 16, training package description information 1 18, reference product category 120, and relevant regulatory guideline(s) 122 in the system 102. The training information 108 may include a plurality of set of training product package images withcorresponding reference product category 120 along with the relevant regulatory guideline(s) 122 mentioned as well.

[0057] At block 304, training package description information are extracted from the training information. In an example, the training engine 1 12 extract training package description information 118 from the training product package image(s) 1 16. For example, the system 102 obtain the training information 108 from the repository 104 which may only include training product package image(s) 1 16, corresponding product category 120, and relevant regulatory guideline(s) 122. In such a case, once the training information 108 is obtained, the training product package image(s) 1 16 are further processed to extract training package description information 118 for training images. In an example, the training package description information 1 18 may be annotated with each image and thereafter extracted or may be detected through image recognition means. In an example, the training package description information 1 18 includes an attribute value corresponding to a regulatory compliance attribute of the training packaged product. Examples of such regulatory compliance attribute include, but are not limited to, presence, size, style, colour, and location, or combination thereof, of various recommended descriptive information on the package of packaged product. The recommended descriptive information is one of details of manufacturer, packer, importer, consumer complaints, relevant dates, sale price in different units, quantity of package in different units, dimension of the product, dimension of the text written on the label, and type of product, or combination thereof, on the package of the packaged product.

[0058] At block 306, the training engine 1 12 process the training package description information 1 18, reference product category 120, and relevant regulatory guideline(s) 120 to make an interrelation between these elements and trains the compliance checking model 114 based on this interrelation. During the training process, the compliance checking model 1 14 learns to associate the package description information and the productcategories with the requirements specified in the relevant regulatory guideline(s) 120. It identifies patterns and relationship between these elements, learning to recognize what constitutes compliance for different product categories under various regulations. When trained, the compliance checking model develops an understanding of the regulatory requirements, distinguishing between essential and non-essential requirements for compliance across various product categories and their respective regulatory guidelines..

[0059] Once trained, the compliance checking model 114 may be used for generating a compliance report indicating compliance results corresponding to a plurality of requirements pertaining to a category of an input packaged product. For example, package description information pertaining to an input packaged product be processed based on the compliance checking model 1 14. In such a case, based on the compliance checking model 1 14, it is determined whether the package description information depicting information regarding details printed on the package of the input packaged product are complying with the regulatory guidelines.

[0060] FIG. 4 illustrates another example method 400 for generating a compliance report for an input packaged product. Based on the present approaches as described in the context of the example method 400, the package description information of the input packaged product is analyzed based on the trained compliance checking model 1 14 to generate the compliance report indicating compliance results of the input packaged product.

[0061] Further, the above-mentioned method 400 may be implemented in a suitable hardware, computer-readable instructions, or combination thereof. The steps of such method may be performed by either a system under the instruction of machine executable instructions stored on a non- transitory computer readable medium or by dedicated hardware circuits, microcontrollers, or logic circuits. For example, the method may be performed by a compliance checking system, such as system 202. In animplementation, the method may be performed under an “as a service” delivery model, where the system 202, operated by a provider, receives programmable code. Herein, some examples are also intended to cover non-transitory computer readable medium, for example, digital data storage media, which are computer readable and encode computer-executable instructions, where said instructions perform some or all the steps of the above-mentioned method.

[0062] At block 402, an input package image or a set of input package images corresponding to an input packaged product is obtained. For example, the analysis engine 210 of the system 202 obtains an input package image or a set of input package images corresponding to the input packaged product and stored them as input package image(s) 212 in the system 202. In an example, the input packaged product is the product which is being checked for compliance with respect to certain regulatory guidelines. Such input package image(s) 212 may be either stored in a database repository, such as repository 204 or may be captured by a camera device in real time. As illustrated in the exemplary view of FIG. 5A, initially a new project is created by the user and once the same is created, the system 202 obtains or captures input package image(s) 212 based on inputs received from the user as depicted in FIG. 5B.

[0063] In another example, the input package image(s) 212 may be obtained with a reference object placed alongside the input packaged product in the image. Examples of such reference object include , but are not limited to, a payment card, a ruler, a paper currency note having standard dimensions, a compact disc, and many more. For example, while clicking a picture of the input packaged product, the user has to place the reference object alongside the input packaged product and then take the picture. As depicted in FIG. 5B, an image 502 of package of an input packaged product 504 includes a reference object 506 placed alongside with the input packaged product 504.

[0064] At block 404, real-world dimension of the input packaged product is determined based on standardized dimension of a reference object. In an example, the analysis engine 210 analyse the input package image(s) 212 that includes both the input packaged product and the reference object to determine the real-world dimensions of the input packaged product. In an example, consider the reference object to be as a payment card having standardized dimensions of 85.60 x 53.98 mm. In such a case, the analysis engine 210 first identifies and measures the payment card in the input package image(s) 212, establishing a pixel-to-millimeter ratio. It then uses this ratio to calculate the actual dimensions of the input packaged product. For example, if the card appears 100 pixels wide in the image, and its known width is 85.60 mm, the analysis engine 210 may determine that each pixel represents 0.856 mm in the real world. Using this scale, the analysis engine 212 may then measure the input packaged product in the input package image(s) 212 and convert those pixel measurements to real-world dimensions.

[0065] At block 406, a first set of questions are rendered on a display device to obtain information corresponding to a category and a sub-category of the input packaged product. For example, once the input package image(s) 212 are obtained, the analysis engine 210 renders a first set of question indicating various product categories and corresponding subcategories for the user to select one of them (as depicted in FIG. 6A-6B).

[0066] At block 408, a first set of user inputs are received from the user selecting a category and a sub-category of the input packaged product. For example, in response to rendered first set of questions, the user provides the first set of user inputs selecting the product category and corresponding sub-category for the input packaged product. These first set of user inputs are received by the analysis engine 210 of the system 202 to perform further processing. Examples of categories include, but are not limited to, cosmetics, drugs, electronics, fabrics, food, fuels, oils, sheets, liquid chemicals, etc. FIG. 6A illustrates an exemplary view of product categoriesdisplayed by the system 202. Further, FIG. 6B illustrates another exemplary view of product sub-categories displayed by the system 202 for receiving first set of user inputs from the user.

[0067] At block 410, a second set of questions are rendered on the display device to obtain information corresponding to a plurality of package related details. For example, upon receiving the first set of user inputs, the analysis engine 210 renders the second set of question inquiring about various product related details from the user (as depicted in FIG. 7). It may be noted that, upon receiving the first set of user inputs indicating the product category and sub-category, the analysis engine 210 may dynamically select and render the second set of questions tailored to the specific product type or category.

[0068] In an example, one or more second set of questions rendered to the user are pre-filled based on the dimensions determined using the reference object. For example, in FIG. 7 a user interface is depicted with value 702 corresponding to the height of the principle image of the product is pre-filled using the dimensions determined by the analysis engine 210 using standard dimension of the reference object in the input image as reference.

[0069] At block 412, a second set of user inputs are received providing values corresponding to a plurality of package related details. For example, in response to rendered second set of questions, the user provides the second set of user inputs providing information for the plurality of package related details. These second set of user inputs are received by the analysis engine 210 to perform further processing. In case there are some pre-filled values present, the user provides their confirmation or modification for the pre-filed values in the form of second set of user inputs.

[0070] At block 414, a third set of questions are rendered on the display device to obtain information pertaining to general details of the input packaged product. For example, upon receiving the second set of user inputs, the analysis engine 210 renders the third set of questions inquiringabout various general details of the product from the user (as depicted in FIG. 8). It may be noted that, upon receiving the second set of user inputs indicating the product related details, the analysis engine 210 may dynamically select and render the third set of questions tailored to the specific product type or category.

[0071] At block 416, a third set of user inputs is received which provides answers corresponding to a plurality of questions related to a general product. For example, analysis engine 210 receives a third set of user inputs answering a plurality of questions related to a general product. Examples of such questions include, but are not limited, whether the product is for retail sale or whole sale, uses of packaged product, range of net quantity of prepackaged commodity, surface on which label information is molded, country of manufacturing, etc. It may be noted that, inputs, such as first, second, and third set of user inputs provided by user helps in narrowing down the processing of the input package image(s) 212. FIG. 8A-8B illustrates exemplary views of questions displayed by the system 202 for receiving third set of user inputs answering the displayed questions.

[0072] At block 418, input package description information is extracted from the set of input package images by analyzing the input package images at pixel level. For example, the analysis engine 210 extracts input package description information 214 from the input package image(s) 212 using image recognition techniques using first, second and third set of user inputs as a basis. It may be noted that, since presence, style, location, color of various descriptive information mentioned on the package is dependent on category, dimension, and other product related details, the analysis engine 210 utilizes these inputs to accurately identify the descriptive information.. In an example, this information is extracted by analysing the input package image(s) 212 on pixel level. In an example, the input package description information 214 includes an attribute value corresponding to a regulatory compliance attribute of the packaged product. Examples of such regulatory compliance attribute include, but are not limited to, presence,size, style, colour, and location, or combination thereof, of various recommended descriptive information on the package of packaged product. The recommended descriptive information is one of details of manufacturer, packer, importer, consumer complaints, relevant dates, sale price in different units, quantity of package in different units, dimension of the product, dimension of the text written on the label, and type of product, or combination thereof, on the package of the packaged product.

[0073] At block 420, a compliance repot is generated by processing the extracted input package description information using a trained compliance checking model. For example, the analysis engine 210 processes the input package description information 214 using a trained compliance checking model, such as compliance checking model 1 14, to obtain compliance evaluation results. In an example, the compliance evaluation result indicates the degree of adherence of the input packaged product to specific regulatory guidelines or requirements. It may include assessments of individual compliance attributes, such as the presence, size, style, color, and location of required information on the product packaging. These results serve as the basis for determining whether the product meets various regulatory recommendations or requirements pertaining to its category.

[0074] The analysis engine 210 then analyzes these compliance evaluation results to determine the compliance status for a plurality of recommendations pertaining to the product category of the input packaged product. Based on this analysis, the analysis engine 210 generates a compliance report, such as compliance report 216, indicating whether each recommendation is met by the input packaged product.

[0075] FIG. 9 illustrates an exemplary view of a compliance report, such as compliance report 216. The compliance report 216 as depicted in FIG. 9 further include certain general recommendations for guiding the user to make the input package product in compliance with respect the regulatory guidelines as depicted in FIG. 10.

[0076] Although examples for the present disclosure have been described in language specific to structural features and / or methods, it is to be understood that the appended claims are not necessarily limited to the specific features or methods described. Rather, the specific features and methods are disclosed and explained as examples of the present disclosure.

Claims

I / We Claim:1 . A system comprising: a processor; and an analysis engine coupled to the processor, wherein the analysis engine is to: obtain an input package image of an input packaged product which is under evaluation for checking compliance with respect to a regulatory guideline; extract an input package description information from the input package image, wherein the input package description information comprises an attribute value corresponding to a regulatory compliance attribute of the packaged product; process the input package description information using a compliance checking model to obtain a compliance evaluation result, wherein the compliance checking model is trained based on a training set of product package images with each image associated with a reference product category; and based on the compliance evaluation results, generate a compliance report.

2. The system as claimed in claim 1 , wherein the regulatory compliance attribute is one of a presence, a size, a style, a color, and a location, or combination thereof, of a recommended descriptive information on the package of the packaged product.

3. The system as claimed in claim 2, wherein the recommended descriptive information is one of a detail of the manufacturer, a packer, an importer, consumer complaints number, relevant dates, sale price in different units, quantity of packaged product in different units, dimension of the packaged product, dimension of the text written on the label, and type of packaged product.

4. The system as claimed in claim 1 , wherein the input packaged product is one of a cosmetic product, drug product, electronic product, fabric product, food product, fuel product, oil product, sheet product, and liquid chemical product.

5. The system as claimed in claim 1 , wherein the analysis engine is to: analyze the compliance evaluation result to determine compliance status for a recommendation pertaining to the product category of the packaged product; and based on the analysis, generate the compliance report indicating whether the recommendation is met by the input packaged product.

6. The system as claimed in claim 1 , wherein to obtain the input package image, the analysis engine is to: obtain the input package image with a reference object of standard dimension placed alongside the input packaged product, wherein the reference object is one of a payment card, standard-sized coin, ruler, and paper currency note; and determine dimensions of the input packaged product using the standard dimensions of the reference object in the input package image.

7. The system as claimed in claim 6, wherein upon obtaining the input package image, the analysis engine is to: render a first set of questions on a display to obtain information from a user indicating category and sub-category of the packaged product; and receive a first set of user inputs corresponding to the first set of questions from the user.

8. The system as claimed in claim 7, wherein upon receiving the first set of user inputs, the analysis engine is to:render a second set of questions to obtain information from the user indicating package related details, wherein one or more second set of questions are pre-filled based on the dimension of the input packaged product determined based on the standard dimension of the reference object, wherein the package related details comprises at least one of dimensions of the package, dimension of the text written on the label, surface area of a principal image of the packaged product, and a generic name of the input packaged product; and receive a second set of user inputs corresponding to the second set of questions from the user.

9. The system as claimed in claim 8, wherein upon receiving the second set of user inputs, the analysis engine is to: render a third set of questions to obtain information from the user indicating whether the packaged product is for retail sale or wholesale, uses of the packaged product, range of net quantity of pre-packaged commodity, surface on which label information is molded, and country of manufacturing; and receive a third set of user inputs corresponding to the third set of questions from the user.

10. The system as claimed in claim 9, wherein upon receiving the third set of user inputs, the analysis engine is to: extract the input package description information from the input package image of the packaged product using the first, second and third set of user inputs.1 1. A method comprising: obtaining a training information comprising a training product package image, a reference product category, and a relevant regulatory guideline;extracting a training package description information from the training product package image, wherein the training package description information comprises an attribute value corresponding to a regulatory compliance attribute of a training packaged product; and training a compliance checking model based on the training package description information, the reference product category, and the relevant regulatory guideline, wherein the compliance checking model, when trained, is to provide compliance evaluation results for an input packaged product.

12. The method as claimed in claim 1 1 , wherein the regulatory compliance attribute is one of a presence, a size, a style, a color, and a location, or combination thereof, of a recommended descriptive information on the package of the packaged product.

13. The method as claimed in claim 12, wherein the recommended descriptive information is one of a detail of the manufacturer, a packer, an importer, consumer complaints number, relevant dates, sale price in different units, quantity of packaged product in different units, dimension of the packaged product, dimension of the text written on the label, and type of packaged product.

14. The method as claimed in claim 1 1 , wherein the compliance evaluation result is used to generate a compliance report indicating compliance status corresponding to a recommendation.

15. The method as claimed in claim 1 1 , wherein the input packaged product is one of a cosmetic product, drug product, electronic product, fabric product, food product, fuel product, oil product, sheet product, and liquid chemical product.

Images