Improvements in food safety

Abstract

A method of detecting an irritant in an ingredient for a food product. The method comprises the steps of receiving an input including an identification of an ingredient; querying a list of stored ingredients with the input, wherein a stored ingredient is associated with one or more components, wherein a component is a component of food product and the stored ingredient is a food product that comprises the one or more components; receiving a response, wherein the response includes at least one stored ingredient that matches the ingredient identified in the input; and the one or more components associated with the stored ingredient; comparing the one or more components of the stored ingredient with one or more components in one or more predetermined component classifications; and transmitting a first notification to food processing system if at least one of the one or more components associated the stored ingredient included in the response matches a component in a component classification.

Description

[0001] Improvements in food safety

[0002] Technical Field

[0003] The present disclosure relates to improvements in food manufacturing safety. In particular, the present the disclosure is directed towards the detection irritants in food products.

[0004] Background

[0005] The food industry has experienced significant growth and diversification in recent years, with an increasing demand for high-quality, safe, and nutritious food products. An important step in the manufacture of a food product is that of ensuring that the ingredients used to make the food product do not cause the production of a harmful product to those purchasing the food product.

[0006] As a result, the need for effective and efficient methods for detecting irritants in food ingredients has become increasingly important. As used in this disclosure, the term ‘irritants’ is intended to include allergens, contaminants, or other substances that may cause adverse reactions to those consuming food products. Ensuring the safety and quality of food products is not only a legal requirement but also a moral obligation for food manufacturers and processors.

[0007] One of the primary challenges in the food industry is the accurate identification and classification of ingredients and their associated components. This is particularly important for the detection of irritants, as the presence of an irritant in a food product can have severe, and possibly fatal, consequences for individuals with allergies. In addition, the accurate identification of ingredients is essential for meeting the needs of consumers with specific dietary requirements, such as vegetarians and vegans.

[0008] At present, detecting irritants in food ingredients involves manual inspection and analysis of ingredient lists and product labels to determine if the ingredients used in a food manufacturing process contain any irritants. This is time-consuming, labour-intensive, and prone to human error.

[0009] Furthermore, the increasing complexity of food products and the globalization of food supply chains have made it more difficult to accurately track and monitor the presence of irritants in food ingredients. This is complicated by ingredients being made up from other sub-ingredients. For example, “Yogurt” is sometimes listed as an ingredient, but the ingredients of yogurt are often not clarified. For example, it is not always clear if the type of milk used to make a yogurt and thus it is unclear if the yogurt is lactose free. For the sake of clarity, sub-ingredients are referred to in this disclosure as components.

[0010] There are also further limitations and challenges associated with existing irritant detection systems. For example, the accuracy and reliability of these systems can be affected by the quality and consistency of the input data. In addition, the effectiveness of these systems can be limited by the complexity of food products and the variability of ingredient composition.

[0011] Moreover, existing systems are not able to effectively handle large volumes of data or adapt to changes in data structures and formats. This can result in delays and inefficiencies in the detection process, as well as increased risk of errors and false positives. Furthermore, the integration of these systems with existing food processing and manufacturing systems can be challenging, as integration may require significant modifications and customizations of ingredients lists required to accommodate the specific needs and requirements of different food processors and manufacturers.

[0012] Improvements are desired to overcome shortcomings of existing implementations, such as enhancing the accuracy and reliability of detection methods, improving the efficiency and scalability of data processing and analysis, and facilitating the seamless integration of irritant detection systems with existing food processing and manufacturing systems.

[0013] Summary

[0014] In general terms, the present disclosure is directed to a method for detecting an irritant in a food ingredient by querying a list of stored ingredients and their associated components, comparing those components to predetermined classifications, and notifying a food processing system if an irritant is detected. Advantageously, this invention addresses the problem of detecting discrepancies between ingredient data and allergen or lifestyle declarations in food products, ensuring accurate information for consumers and preventing potential health risks.

[0015] According to an aspect of the invention, there is provided a method of detecting an irritant in an ingredient for a food product. The method comprises the steps of: receiving an input including an identification of an ingredient; querying a list of stored ingredients with the input, wherein a stored ingredient is associated with one or more components, wherein a component is a component of food product and the stored ingredient is a food product that comprises the one or more components; receiving a response, wherein the response includes at least one stored ingredient that matches the ingredient identified in the input and the one or more components associated with the stored ingredient; comparing the one or more components of the stored ingredient with one or more components in one or more predetermined component classifications; and transmitting a first notification to food processing system if at least one of the one or more components associated the stored ingredient included in the response matches a component in a component classification.

[0016] Optionally, the input includes a list of one or more components associated with the ingredient. The one or more components associated with the ingredient in the input are compared with the one or more components associated with the stored ingredients. If the one or more components associated with the ingredient in the input do not match the one or more components associated with the stored ingredient, a second notification is transmitted.

[0017] Optionally, the method comprises adding at least one of the one or more components associated with the ingredient in the input to the one or more components associated which a stored ingredient.

[0018] Optionally, the method further includes reformatting the input, wherein reformatting the input includes reformatting the identification of the ingredient, or a component of the ingredient, to a predetermined format, and the step of querying uses the reformatted input to query the list of stored ingredients.

[0019] Optionally, the step of receiving an input includes receiving a list of ingredients of an item and generating an input for an ingredient in the list, wherein the input includes information identifying the ingredient and information identifying the ingredient’s position in the list.

[0020] Optionally, the step of querying a list of stored ingredients with the input comprises using an artificial intelligence engine to generate a probability score for a stored ingredient. The probability score is indicative of the likelihood of the stored ingredient matching the ingredient identified in the input.

[0021] Optionally, the method further comprises selecting a stored ingredient as a match of the ingredient identified in the input. Selecting comprises selecting a stored ingredient with a probability score which crosses a predetermined threshold.

[0022] Optionally, the irritant is a food item that causes a food allergy, and the one or more component classifications includes a classification of components that are a cause of a food allergy or a food intolerance. Optionally, the irritant is a food item that is contrary to a predetermined diet, and the one or more component classifications includes a classification of components that are contrary to the predetermined diets.

[0023] Optionally, the predetermined diet is any one of vegetarian, vegan, pescetarian, pollotarian, paleo, Islamic, kosher, Jain, and Mormon.

[0024] Optionally, the method comprises transmitting a third notification to the food processing system if at least one of the one or more components associated with the stored ingredient included in the response does not match any components in a component classification. The third notification identifies the component classification having no components which the one or more components associated with the stored ingredient.

[0025] Optionally, the notification includes information identifying the stored ingredient included in the response.

[0026] Optionally, the step of transmitting the notification includes transmitting information identifying the stored ingredient and the one or more components associated the stored ingredient included in the response that match a component in a component classification.

[0027] According to an aspect of the invention, there is provided a computer program product comprising instructions. When executed by a computing system, these instructions cause the system to perform a method as described above.

[0028] According to an aspect of the invention, there is provided a food processing system. The food processing system comprises means for detecting an irritant in an ingredient for a food product. The food processing system comprises an input device for receiving an input including an identification of an ingredient, a database engine for querying a list of stored ingredients with the input, wherein each stored ingredient is associated with a stored set of one or more components which are contained within the stored ingredient, a processor for receiving a response, wherein the response includes at least one stored ingredient that matches the ingredient identified in the input and the one or more components associated with the stored ingredient, a processor for comparing the one or more components of the stored ingredient with one or more components in one or more predetermined component classifications, and an output device for providing a first notification if at least one of the one or more components associated the stored ingredient included in the response matches a component in a component classification. Brief Description of the Drawings

[0029] The invention will be more clearly understood from the following description of an embodiment thereof, given by way of example only, with reference to the accompanying drawings, in which:-

[0030] Figure 1 shows a flow chart illustrating a method in accordance with the present disclosure; and

[0031] Figure 2 shows a system in accordance with the present disclosure.

[0032] Detailed Description

[0033] Figure 1 shows a flowchart for performing a method 1000 according to the present disclosure. The method 1000 includes receiving an input 1100 including an identification of an ingredient, querying 1200 a list of stored ingredients with the input 1100, receiving a response 1300, comparing 1400 the one or more components of the stored ingredient with one or more components in one or more predetermined component classifications, and transmitting 1500 a first notification to a food processing system if at least one of the one or more components associated with the stored ingredient included in the response 1300 matches a component in a component classification.

[0034] Advantageously, the method 1000 provides an efficient and precise approach to identifying ingredients and their associated components within a food processing system. This is achieved, at least in part, by including data from a plurality of sources or systems. The data may be obtained from different stages of a food production process, such as e.g. ingredient suppliers, manufacturing facilities, quality control laboratories, distribution centres, etc. By including data from multiple sources, method 1000 has access to more up-to-date and accurate information on ingredients and their components than prior art solutions. Data may also be obtained using automated means, e.g. using a web-crawler to extract ingredient information from an ingredient supplier’s website.

[0035] The input 1100 data received from these sources is subjected to a pre-processing and normalization procedure. The pre-processing and normalization procedure is configured to convert the data into a first predetermined format such that a computerised system can be controlled to compare data relating to ingredients from different data sources. In particular, pre-processing and normalization procedure preferably eliminates customer-specific formatting, or extraneous data that may not be relevant to the detection mechanisms, or both. This pre-processing step is used to standardise the data to facilitate querying 1200 the list of stored ingredients, thereby significantly enhancing the accuracy and reliability of the method 1000.

[0036] Examples of pre-processing techniques employed in the method 1000 include the removal of invalid characters, such as special symbols or non-alphanumeric characters that may be present in the input 1100 data due to typographical errors or inconsistencies in data entry. Additionally, the pre-processing and normalization procedure may be configured to eliminate line-breaks which may be present in the input 1100 data as a result of formatting inconsistencies or data transfer issues between different systems or platforms and incorrectly interpreted as delimiting a data field. This ensures that the input 1100 data is stored in a standardized form, facilitating more accurate comparisons with the stored ingredients than possible using prior art solutions.

[0037] The method 1000 preferably also addresses non-standard accent characters, which may be present in the input 1100 data due to variations in language or regional dialects. By normalizing these accent characters to a standardized format, the method 1000 ensures that the input 1100 data is consistent and easily comparable with the stored ingredients, regardless of the original language or dialect in which the data was entered.

[0038] Furthermore, the method 1000 effectively handles multiple whitespaces that may appear in quick succession within the input 1100 data. These extraneous whitespaces can be the result of formatting inconsistencies, data entry errors, or data transfer issues between different systems. By consolidating these multiple whitespaces into a single whitespace, the method 1000 further improves the standardization of the input 1100 data, further improving the accuracy of the ingredient identification process.

[0039] In addition to these pre-processing techniques, the method 1000 preferably is configured to checks for any inverse declarations within the input 1100 data, such as instances where an ingredient is declared as being absent when it is actually present, or vice versa. These inverse declarations can be the result of data entry errors, miscommunication between different systems, or other inconsistencies in the data.

[0040] In addition, if these inconsistencies cannot be resolved by the system, they are preferably flagging for review by e.g. an integration team. As a result, the method 1000 ensures that these issues are promptly addressed and resolved, further enhancing the accuracy and reliability of the ingredient identification process. The input 1100 may further include a list of one or more components associated with the ingredient included in the input 1100. The method 1000 may further comprise comparing 1400 the one or more components associated with the ingredient in the input 1100 with the one or more components associated with the stored ingredients, and transmitting 1500 a second notification if the one or more components associated with the ingredient in the input 1100 do not match the one or more components associated with the stored ingredient.

[0041] Advantageously, this allows for a more comprehensive comparison of the components associated with the input 1100 ingredient and the stored ingredients, which in turn facilitates the identification of discrepancies between the stored ingredient components and the input 1100 ingredient components, ensuring that any potential issues or inconsistencies are promptly addressed. This feature contributes to the platform's efficiency in dealing with potential issues, allowing them to be rectified straight away before the data ends up downstream with consumers. The notification can be presented on a web dashboard, enabling users to respond and optionally upload fact sheets or data on the input 1100 ingredient components to agree or disagree with the stored ingredient components. Any disagreements can then be fed back into the system to enhance the detection mechanisms where possible.

[0042] The method 1000 uses a deterministic approach to interrogate the ingredient data directly, detecting differences between the product data and the declaration. This step helps prevent missing declarations when the ingredients are complete. The method 1000 also employs a probabilistic approach using a machine learning service for detections when ingredients or allergens have been omitted from the ingredients data entirely. This combination of deterministic and probabilistic approaches ensures a more accurate and reliable identification of ingredients and their associated components.

[0043] Optionally, the method 1000 comprises adding at least one of the one or more components associated with the ingredient in the input 1100 to the one or more components associated with a stored ingredient. Advantageously, this allows for the continuous updating and improvement of the stored ingredients list, ensuring that the method 1000 remains accurate and up- to-date as new ingredients and components are identified. This updating process also helps enhance the detection mechanisms of the method 1000 by incorporating any disagreements from customers into the system.

[0044] Optionally, the method 1000 further includes reformatting the input 1100, wherein reformatting the input 1100 includes reformatting the identification of the ingredient, or a component of the ingredient, to a predetermined format. The step of querying 1200 then uses the reformatted input 1100 to query the list of stored ingredients. This ensures that the input 1100 data is in a consistent and standardized format, improving the accuracy and efficiency of the querying 1200 process. The reformatting step can include removing extraneous data, such as non-standard accent characters, html entities or tags, and multiple whitespaces in quick succession. This standardized format allows for more accurate comparisons between the input 1100 ingredient and the stored ingredients, ultimately leading to more reliable identification of ingredients and their associated components.

[0045] Optionally, receiving an input 1100 includes receiving a list of ingredients of an item and generating an input 1100 for an ingredient in the list, wherein the input 1100 includes information identifying the ingredient and information identifying the ingredient’s position in the list. This facilitates efficient processing of multiple ingredients within a single item, streamlining the method 1000 for identifying and analysing each ingredient and its associated components. By including information identifying the ingredient's position in the list, the method 1000 can better manage the processing of each ingredient and ensure accurate comparisons with the stored ingredients. This approach also enables the system to handle large updates or peak hours effectively, as the input 1100 data can be queued for processing to manage server load.

[0046] The step of querying 1200 a list of stored ingredients with the input 1100 may comprise using an artificial intelligence (Al) engine to generate a probability score for a stored ingredient, the probability score indicative of the likelihood of the stored ingredient matching the ingredient identified in the input 1100. By leveraging Al, the accuracy and efficiency of ingredient identification can be improved. The Al engine uses a logistic regression model that classifies products based on a bag of words representation of the product name, which is trained using any suitably verified test data. This approach allows the system to predict potential allergens or components related to certain words or n-grams, enhancing the overall detection capabilities of the method 1000. Preferably, selecting a stored ingredient as a match of the ingredient identified in the input 1100. Selecting comprises selecting a stored ingredient with a probability score which crosses a predetermined threshold. This ensures that only stored ingredients with a high probability of matching the input 1100 ingredient are selected, improving the overall accuracy and reliability of the method 1000. By setting a predetermined threshold, the system can filter out potential false positives and focus on the most likely matches, ultimately leading to more accurate identification of ingredients and their associated components.

[0047] Where the irritant is a food item that causes a food allergy, the one or more predetermined component classifications can include a predetermined classification of components that cause a food allergy or a food intolerance. This facilitates the targeting and identification of ingredients and components that may cause food allergies or intolerances. By focusing on these classifications, the system can provide valuable information to stakeholders, enabling them to improve food safety. This approach contributes to the platform's efficiency in dealing with potential issues, allowing for early identification and prevention of adverse effects on consumers. Where the irritant is a food item that is contrary to a predetermined diet, the one or more predetermined component classifications can include a classification of components that are contrary to the predetermined diet. This enables identification of ingredients and components that may not be suitable for specific diets. By classifying components based on their compatibility with predetermined diets, the system can provide valuable information to stakeholders and consumers, ensuring that dietary restrictions and preferences are respected and maintained.

[0048] A third notification may be sent to the food processing system if the at least one of the one or more components associated with the stored ingredient included in the response 1300 does not match any components in a component classification, wherein the third notification identifies the component classification having no components which match the one or more components associated with the stored ingredient. This allows a user to identify if a product is suitable for a particular diet. For example, a snack bar designed to comply with a paleo diet may be found to contain no components derived from animals, in which case the method 1000 can alert a user that the snack bar is also suitable for vegans.

[0049] The input 1100 may include information identifying an ingredient’s position in a list of ingredients of a food product. This provides additional information to the food processing system, allowing for a more comprehensive understanding of the stored ingredient and its associated components. This feature enhances the review process for stakeholders by providing a clear pointer to the offending ingredient, by including a "snippet" of the source data where the ingredient was found. This additional information can help stakeholders make more informed decisions when addressing potential issues or discrepancies.

[0050] Preferably, the first notification includes information identifying the stored ingredient included in the response 1300 and the one or more components associated with the stored ingredient included in the response 1300 that match a component in a component classification. This provides a more detailed notification to the food processing system, allowing for a better understanding of the stored ingredient and its associated components that match a component classification. This feature further enhances the review process for stakeholders by providing more comprehensive information about the stored ingredient and its matching components, ultimately leading to more accurate and reliable identification of ingredients and their associated components. Figure 2 illiterately shows a food processing system 2000 according to the present disclosure. The system 2000 includes an input device 2100 for receiving an input 1100 including an identification of an ingredient, a database engine 2200 for querying 1200 a list of stored ingredients with the input 1100, and a processor 2300 for receiving a response 1300. The response 1300 includes at least one stored ingredient that matches the ingredient identified in the input 1100 and the one or more components associated with the stored ingredient. The processor 2300 then compares the one or more components of the stored ingredient with one or more components in one or more predetermined component classification. Finally, an output device 2400 provides a first notification if at least one of the one or more components associated with the stored ingredient included in the response 1300 matches a component in a component classification.

[0051] Advantageously, the food processing system 2000 enables efficient and accurate identification of ingredients and their associated components by integrating various sources of data from different stages of the food production process. The system 2000 preprocesses and normalizes the input 1100 data to remove customer-specific formatting or extraneous data not relevant to the detection mechanisms, such as invalid characters, new lines, non-standard accent characters, html entities or tags, and multiple whitespaces in quick succession. This preprocessing step ensures that the data is in a suitable format for querying 1200 the list of stored ingredients and improves the accuracy of the system 2000. The system 2000 also interrogates any inverse declarations and flags any unexpected values for review by the integration team.

[0052] The food processing system 2000 further employs a combination of deterministic and probabilistic approaches to accurately identify ingredients and their associated components. The deterministic approach involves directly interrogating the ingredient data and detecting differences between the product data and the declaration. The probabilistic approach uses a machine learning service for detections when ingredients or allergens have been omitted from the ingredients data entirely. This combination of approaches ensures a more accurate and reliable identification of ingredients and their associated components.

[0053] In addition, the food processing system 2000 allows for continuous updating and improvement of the stored ingredients list by incorporating any disagreements from customers into the system 2000. This updating process enhances the detection mechanisms of the system 2000 and ensures that it remains accurate and up-to-date as new ingredients and components are identified. Furthermore, the food processing system 2000 provides alerts to stakeholders based on the results of the ingredient identification process, with "dangerous" issues being alerted immediately. This contributes to the platform's efficiency in dealing with potential issues so they can be rectified straight away before the data ends up downstream with consumers. The issues are presented on a web dashboard so customers can respond, optionally uploading fact sheets or data to agree or disagree with the claims. This feedback loop helps enhance the detection mechanisms of the system 2000 where possible.

[0054] Furthermore, the information provided the system 2000 can be used to control portions of a production line. For example, data which has been converted to the first predetermined format may be provided to a labelling system to automatically update the labelling of a product based on a change in an ingredient used for the production of a food product.

[0055] It will be understood that the term "artificial intelligence engine" as used herein may refer to a computer system 2000 or software application that utilizes machine learning algorithms, data analysis, and pattern recognition to process and analyse large amounts of data, such as the list of stored ingredients and their associated components, in order to make predictions or decisions related to the detection of irritants in food ingredients.

[0056] It will be understood that the term "probability score" as used herein may refer to a numerical value or percentage that represents the likelihood of a particular food ingredient containing an irritant, based on the comparison of its components to predetermined classifications. It will be understood that the term "predetermined threshold" as used herein may refer to a specific limit or value set in advance, which, when exceeded or met by the concentration or presence of an irritant in a food ingredient, triggers a notification to the food processing system 2000.

[0057] It will be understood that the term "predetermined diet" as used herein may refer to a specific set of food items, ingredients, or nutritional guidelines that have been established in advance to meet certain dietary requirements, restrictions, or preferences, such as allergen-free, gluten-free, or low-sodium diets.

[0058] It will be understood that the term "computer program product" as used herein may refer to a non-transitory, tangible medium containing software code or instructions, such as a CD-ROM, USB drive, or downloadable file, which when executed by a computer or processor 2300, enables the implementation of the disclosed method 1000 for detecting irritants in food ingredients.

[0059] It will be understood that the term "component" as used herein may refer to any component of an ingredient, including food additives such as those classified using European numbering (e numbers) and other such chemicals. In some cases, an ingredient (e.g. water) will only contain a single component.

[0060] It will be appreciated by the person of skill in the art that various modifications may be made to the above-described examples without departing from the scope of the invention as defined by the appended claims.

Claims

Claims1. A computer implemented method of detecting an irritant in an ingredient for a food product comprising: receiving an input including an identification of an ingredient; querying a list of stored ingredients with the input, wherein a stored ingredient is associated with one or more components, wherein a component is a component of a food product and the stored ingredient is a food product that comprises the one or more components; receiving a response, wherein the response includes: at least one stored ingredient that matches the ingredient identified in the input; and the one or more components associated with the stored ingredient; comparing the one or more components of the stored ingredient with one or more components in one or more predetermined component classifications; and transmitting a first notification to a food processing system if at least one of the one or more components associated with the stored ingredient included in the response matches a component in a component classification.

2. The method of claim 1, wherein: the input further includes a list of one or more components associated with the ingredient included in the input; comparing the one or more components associated with the ingredient in the input with the one or more components associated with the stored ingredients; and transmitting a second notification if the one or more components associated with the ingredient in the input do not match the one or more components associated with the stored ingredient.

3. The method of claim 2, comprising adding at least one of the one or more components associated with the ingredient in the input to the one or more components associated with a stored ingredient.

4. The method of any preceding claim, further including:reformatting the input, wherein reformatting the input includes reformatting the identification of the ingredient, or a component of the ingredient, to a predetermined format, wherein: the step of querying uses the reformatted input to query the list of stored ingredients.

5. The method of any preceding claim, wherein receiving an input includes: receiving a list of ingredients of an item; and generating an input for an ingredient in the list, wherein the input includes: information identifying the ingredient; and information identifying the ingredient’s position in the list.

6. The method of any preceding claim, wherein the step of querying a list of stored ingredients with the input comprises using an artificial intelligence engine to generate a probability score for a stored ingredient, the probability score indicative of the likelihood of the stored ingredient matching the ingredient identified in the input.

7. The method of claim 6, further comprising selecting a stored ingredient as a match of the ingredient identified in the input, wherein selecting comprises selecting a stored ingredient with a probability score which crosses a predetermined threshold.

8. The method of any preceding claim, wherein the irritant is a food item that causes a food allergy, and the one or more component classifications includes a classification of components that are a cause of a food allergy or a food intolerance.

9. The method of any preceding claim, wherein the irritant is a food item that is contrary to a predetermined diet, and the one or more component classifications includes a classification of components that are contrary to the predetermined diets.

10. The method of claim 9, wherein the predetermined diet is any one of the following diets: vegetarian, vegan, pescetarian, pollotarian, paleo, Islamic, Kosher, Jain, and Mormon.

11. The method of any preceding claim, comprising transmitting a third notification to a food processing system if at least one of the one or more components associated with the stored ingredient included in the response does not match any components in a component classification, wherein the third notification identifies the component classification having no components which match the one or more components associated with the stored ingredient.

12. The method of any preceding claim, wherein the first nonfiction includes information identifying the stored ingredient included in the response.

13. The method of any preceding claim, wherein the step of transmitting the first notification includes transmitting information identifying: the stored ingredient included in the response; and the one or more components associated with the stored ingredient included in the response that match a component in a component classification.

14. A computer program product comprising instructions which, when executed by a computing system, cause the system to perform a method according to any preceding claim.

15. A food processing system comprising means for detecting an irritant in an ingredient for a food product comprising: an input device for receiving an input including an identification of an ingredient; a database engine for querying a list of stored ingredients with the input, wherein each stored ingredient is associated with a stored set of one or more components which are contained within the stored ingredient; a processor for a receiving a response, wherein the response includes: at least one stored ingredient that matches the ingredient identified in the input; and the one or more components associated with the stored ingredient; a processor for comparing the one or more components of the stored ingredient with one or more components in one or more predetermined component classifications; and an output device for providing a first notification if at least one of the one or more components associated with the stored ingredient included in the response matches a component in a componentclassification.

Images