RU2800013C2 - Devices and methods for presenting pet food recommendations - Google Patents

Abstract

FIELD: pets; computer technologies.

SUBSTANCE: invention relates to methods and apparatus for presenting pet food recommendations. An image of a pet is analysed using a pet image recognition model to determine one or more characteristics. Behavioural data of the pet are obtained. Using the pet food recommendation model, pet characteristics, and behavioural data, pet food recommendations and/or pet feeding recommendations are generated.

EFFECT: presenting a recommendation for a food product for a pet and/or a recommendation for feeding a pet.

20 cl, 3 dwg, 6 tbl

Description

Cross-reference to related applications

This application claims priority over U.S. Provisional Application No. 62/800,199, filed February 1, 2019, the disclosure of which is incorporated herein by reference in its entirety.

Technical field

The present description relates generally to methods and apparatuses for pet foods. More specifically, the present disclosure relates to methods and apparatus for presenting a pet food recommendation and/or a pet food recommendation.

Prerequisites for the creation of the invention

Many pet owners take good care of their pets and understand that decisions about what and how much to feed their pets can contribute to a longer, happier, and healthier life together. However, the characteristics of each pet and the conditions and preferences of the pet and owner are unique. Moreover, these characteristics, circumstances and preferences change over time. For this reason, consumers are looking for choices about what to feed their pets and information to help them make decisions. However, the abundance of products and sources of information can complicate this process and leave consumers with open questions and emotions. Lack of information or changing conditions can also mean that a feeding decision at one stage in a pet's life may not be optimal at another stage.

Existing solutions are mostly provided by retailers or manufacturers. Retailers' decisions are based on user-defined user inputs and are primarily based on filtering criteria, statistics, feedback, and template data. Filtering processes may narrow down the selection based on the characteristics of the product available, but do not take into account sufficient information to provide the optimal solution for a particular pet. Moreover, existing solutions do not provide preventive recommendations that take into account changes in the body of a pet over time. Product selectors provided by manufacturers have many of these characteristics and disadvantages. Even in cases where a manufacturer can gather more information about a particular pet, the recommendations may still only be product recommendations and may not provide practical advice to owners. Also, these recommendations may only be appropriate for a specific point in time (i.e. recommendations are not preventive and/or do not change over time). In addition, recommendations tend to be limited to the product range of a single pet food manufacturer.

The essence of the invention

The present disclosure proposes advantages and solutions to problems associated with current pet food recommendation technologies. As such, the pet food recommendation system may contain the most up-to-date information about the pet owner and the pet in order to standardize the decision-making process and provide personalized pet food recommendations and/or pet feeding recommendations. Personalized pet food recommendations and/or pet feeding recommendations may be provided initially and/or proactively throughout the life of the pet. For example, a pet food and/or pet food recommendation can help a pet owner move from their current diet and feeding advice to a diet that can help maintain or change physical condition, with the ultimate goal of achieving that particular pet's ideal physical condition.

In one embodiment, personalized pet food recommendations and/or pet feeding recommendations may take into account the physical condition of the pet, deviation from the expected weight of the pet, activity level of the pet, anatomy of the pet, breed and/or breed size of the pet. For example, the breed of a pet may provide information about certain characteristics and general parameters that can be expected in relation to those characteristics, such as weight range and activity levels, but each individual pet will have its own unique weight and/or activity level. Personalized recommendations for feed products for pets and/or recommendations for feeding pets can additionally take into account the preferences of the owner of the home animal, the needs of the domestic animal and/or preferences of the pets (for example, the presence and/or the absence of certain ingredients, such as cereals or protein, in pets feed, a degree that the product is made of natural implerance, incompatibility Due to a potential allergic reaction or sensitivity and retail sales of a fodder product for pets), to personalize the recommendation for a pets and the owner of a pet, due to a potential allergic reaction or sensitivity. In addition, pet food recommendations and/or pet feeding recommendations may additionally take into account conformation and/or morphometric characteristics as described in US Pat.

In one embodiment, the pet may be a dog, cat, bird, rodent, horse, pig, fish, reptile, and/or any other domestic and/or domesticated animal.

Accordingly, one aspect of the description relates to a method including: obtaining an image showing a pet; analyzing, with the pet image recognition model, the pet image to determine one or more characteristics of the pet; receiving one or more types of input associated with a pet; analyzing, with the pet food recommendation model, the characteristics of the pets and the input data associated with the pets to generate a pet food recommendation and/or a pet feeding recommendation. In some embodiments, the characteristics of the pet include one or more of breed, breed size, pet size, physical condition, developmental stage, activity level, pet sex, pet sex status, and pet weight. In some embodiments, the pet-related input may comprise one or more of pet preferences, user preferences, and/or pet characteristics. In one embodiment, pet and/or user preference includes one or more of cereal preference, protein preference, food consistency, natural ingredient preference, and purchase method preference. In another embodiment, the pet and/or user preference is selected from the group consisting of cereal preference, protein preference, food consistency, natural ingredient preference, and purchase method preference. In some embodiments, the pet-related input contains characteristics of the pet. In one embodiment, one or more pet characteristics are determined by the pet image recognition model, and one or more additional pet characteristics are input by the user.

In some embodiments, the pet image is obtained from the user device. In one embodiment, the user device is a mobile device, such as a mobile phone. In another embodiment, the user device is a laptop or desktop computer.

In some embodiments, analyzing the pet image further includes: analyzing the pet image to detect one or more characteristics of the pet image, and determining one or more characteristics of the pet based on the image characteristics. In some embodiments, the image features are selected from the group consisting of the pet outline, the area covered by the pet outline, the percentage of pet image pixels per pet, one or more pet geometries in the pet image. In some embodiments, the dimensions are selected from length, height, leg-to-shoulder distance, back-to-belly distance, chest width, chest depth.

In some embodiments, the method further includes: filtering the pet food list based on at least one user preference to create a filtered pet food list. In some embodiments, at least one user preference includes at least one preference selected from the group consisting of grain preference, protein preference, food texture preference, natural ingredient preference, and purchase method preference.

In some embodiments, the pet characteristics include one or more characteristics selected from the group consisting of breed, breed size, pet size, physical condition, developmental stage, activity level, pet sex, pet sex status, and pet weight.

In some embodiments, the method further includes calculating one or more variations selected from the group consisting of breed size variation, pet size variation, physical condition variation, developmental stage variation, activity level variation, and pet weight variation.

In some embodiments, the method further includes calculating the plurality of variations to generate the plurality of calculated variations and evaluating the food list based on the sum of the calculated variations to identify one or more recommended pet foods.

In some embodiments, the method further includes providing the user with a pet food recommendation, wherein the pet food recommendation contains a plurality of pet foods; receiving information from the user about a selected pet food product from a plurality of pet food products; and generating a feeding recommendation for the pet based on the selected pet food product. In one embodiment, the pet food recommendation is generated based on the calorie content of the selected pet food product.

In some embodiments, the method further includes storing behavioral data of the pet and/or user in the history database and training one or both of the pet image recognition model and the pet food recommendation model based on the behavioral data stored in the history database.

Another aspect of the description relates to a system comprising: a processor and a storage device that stores instructions that, when executed by the processor, cause the processor to obtain a pet image, which depicts a pet; parsing, with the pet image recognition model, an image of the pet to determine one or more characteristics of the pet; accepts one or more pet-related input types; analyzes, with the pet food recommendation model, pet characteristics and input data associated with the pet to generate a pet food recommendation and/or a pet feeding recommendation.

In some embodiments of the system, additional instructions are stored in memory which, when executed by the processor, cause the processor to: perform an analysis of the pet image to detect one or more characteristics of the pet image, and determine one or more characteristics of the pet based on the image characteristics.

In some embodiments of the system, additional instructions are stored in memory which, when executed by the processor, cause the processor to: filter the pet food list based on at least one user preference to create a filtered pet food list.

In some embodiments, pet characterization systems include one or more characteristics selected from the group consisting of breed, breed size, physical condition, developmental stage, activity level, pet size, pet sex, pet sex status, and pet weight.

In some embodiments of the system, additional instructions are stored in memory which, when executed by the processor, cause the processor to: calculate one or more variations selected from the group consisting of a breed size variation, a pet size variation, a physical condition variation, a developmental stage variation, an activity level variation, and a pet weight variation.

In some embodiments of the system, additional instructions are stored in memory which, when executed by the processor, cause the processor to: calculate a plurality of variations to generate a plurality of calculated variations, and evaluate a list of pet foods based on the sum of the calculated variations to identify one or more recommended pet foods.

In some embodiments of the system, additional instructions are stored in memory which, when executed by the processor, cause the processor to: present a pet food recommendation to the user, the pet food recommendation comprising a plurality of pet food products; receives information about a selected pet food product from a plurality of pet food products from a user, and generates a feeding recommendation for the pet based on the selected pet food product.

In some embodiments of the system, additional instructions are stored in memory which, when executed by the processor, cause the processor to: store the behavioral data of the pet and/or user in the history database and train one or both of the image recognition model and the pet food recommendation model based on the behavioral data stored in the history database.

Another aspect of the invention relates to a computer-readable physical storage medium that stores instructions that, when executed by a processor, cause the processor to: receive a pet picture in which the pet is depicted; analyzes the image of the pet with the image recognition model to determine one or more characteristics of the pet; accepts one or more pet-related input types; analyzes, using the pet food recommendation model, the characteristics of the pet and input data to generate a pet food recommendation.

In one embodiment, the pet food recommendation contains one or more pet foods. In one embodiment, the pet food recommendation includes a composition intended for consumption by a pet. In some embodiments, the pet food recommendation contains one or more pet food products that the pet consumes as the main food, including, without limitation, dry, soft, semi-moist, wet, and liquid food compositions. In another embodiment, the pet food recommendation includes a pet treat. In one embodiment, the treat is food intended to be consumed by the pet as an occasional reward or reward, rather than the pet's sole source of nutrition. In another embodiment, the pet food recommendation includes a pet supplement. In one embodiment, the pet supplement is an oral composition provided separately from the food formula or treats, and is intended to feed a pet in addition to normal nutritional requirements. For example, complementary pet products may include products that provide joint, skin, coat, and/or digestive benefits.

In one embodiment, personalized pet food recommendations and/or pet feeding recommendations may use knowledge of pet food characteristics and attributes of pets and/or pet breeds to determine the most appropriate product and feeding recommendations.

In one embodiment, the personalized pet food recommendation and/or pet feeding recommendations may use image recognition technology to determine a variety of pet factors (i.e., pet breed, pet size, pet physical condition) in conjunction with machine learning models capable of generating food recommendations based on data received from the end user, data obtained from image recognition tools, and/or other data as a result of applying algorithms to account for the activity and stage of development of the pet. In one embodiment, the image used by the image recognition tools is a static or still image (eg, a digital photograph). In another embodiment, the image recognition technology uses multiple images or a dynamic image (eg, multiple individual photos or videos).

In one embodiment, the personalized pet food recommendations and/or pet feeding recommendations may be personalized and individualized, including pet food selection, as well as any previously entered pet variables (i.e., pet breed, pet size, pet physical condition) to form a daily pet food intake recommendation. In one embodiment, the pet's daily food intake may be adjusted (eg, to increase the pet's weight, decrease the pet's weight, or maintain the pet's weight).

In one embodiment, to obtain personalized pet food recommendations and/or pet feeding recommendations, artificial intelligence (e.g., image recognition) may be used to determine the type of pet and/or the breed of the pet for inclusion in the pet food recommendation and/or the pet feeding recommendation.

In one embodiment, to obtain personalized pet food recommendations and/or pet feeding recommendations, machine learning can be used to optimize the pet food recommendation and pet feeding recommendation over time. For example, machine learning models may be configured to generate one or both of a pet food recommendation and a pet feeding recommendation, and may include feedback on the quality of the recommendations and actual performance information provided by the pet owner and/or another user (e.g., a veterinarian).

In one embodiment, the device may be configured to provide personalized pet food recommendations and/or pet feeding recommendations. In such an embodiment, the device may be further configured to track the achievement of goals related to the pet's physical condition over time and influence pet food selection and the pet's feeding schedule based on the goals.

In one embodiment, to receive personalized pet food recommendations and/or pet feeding recommendations, connected pet devices (e.g., collars, bowls, scoops, litter boxes, drinkers, cages, fences, scales, baskets, mats, beds) may be used to automatically collect and transmit data used to provide personalized pet food recommendations and/or pet feeding recommendations. The data used to provide personalized pet food recommendations and/or pet feeding recommendations may be relevant to detecting changes in pet food or pet food preferences over time. Such a configuration may eliminate sources of error or time delays associated with manual data entry (e.g., user and/or pet owner input), which may allow more proactive personalization of pet food recommendations and pet feeding recommendations.

In one embodiment, the personalized pet food recommendations and/or pet feeding recommendations may use artificial intelligence and machine learning to evaluate pet waste characteristics and convert the pet waste characteristics into personalized pet food recommendations and/or pet feeding recommendations.

The features and advantages described herein are not intended to be exhaustive, and in particular, many additional features and advantages will be apparent to one of ordinary skill in the art in view of the figures and description. Moreover, it should be noted that the language used in the description has been chosen primarily for readability and methodological purposes, and not to limit the scope of the subject matter of the invention.

Brief description of graphic materials

In FIG. 1 is a block diagram of an embodiment of a system and method for presenting personalized pet food recommendations and/or pet food recommendations in accordance with the present disclosure.

In FIG. 2 shows an embodiment of an application scheme in accordance with the present disclosure.

In FIG. 3 shows an embodiment with summary information data for presenting personalized pet food recommendations and/or pet feeding recommendations in accordance with the present disclosure.

Detailed description of exemplary embodiments

This document provides detailed embodiments of devices and methods. However, it should be understood that the described embodiments are only examples of devices and methods that can be implemented in various forms. Thus, the specific functional details described herein should not be interpreted as a limitation, but only as a typical example for various uses of the present description by those skilled in the art.

As used herein, the terms "about", "about" and "substantially" mean numbers in the numerical range, for example, in the range of -10% to +10% of the specified number, preferably in the range of -5% to +5% of the specified number, more preferably -1% to +1% of the specified number, and most preferably -0.1% to +0.1% of the specified number. It should be understood that all numerical ranges herein include all integers and fractional numbers included in the range. Moreover, these numerical ranges should be considered as supporting a claim referring to any number or subset of numbers within the given range. For example, the description "1 to 10" should be considered to include the range "1 to 8", "3 to 7", "1 to 9", "3.6 to 4.6", "3.5 to 9.9", etc.

In this document, the singular forms include references to the plural, unless the context clearly indicates otherwise. For example, the use of the term "ingredient" or "method" includes many such "ingredients" or "methods". The term "and/or" used in the context of "X and/or Y" should be interpreted as "X" or "Y" or "X and Y". Similarly, "at least one of X or Y" should be interpreted as "X" or "Y", or "X and Y".

Preferred embodiments relate to devices and methods that provide personalized pet food recommendations and/or pet feeding recommendations to a user and/or pet owner.

As described herein, reliable pet and human level data, along with artificial intelligence and machine learning, enable the devices and methods described herein to be facilitated. In broad terms, the term "artificial intelligence" refers to technologies that allow computers to mimic human abilities to perceive, solve problems, act, adapt, and improve over time with experience. Image recognition and natural language processing can be seen as two specific examples or types of artificial intelligence that transform aspects of their respective environment into pieces of information that can be provided as input to computerized algorithms. Machine learning technologies can help adapt and improve the performance of an algorithm on a task or the ability to achieve a desired result over time based on statistical data.

As used herein, the term "species" refers to the standard association name for a species of domestic animal. For example, a golden retriever is classified as a dog, a Maine Coon is a cat, and a blue tang is a fish.

As used herein, the term "developmental stage" refers to the relative developmental stage of a pet in the pet's life cycle. For example, dogs may have stages of development such as "puppy", "adult dog" and "old dog". Cats can have stages of development such as "kitten", "adult cat" and "old cat".

As used herein, the term "breed size" refers to the relative size of a given breed of pet compared to other breeds of pets of the same species. For example, breed sizes may include dwarf, small, medium, large, and giant. For example, the Chihuahua is considered a dwarf breed, while the Great Dane is considered a giant breed. In some cases, a pet is difficult to attribute to a specific breed. In such cases, the pet may be more generally classified by size (eg, small, medium, or large pet).

As used herein, the terms "breed size variation" or "pet size variation" refers to the extent to which a given pet food product matches the size of a particular breed or the size of a particular pet. Breeds of pets of different sizes or pets of different sizes may have different needs. For example, dwarf or small dogs may require a lower calorie intake, usually expressed as caloric intake per unit time, eg 150 kcal/day. In addition, the calorie content of a food, which reflects the caloric value in a given volume, for example, 150 kcal/cup, influences the recommendation of how much food should be fed at a given point in time to achieve the caloric intake. By comparison, large or giant breed dogs may require food that addresses certain health characteristics specific to their size (for example, foods supplemented to address hip or digestive issues that are typically found in large and giant dog breeds).

For example, breed size variation (BSV) for a pet food product can be calculated given breed size using the formula:

where PF is a priority factor and where M is determined by relating the size of the breed for which the pet food product is intended to the size of that breed. For example, for a specified breed size, foods intended for a specified breed size may be assigned an M of 0, foods intended for all breed sizes may be assigned an M of 1, and foods intended for breed sizes bordering a specified breed size may be assigned an M of 2. An M value of 0, pet foods intended for all breed sizes will be assigned an M value of 1, and pet foods intended for dwarf and medium breed sizes will be assigned an M value of 2, as shown in the table below:

Breed size M value small 0 All breeds 1 small and medium 2

The priority factor may reflect the relative importance of breed size variation in subsequent calculations (for example, when evaluating pet food to form a pet food recommendation). Similarly, you can calculate the variation in the size of pets.

In certain implementations where breed size variation or pet size variation is very important to subsequent calculations, the priority factor may be set to 10 (out of a maximum value of 10). In this embodiment, continuing with the previous example, pet foods intended for small breeds will have a breed size variation of BSV=10/(0+1)=10, pet foods intended for all breed sizes will have a breed size variation of BSV=10/(1+1)=5, and pet foods intended for toy and medium breeds will have a breed size variation of BSV=10/(2+1)=3, 33. Pet foods intended for breeds that are significantly larger than those indicated (i.e. large and giant breeds when calculated for small breeds) may be excluded from the calculation of breed size variation.

As used herein, the term "physical condition variation" refers to how accurately a given pet food product is designed for pets of a given physical condition. For example, overweight pets may require a lower calorie content than ideal weight pets, so pet foods formulated for overweight pets may have a lower calorie content. Ceteris paribus, a lower caloric intake and/or lower caloric intake may help a pet lose weight to restore the pet's ideal body weight. Similarly, underweight pets may require a higher caloric intake than ideal weight pets. Accordingly, pet foods formulated for underweight pets may have a higher calorie content to help underweight pets gain weight and regain ideal body weight. Thus, the variation in physical condition for a given pet food product can be calculated based on the calorie content of the pet food product.

For example, the Physical Condition Variation (BCV) for a given pet food relative to the calorie content of the pet food can be calculated using the formula:

where PF is a priority factor, FAV is a value assigned to the pet food based on the calorie content of the pet food, and M is a value selected based on a given physical condition. For example, the FAV can be set based on the calorie content (calories/bowl) of the pet food according to the table below:

calories/bowl 100 125 200 300 FAV 0 0.25 1 2

Similarly, the value of M can be calculated based on a given physical state according to the table below:

Physical state M value Overweight 0 Ideal mass 1 Mass deficiency 2

The priority factor may reflect the relative importance of variation in physical condition in subsequent calculations (eg, when evaluating a pet food product to make a recommendation for a pet food product). In certain implementations where physical state variation is important to subsequent calculations, but not as important as rock size variation, the priority factor may be set to 9 (out of a maximum value of 10). In such embodiments, for an overweight pet (i.e., M=0), the highest calorie pet food product will have a BCV=9/(+1)=4.5 physical state variation, and the lowest calorie pet food product will have a BCV=9/(+1)=9 physical state variation.

As used herein, the term "developmental stage variation" refers to a comparison between the developmental stage of a pet and the developmental stage of a pet for which the pet food product is intended. For example, younger pets may require a higher calorie diet because such pets are still growing and tend to be very active, while older pets may require a lower calorie diet because older pets tend to be less active. In addition, older pets may need food that addresses the health concerns of older pets (eg, joint problems, digestive problems, heart problems). Thus, developmental stage variation can be calculated relative to a given developmental stage of the pet.

For example, developmental stage variation (LSV) can be calculated using the formula:

where PF is a priority factor and where M is determined based on the stage of development of the pet and the stage of development for which the pet food product is intended. For example, similar to the breed size variation described above, for a given developmental stage, foods destined for the same developmental stage may be assigned an M value of 0, pet foods destined for all developmental stages may be assigned an M value of 1, and pet foods destined for developmental stages bordering a specified developmental stage may be assigned an M value of 2. developments other than those described in this document may be excluded from the calculation. As another example, in such an embodiment, if the indicated developmental stage is puppy or kitten, then pet foods intended for puppies or kittens are assigned an M value of 0, pet foods intended for all developmental stages are assigned an M value of 1, and pet foods intended for the adult developmental stage are assigned an M value of 2, as shown in the table below:

Development stage M value Puppy 0 All 1 adult animal 2

Pet foods intended for the developmental stage of older animals may be excluded from the calculation and from potential pet food recommendations. The priority factor may reflect the relative importance of developmental stage variation in subsequent calculations (eg, when evaluating a pet food product to make a recommendation for a pet food product). In certain implementations where developmental stage variation is important to subsequent calculations, but not as important as rock size variation and physical condition variation, the priority factor may be set to 8 (out of a maximum value of 10). In such embodiments, and continuing with the preceding example, pet foods intended for puppies or kittens will have a developmental stage variation of LSV=8/(0+1)=8, pet foods intended for all developmental stages will have a developmental variation of LSV=8/(1+1)=4, and pet foods intended for adult stages of development will have a developmental variation of LSV=8/(2+1) =2.67.

As used herein, the term "activity variation" refers to a comparison between the activity level of a pet and the energy value of a pet food product. More active pets may require a substantially higher calorie intake to prevent excessive weight loss, and therefore may require higher calorie pet foods. Similarly, less active pets may require substantially fewer calories to avoid gaining excess weight, so they may require lower calorie pet foods. Thus, the variation in the degree of activity for a pet food product can be calculated relative to the indicated degree of activity.

For example, the activity level variation (ALV) for a given pet food product relative to the level of activity can be calculated using the formula:

where PF is a priority factor, FAV is an assigned nutritional value that is determined based on the calorie content of the pet food product, and M is selected based on a given degree of activity. For example, the FAV may be set based on the calorie content (calories/bowl) of the pet food according to the table above, taking into account the variation in physical condition. Similarly, the value of M can be calculated based on a given degree of activity according to the table below:

Degree of activity M value Inactive 0 Moderately active 1 active 2 Very active 3

The priority factor may reflect the relative importance of activity variation in subsequent calculations (eg, when evaluating a pet food product to recommend a pet food product). In certain embodiments in which activity variation is less important than breed size variation, physical condition variation, and developmental stage variation, the priority factor for activity variation may be set to 5 (out of a maximum value of 10).

In such embodiments, for a less active pet (i.e., M = 0), the activity level variation for the highest calorie pet food product would be ALV=5/(+1)=1.25, and the activity level variation for the lowest calorie pet food product would be ALV=5/(+1)=5.

As used herein, the term "pet weight variation" refers to a comparison of the actual weight of the pet and the estimated weight of the pet of the respective breed(s) and the energy value of the pet food. Pets that are overweight for their breeds may require lower calorie pet foods than pets that are ideal for their breed(s). Similarly, underweight pets for their breed(s) may require higher calorie foods than ideal weight pets. Accordingly, it is possible to calculate the variation in the weight of the pet relative to the specified weight of the pet, inherent in its breed.

For example, the deviation of pet weight (PWV) from pet weight can be calculated using the formula:

where PF is a priority factor, FAV is an assigned feed value that is determined based on the calorie content of the pet food product, and M is a value selected based on a given weight of the pet. For example, the FAV may be set based on the calorie content (calories/bowl) of the pet food according to the table above, taking into account the variation in physical condition. Similarly, the value of M can be calculated based on a given degree of activity according to the table below:

Mass status M value Overweight 0 Ideal mass 1 Mass deficiency 2

In certain embodiments, the pet weight variation calculation may further include the pet weight status calculation. For a given pet weight (PW), the weight status (WC) can be calculated using the formula:

where ACW is the average calculated weight of the breed of the pet. For domestic animals of the same breed, the ACW may be equal to the average of the minimum and maximum normal weights for the breed. For pets with more than one breed, the ACW may be equal to the average of the minimum and maximum weights of each breed. If the WC is greater than 1, then the pet is considered to be overweight, if the WC is less than 1, then the pet is considered to be underweight, and if the WC is approximately 1 (for example, 0.95–1.05), then the pet is considered to be of ideal weight. The priority factor may reflect the relative importance of variation in pet weight in subsequent calculations (eg, when evaluating pet food to form a pet food recommendation). In certain embodiments where pet weight variation is less significant than breed size variation, physical condition variation, developmental stage variation, and activity level variation, the pet weight variation priority factor may be assigned a value of 1 (out of a maximum value of 10).

In such embodiments, for an overweight pet (i.e., M=0), the highest calorie pet food product will have a pet weight variance of PWV=1/(+1)=0.33 and the lowest calorie pet food product will have a pet weight variance of PWV=1/(+1)=1.

As used herein, the term "grain preference" refers to a user's or pet's preference for grain content in a pet food product. In certain embodiments, grain preference may include (i) that cereal-containing pet foods are preferred, (ii) that cereal-containing pet foods are undesirable, or that cereal-free pet foods are preferred, (iii) that there is no preference for cereals in pet foods.

As used herein, the term "protein preference" refers to a user's or pet's preference for the protein content of a pet food product. In certain embodiments, the protein preference may include (i) that pet foods containing poultry (e.g., chicken, duck, or turkey) are preferred, (ii) that pet foods containing beef are preferred, (iii) that pet foods containing pork are preferred, (iv) that pet foods containing fish are preferred, (v) that pet foods containing vegetable protein, (vi) that there is no preference for protein content in pet foods.

As used herein, the term "food consistency preference" refers to a user's or pet's preference for the consistency of a pet food product. In certain embodiments, food texture preference may include (i) that wet pet food is preferred, (ii) that dry pet food is preferred, (iii) that there is no preference for pet food consistency.

As used herein, the term "natural ingredient preference" refers to a user's or pet's preference for the texture of a pet food product. In certain embodiments, preference for natural ingredients may include (i) preference for pet foods that contain natural ingredients, (ii) preference for pet foods that do not contain natural ingredients, (iii) no preference for natural ingredients in pet foods.

As used herein, the term "purchase preference" refers to a user's preference for the availability of purchasing a pet food product. In certain embodiments, the purchasing method preference may include (i) that there is a preference for products available online (or both online and in stores), (ii) that there is a preference for products available in stores (or both in stores and online), (iii) that there is no purchase preference for pet food products.

Preferred embodiments relate to methods and apparatus for determining a pet food and/or a pet diet.

In FIG. 1 is an outline of an embodiment of a system 5 for determining a recommended food 21 for pet 1. For example, user 2, who may own pet 1, may wish to purchase food for pet 1. However, user 2 may not know the exact type of food that is best suited to the needs of pet 1 and/or the recommended feeding schedule for pet 1. Accordingly, user 2 may use system 5 to determine a pet food recommendation and/or recommendation 2 1 on feeding a pet 1.

System 5 may include an interface 22. In some embodiments, the interface is a web interface. The web interface 22 may be implemented by a web interface server 23 . For example, web interface server 23 may provide web interface 22 upon receipt of a request from user device 11. User device 11 may be implemented as, for example, a mobile phone, smartphone, smart watch, laptop, tablet, personal digital assistant (PDA), and/or desktop computer. The user device 11 may be configured to acquire an image 3 of a pet. In another embodiment, the user device 11 may additionally or alternatively store an image 3 of the pet 1. The interface 22 of the system 5 may be configured to receive an image 3 of the pet. For example, user 2 may be able to upload a pet image 3 to interface 22. User device 11 may communicate with interface 22 and/or interface server 23 via a network connection, such as a network connection over the Internet. In certain embodiments, the user device 11 may communicate over a wired or wireless network connection (eg, Wi-Fi network connection, Bluetooth, Ethernet, cellular data connections). In further implementations, the functionality of one or both of the interface 22 and interface server 23 may be implemented by the user device 11 (eg, as a program or application running on the user device). In other embodiments, the interface is a voice interface, mobile phone application, chatbot integration, or other user interfaces for web and/or cloud solutions.

Interface 22 may receive additional or alternative inputs. For example, interface 22 may additionally or alternatively obtain information about the pet's breed, pet's weight, pet's activity level, pet's gender, pet's sexual status (e.g., neutered/neutered), pet's environment/climate, current/past feed additive intake, pet's current/past diet, nutritional habits, desired health outcomes, user and/or pet's ingredient preference, and/or purchase method preference. Additional information about the pet may be provided by the user (eg, pet owner or veterinarian). Additional information can also be obtained from a statistical database, or even remotely from devices located on the pet (eg, a collar) or from the pet's environment (eg, from a feeding bowl or litter box).

In the embodiment in which the interface 22 receives the pet image 3, the pet image 3 may be passed to the pet image recognition model 33 . The pet image recognition model 33 may analyze the pet image 3 to determine the pet species/breed, the pet's stage of development, and/or the size of the pet by analyzing the pet image 3 for one or more characteristics of the pet. The pet image recognition model 33 may output a pet image recognition result 32 . The pet image recognition result 32 may comprise a pet breed/size variation, a physical condition variation, an activity degree variation, a pet weight variation, and/or a developmental stage variation. The pet image recognition result 32 can then be fed back to the interface 22.

In certain implementations, as shown in FIG. 1, the pet image recognition model 33 may be implemented on the cloud server 31. For example, the pet image recognition model 33 may be implemented via a cloud server provided by a cloud service provider (eg, Google Cloud Platform Services).

The system 5 may then pass the pet image recognition result 32 to the pet food recommendation model 63 . In certain implementations, such data transfer may be indirect. For example, as shown in FIG. 1, the cloud server 31 and/or the image recognition model 33 may transmit the pet image recognition result 32 to the pet food recommendation model 63 indirectly via the interface 22 and/or the interface server 23. In certain embodiments, the cloud server 31 and/or the image recognition model 33 may transmit the pet image recognition result 32 to the pet food recommendation model 63 directly (for example, over a communication channel between the cloud server 31 and the cloud server 61).

The pet food recommendation model 63 can then use the pet image recognition result 32 and/or other information provided by the user or stored in the pet food recommendation model 63 to determine the pet food recommendation result 62 . The pet food recommendation result 62 can be selected by the pet food recommendation model 63 from a list of pet foods, such as pet foods 62a, 62b, 62c, and 62d, by analyzing the image recognition result 32. For example, as described in more detail below, pet food recommendation model 63 may calculate one or more of breed/pet size variation, physical condition variation, developmental stage variation, activity degree variation, and pet weight variation based on pet characteristics included in image recognition result 32 or provided by user 2.

In one embodiment, pet food 62a may include high fat pet food, pet food 62b may include gluten-free pet food, pet food 62c may include low sugar pet food, and pet food 62a may include meat-free pet food. If the pet food recommendation result 62 is determined by the pet food recommendation model 63, then the pet food recommendation result 62 may then be passed to the interface 22. In other embodiments, the pet food recommendation result 62 may include more than one pet food 62a-d. For example, the pet food recommendation model 63 may select more than one pet food 62a, 62b, 62c, and 62d corresponding to the user's provided preferences, and may include the selected pet foods 62a, 62b, 62c, and 62d in the pet food recommendation result 62 for the user to select. In some embodiments, when a plurality of pet food products are presented within a pet food recommendation, the products may be prioritized based on documented logic in the model.

As shown in FIG. 1, in certain embodiments, the pet food recommendation model 63 may be implemented on a cloud server 61. For example, the pet food recommendation model 63 may be implemented on a cloud server 61 provided by a cloud service provider (e.g., Google Cloud Platform Services).

The interface 22 may use the pet food recommendation result 62 to suggest the recommended pet food 21 to the user 2. In addition, along with the recommended feeding schedule based on the pet image recognition result 32 and the pet food recommendation result 62, a link to purchase the recommended pet food 21 can be provided in the interface 22.

The interface 22 may collect data regarding the pet image recognition result 32 and the pet food recommendation result 62 . For example, user 2 may be able to provide feedback on the recommended pet food 21, such as the effect of the recommended pet food 21 on weight gain and/or loss of pet 1, changes in pet 1's appearance, and/or whether pet 1 liked the recommended pet food 21. In addition, interface 22 may receive data from other sources, including smart collars used to track pet 1's movements, smart food bowls pet 1 feeding scales, pet 1 veterinary card, subsequent image recognition results 32 of pet 1 and/or actual preferences for pet 1 entered by user 2. Such feedback can essentially be considered as “statistical data” for training and optimizing the pet food recommendation model 63.

Statistics for training and optimizing the pet food recommendation model may be stored in the statistics database 101 . To ensure that the recommended pet food 21 is most likely to meet the needs and preferences of the user 2 and/or the needs of the pet 1, the statistics for training and optimizing the pet food recommendation model can be used in the pet food recommendation model 63, as described in more detail below.

In certain embodiments, system 5 may include one or more components implemented on a computer system including a central processing unit (CPU) and a storage device. For example, one or more components of system 5 may include a memory that stores instructions that, when executed by the CPU, cause the CPU to implement one or more of the component's functions. As another example, one or more of the user device 11, the interface server 23, the cloud servers 31 and 61, and the statistics database 101 may be implemented as at least one computer system. In certain embodiments, each of these components may be implemented on a separate computer system. In other embodiments, one or more components may be implemented on a single computer system. For example, the interface server 23 and the user device 11 may be implemented on the same computer system. As another example, in certain embodiments, the cloud servers 31 and 61 may be implemented on the same computer system.

In addition, one or more components of system 5 may communicate over a network, such as a public or private network. For example, the user device 11, the interface server 23, the cloud servers 31 and 61, and the statistics database 101 can communicate over the Internet. The components may communicate over one or more wired or wireless network interfaces, including but not limited to Ethernet, Wi-Fi, Bluetooth, and cellular networks.

In FIG. 1 also outlines an embodiment of a method for providing a recommended pet food 21 to a user 2. The method may be implemented on a computer system such as system 5. For example, the method may be implemented using a user device 11, an interface server 23, cloud servers 31 and 61, and a history database 101. The method may also be implemented using a set of instructions stored on a computer-readable medium that, when executed by a processor, causes the computer system to execute the method. For example, all or part of the method may be implemented using a CPU and a storage device.

Although the following examples are described with reference to the block diagram shown in FIG. 1, many other ways of performing the operations associated with FIG. 1. For example, you can change the order of some steps, combine certain steps with others, repeat one or more steps, and some of the steps described may be optional. In certain embodiments, the steps of the method may preferably be performed in the order shown, with no intermediate steps.

In step 10, the interface 22 may receive a notification from the user 2 (eg, a notification sent via the user device 11) that the user 2 wants to be provided with the recommended pet food 21.

In step 20, the interface 22 may obtain preference information such as user 2 or pet 1 preference for grains, protein preference, food texture preference, natural ingredient preference, and/or purchase method preference. The interface 22 may use this information to personalize the pet food recommendation result 62 for the pet and pet owner and/or other pet food preferences and/or pet food preferences. The interface 22 may also be configured to obtain an image 3 of the pet 1. As further described below, the interface 22 may also receive data about the pet 1, such as data associated with one or more of the sex, breed, or age of the pet 1.

At step 30, the interface 22 may transmit the pet image 3 to the pet image recognition model 33 . The pet image recognition model 33 may analyze the pet image 3 to determine the pet image recognition result 32 based on the pet image 3 . For example, the pet image recognition model 33 may analyze the pet image 3 to detect one or more characteristics of the pet image 3. For example, the image recognition model 33 may be implemented as a machine learning model such as a 2D machine learning model (eg, a recurrent neural network or a convolutional neural network). The pet image recognition model 33 may be trained on training pet images that may be prepared or otherwise obtained from previously provided pet images 3, as will be described below. In the learning process, the pet image recognition model 33 may be configured to detect image characteristics or otherwise determine image characteristics to detect a pet in the image 3 based on the learning process. Examples of image characteristics may include the outline of pet 1, the area bounded by outline of pet 1, the percentage of pixels in pet image 3 occupied by pet 1, and one or more geometric dimensions of pet 1 in pet image 3 (e.g., length, height, distance from paws to withers, distance from back to abdomen, chest width, chest volume).

Thereafter, the characteristics of the image can be analyzed to generate an image recognition result 32 . Image recognition result 32 may include one or more characteristics of pet 1, such as height, weight, length, breed, breed size, size of pet, activity level, developmental stage, and/or physical condition of pet 1. Image characteristics suggest or otherwise indicate the presence of one or more of pet 1 characteristics. For example, a larger border may be indicative of either or both of the increased weight of pet 1 or a larger size of pet 1 breed. In another embodiment, to determine the breed of the pet, the pet image recognition model may take into account other features of the image, such as coat characteristics and/or facial features.

The pet image recognition model 33 may be configured to generate or predict certain characteristics of the pet depending on the characteristics of the pet to include in the pet image recognition result 32 that are relevant to subsequent processing of the pet food recommendation model 63. For example, the pet food recommendation model 63 may calculate breed size variation, and therefore, may rely on the breed size estimate of pet 1 when generating the pet food recommendation result. Thus, in such an implementation, the pet image recognition model 33 can be configured to estimate the breed size of pet 1 based on the pet image 3 and include the breed size estimate in the image recognition result 32.

In certain embodiments, the pet image recognition model 33 may be implemented as more than one machine learning model. For example, the pet image recognition model 33 may include a first machine learning model configured to detect characteristics of the pet image 3 and a second machine learning model configured to analyze the image characteristics and detect the characteristics of the pet 1. In other further embodiments, the pet image recognition model 33 may be implemented in a single machine learning model. For example, the pet image recognition model 33 may include one machine learning model capable of both detecting image characteristics in the pet image 3 and estimating the characteristics of the pet 1 based on the image characteristics. In yet another example, the pet image recognition model 33 may not estimate image characteristics and may instead include one machine learning model capable of estimating the characteristics of pet 1 based on the pet image 3 directly.

In step 40, the image recognition result 32 of the pet may be passed to the interface 22. In one embodiment, the image recognition result 32 may be passed to the pet food recommendation model 63 without first being passed to the interface 22.

At step 50, the pet food recommendation model 63 may receive a pet image recognition result 32 . The pet image recognition result 32 may be obtained from the interface 22, the pet image recognition model 33, and/or from another location.

At step 60, the pet food recommendation model 63 may analyze the pet image recognition result 32 to determine the pet food recommendation result 62 . The pet food recommendation model 63 can analyze the characteristics of the pet included in the pet image recognition result 32 . The pet food recommendation model 63 may analyze the characteristics of the pet included in the pet image recognition result and the input passed to the interface 22 to determine the pet food recommendation result. For example, image recognition result 32 may include one or more of breed, breed size, pet size, physical condition, developmental stage, activity level, and weight of pet 1, and input data may include one or more of pet sex, pet sex status, pet weight, pet sensitivity response, pet health status, pet environment, current/past feed additive intake, current/previous diets, dietary habits, and/or desired health outcomes that recommendation model 63 can use. pet food product to calculate one or more variations such as breed/pet size variation, physical condition variation, developmental stage variation, activity level variation, and/or pet weight variation. The pet food recommendation model 63 can then evaluate the pet food based on the sum of one or more variations.

In certain embodiments, the pet food recommendation model 63 can be configured to (i) filter the pet food list to create a filtered pet food list based on one or more user preferences, and (ii) sort the filtered pet food list. For example, user 2 may provide one or more preferences for filtering pet food products presented as a result of recommending a pet food product. In some embodiments, user 2 may provide a grain preference, a protein preference (e.g., chicken, beef, pork), a food consistency preference (e.g., wet or dry), a natural ingredient preference, a purchase method preference, and/or other additional variables that may be useful in generating a filtered list of pet foods. The filtered list of pet food products may then be sorted using one or more criteria selected by the user, such as price, availability, and/or brand.

In additional embodiments or in addition to the above embodiments, pet food recommendation model 63 may include a machine learning model generating a pet food recommendation result 62 . For example, a machine learning model can be trained to prioritize one or more variations.

The result of 62 pet food recommendations may be a recommended pet food 21 selected from pet foods 62a-d. In other embodiments, as described above, pet food recommendation result 62 may include a plurality of pet foods 62a-d of user 2's choice.

In certain embodiments, the pet food recommendation result 62 may also include a pet food recommendation. For example, after pet food 62a-d is determined, pet food recommendation model 63 may determine a pet food recommendation based on the calorie content of pet food 62a-d. For example, the pet food recommendation model 63 may determine a feeding recommendation based on a combination of factors, including the desired calorie intake and/or calorie content of the product selected for pet 1, based on pet 1 characteristics (e.g., pet 1 breed, breed size, physical condition, developmental stage, activity level, and/or weight of pet 1). The pet food recommendation model 63 can then determine how much of each pet food 62a-d is required to achieve the desired caloric intake for pet 1 based on the calorie content of the food (e.g., by dividing the desired caloric intake by the calorie content of pet foods 62a-d). Thereafter, the pet food recommendation model 63 may divide the food into a certain number of meals of the desired or recommended pet food 1 per day (eg, 2 meals per day). In certain embodiments, the pet food recommendation model may not generate a pet food recommendation until the user selects pet food 62a-d from pet foods 62a-d included in the pet image recognition result. For example, a recommendation for pet food selection and pet feeding may aim to achieve, over a specified period of time (e.g., 6 weeks or 10 weeks), an ideal physical condition score as measured by weight maintenance, weight loss, or weight gain.

At step 70, the pet food recommendation result 62 can be obtained by the interface 22. In one embodiment, the interface is a web interface. In other embodiments, the interface is a chat, voice, or mobile application.

At step 80, a pet food recommendation result 62 may be transmitted from the interface 22 to the user device 11 and/or user 2. The pet food recommendation result 62 may be in the form of a recommended pet food 21. In one embodiment, the recommended pet food 21 is two or more different pet foods. In one embodiment, the recommended pet food 21 is represented by two or more pet food brands. In one embodiment, the recommended pet food 21 is a single pet food and/or is represented by a single brand of pet food.

At step 90, the statistics database 101 may receive behavioral data for training and optimizing the pet food recommendation model 63 and/or the image recognition model 33 . For example, the image 3 of a pet can be stored in the statistical database and manually marked or analyzed and transferred to the image recognition model 33 for training. In certain embodiments, the subject matter expert may view the pet image 3 to determine one or more characteristics of the pet 1 or the characteristics of the pet image 3 and may label the pet image 3 with the identified pet characteristics. Thereafter, the image recognition model 33 can be trained on the pet image 3 in combination with a plurality of other pet images 3. For example, the image recognition model 33 may adjust its parameters (eg, one or more scores or coefficients in the image recognition model) to maximize the accuracy of the detected image features or pet features.

In another example, the food recommendation model 63 may analyze user 2 or pet 1 behavioral data (e.g., that user 2 purchased a recommended pet food 21, pet 1's subsequent weight loss or gain, pet 1's subsequent activity level) and may update one or more parameters of the model (e.g., priority factors for variations) to provide the most accurate selection of recommended pet food 21. Such data may come from a user's interaction with the solution through web activity monitoring data, cookies such as user surveys, purchase records data and subsequently submitted images 3 of a pet. The food recommendation model 63 may also analyze behavioral data about pet 1 (eg, whether pet 1 has achieved desired changes in physical condition, whether pet 1 likes the recommended pet food 21). Such data may come from connected pet devices (e.g., collars, bowls, scoops, litter trays, drinkers, cameras, scales, bins, mats, beds) that automatically collect and transmit data used to provide personalized pet food recommendations and/or pet feeding recommendations. In certain embodiments, the food recommendation model 63 can be optimized to improve one or more of whether the user 2 purchased the recommended pet food 21, whether the desired change in weight of the pet 1 occurred after consuming the recommended pet food 21, whether the pet 1 liked the recommended pet food 21, and whether the user continues to use the system 5 to generate pet food recommendation results 62.

At step 100, behavioral data for training and optimizing the pet food recommendation model 63 may be fed into the pet food recommendation model 63 to ensure that the recommended pet food 21 is most likely to meet the needs and preferences of the user 2 and/or the needs or preferences of the pet 1.

In FIG. 2 is an outline of an implementation in the form of an application diagram 200 . Application diagram 200 may show how user 2 interacts with user device 11 shown in FIG. 1 to provide input to the interface 22 shown in FIG. 1.

As shown in FIG. 2, the introduction 210 may be configured to receive additional or initial information from the user 2. For example, the introduction from user 1 may receive login information, initial account settings, required legal and/or disclosure documents, and/or other general information. The introduction 210 may only be used on initial login to the interface 22 and may not be displayed on subsequent logins after the initial user profile has been created.

The interface 22 may provide a menu 220. Upon entering the interface 22, the user 2 is presented with a menu 220, which may have a "home", "base", or initial interface. Menu 220 may include navigation elements that allow the user 2 to navigate to other elements of the interface 22.

The interface 22 may be provided with one or more menu items 230. Menu items 230 may be accessed via menu 220. Menu items 230 may allow user 2 to change, set, and/or adjust settings in interface 22 and/or access other information related to interface 22. For example, menu items 230 may contain screens and/or documents related to FAQs, contacts and/or contact information, feedback pages (e.g., user editable feedback or feedback already left by other users). and for viewing by the user 2), privacy policy, terms and/or explanation related to promotional materials in the interface 22.

Interface 22 may provide one or more profile inputs 240 . Profile input 240 may be accessed via menu 220. Profile input 240 may allow user 2 to enter information related to the pet and/or user 2's preferences, which can be used to determine a recommended pet food 21. For example, the profile input 240 may include: a start screen, which may provide instructions on how the user should perform 2 further actions; basic information screen; a photo upload screen where the user 2 can upload a pet image 3 for later use in the image recognition model 33 shown in FIG. 1; a breed information screen where the user 2 can enter breed information of the pet 1; a weight information screen where the user 2 can input information about the weight of the pet 1; an activity degree screen where the user 2 can input information about the activity degree of the pet 1; an ingredient preference screen where user 2 can enter ingredient preference information for pet 1 and/or user 2; a protein preference screen where user 2 can enter protein preference information for pet 1 and/or user 2; and/or a purchase method preference screen where user 2 can enter information about the purchase method preference for user 2.

Upon completion of the entry of at least one of the elements in the profile input 240, a complete profile 250 can be generated using the information provided in the profile input 240 and/or any information generated by the pet image recognition model 33, the pet food recommendation model 63, and/or the statistics database 101, as shown in FIG. 1. Full profile 250 can be used to determine the recommended 21 pet food. For example, the full profile 250 may include one or more features included in the image recognition result 32. Thereafter, the pet food recommendation model 63 may parse one or both of the image recognition result 32 and the full profile 250 to generate a pet food recommendation result 62 as described above.

In one embodiment, if a photo upload is available when entering the profile input 240, the profile input 240 may further include a file download window in which the user can upload the pet image 3 and/or read photo parameter messages if it does not meet certain predefined criteria. In certain embodiments, the pet image 3 may require one or more photographic characteristics. For example, in certain embodiments, image recognition model 33 may be configured to provide the most accurate results based on pet images 3 taken while pet 1 is facing the camera (e.g., image recognition model 33 may be trained primarily on pet images 3 in which pet 1 is facing the camera). In other examples, image recognition model 33 may be configured to provide the most accurate results based on pet images 3 when pet 1 is shown in profile (for example, image recognition model 33 may be trained primarily on pet images 3 in which pet 1 is profiled towards the camera). The uploaded photo may indicate characteristics of the photo necessary or preferred by the image recognition model 33 in order to obtain an accurate image recognition result 32 . Additionally or alternatively, when the pet image 3 is processed by the image recognition model 33 to determine the pet image recognition result 32, the user may see a visual displaying the photo analysis.

The interface 22 may include a visual element 260 that displays the process of thinking. The user can see a visual element 260 depicting the process of thinking as the pet image recognition model 33 processes the pet image 3 and/or any other input data to determine a pet image recognition result 32, a pet food recommendation result 62, and/or a recommended pet food 21.

The results 270 may be displayed on the interface 22. The results 270 may be provided to the user 2 after the pet image recognition model 33 processes the pet image 3 and/or any other input to determine a pet image recognition result 32 and/or a pet food recommendation result 62. The results 270 may include a recommended pet food 21. Results 270 may include a link to access a feeding guide 280. As described above, in certain embodiments, pet food recommendation result 62 may include a pet food group 62a-d from which a user can select a pet food recommendation 21 from pet food 62a-d. In such embodiments, the results may include a selection of foods from which the user selects a recommended pet food.

The feeding guide 280 may include a feeding guide depending on the recommended pet food 21, the pet food selected by the user, the pet image 3, and/or any other input to determine the pet image recognition result 32 and/or the pet food recommendation result 62.

Tips 290 may be provided in interface 22. Tips 290 may be embedded in any of the other visual elements described herein and may provide user 2 with recommendations or explanations as user 2 navigates through the options included in interface 22.

In some embodiments, a user may create a profile to store information about their pet, preferences, and purchase history. This profile can be used to run algorithm based programs that initiate subsequent data transfer and user interaction. For example, the communication may be in the form of automatic email or text messages with instructions for the user to change the pet's food or feed recommendations when the pet reaches a certain stage of development (eg, adulthood). Machine learning capabilities can be applied to device-derived data, such as data from a web-enabled device located on a pet, or user-generated inputs in response to automated messages to continually improve the quality of the pet food and personalize the content of pet feeding recommendations for a particular pet and/or pet owner.

In FIG. 3 is an outline of an embodiment of a personalized feeding guide 330 for providing personalized pet food recommendations and/or pet feeding recommendations. As shown in FIG. 3, the personalized feeding guide 330 is capable of being used in a web and/or mobile application, but the personalized feeding guide 330 may be provided in other media. The personalized feeding guide 330 may include: a selected pet food 301; pet profile 302; pet feeding instructions 303; instructions 304 for switching to a different pet food; brief information 305 about the product; information 306 on the composition and nutritional value of the product; and other recommendations 307 on pet foods. The selected pet food data may include the pet's name, brand name, product name, and a link to a website to purchase the pet food. Selected pet food data 301 may also reflect the unit size and/or design of the pet food selected by user 2. Pet profile 302 may reflect input data and generated information regarding pet 1. For example, such information may include pet breed, pet weight, pet activity level, user and/or pet ingredient preference, pet protein preference, and/or purchase method preference. The pet feeding instructions 303 may include instructions for feeding the pet 2. The pet feeding instructions 303 may take into account the pet image 3 and/or any other input to determine the pet image recognition result 32 and/or the pet food recommendation result 62, as well as the pet image recognition result 32 and/or the pet food recommendation result 62.

The pet food transition instructions 305 may include instructions for safely and successfully transitioning the pet 1 from the first food to the second food. Other pet food recommendations 307 may include other pet foods 62 a, 62 b, 62 c, and 62 d defined by the pet food recommendation model 63 but not selected by the user 2.

It should be understood that various changes and modifications to the examples described herein will be apparent to those skilled in the art. Such changes and modifications may be made without departing from the spirit and scope of the present subject matter and without diminishing its intended benefits. Therefore, the appended claims are intended to cover such changes and modifications. Thus, the present embodiments should not be limited to the exact details of the methodology or implementation set forth above, as such variations and modifications are intended to be included within the scope of the present disclosure. Moreover, unless specifically stated otherwise, any use of the terms "first", "second", etc. does not indicate any order or precedence, and the terms "first", "second", etc. are used solely to distinguish one element from another.

Claims (66)

1. A method of presenting a recommendation for a pet food product, comprising: receiving by the web interface server an image of a pet, which depicts a pet; analysis by the server of web interfaces, using the pet image recognition model, of the pet image to determine one or more characteristics of the pet, the analysis includes: training the pet image recognition model to detect image characteristics based on the percentage of pixels in the pet image that are in the pet; And identifying one or more characteristics of the pet using the trained pet image recognition model; the web interface server receiving one or more pet-related input types; storing behavioral data of the pet in a statistical database; learning by the web interface server the pet food recommendation model based on the behavioral data stored in the statistics database; And the server parsing the web interfaces using the pet food recommendation model, the pet characteristics, and the pet-related raw data to generate a pet food recommendation and/or a pet feeding recommendation. 2. The method of claim. 1, in which the image of the pet is received from the user device. 3. The method of claim 1, wherein the pet image analysis further comprises: parsing by the server of the web interfaces of the pet image to determine one or more characteristics of the pet image; And determining, by the web interface server, one or more characteristics of the pet based on the characteristics of the image. 4. The method of claim 3, wherein the characteristics of the pet include one or more characteristics selected from the group consisting of breed, breed size, pet size, physical condition, developmental stage, activity level, sex of the pet, sex status of the pet, and weight of the pet. 5. The method of claim. 1, further comprising: filtering by the web interface server the pet food list based on at least one user preference to generate the filtered pet food list. 6. The method of claim 5, wherein the at least one user preference includes at least one preference selected from the group consisting of grain preference, protein preference, food consistency preference, natural ingredient preference, and purchase method preference. 7. The method according to p. 1, further comprising: calculation by the web interface server of one or more variations selected from the group consisting of a breed size variation, a pet size variation, a physical condition variation, a developmental stage variation, an activity degree variation, and a pet weight variation. 8. The method of claim 1, further comprising: calculation by the web interface server of a plurality of variations to generate a plurality of calculated variations; And evaluation by the server of the web interfaces of the pet food list based on the sum of the calculated variations to identify one or more recommended pet food products. 9. The method of claim 1, further comprising: providing, by the web interface server, a pet food recommendation to a user, the pet food recommendation comprising a plurality of pet food products; receiving, by the web interface server, from the user information about a selected pet food product from a plurality of pet food products, and generation by the web interface server of recommendations for feeding the pet based on the selected pet food product. 10. The method of claim 9, wherein the pet food recommendation is based on the calorie content of the selected pet food product. 11. The method of claim 1, further comprising: training a pet image recognition model based on behavioral data stored in a statistical database. 12. A pet food recommendation system comprising: CPU; And memory that stores instructions that, when executed by the processor, cause the processor to: obtains an image of a pet, which shows a pet; analyzes, using a pet image recognition model, an image of a pet to determine one or more characteristics of the pet, the analysis including: training the pet image recognition model to detect image characteristics based on the percentage of pixels in the pet image that are in the pet; And identifying one or more characteristics of the pet using the trained pet image recognition model; accepts one or more pet-related input types; stores the behavioral data of the pet in the statistical database; performs training of the pet food recommendation model based on the behavioral data stored in the statistics database; And analyzes, using the pet food recommendation model, the characteristics of the pet and the raw data associated with the pet to form a pet food recommendation and/or a pet feeding recommendation. 13. The system of claim 12, wherein the storage device stores additional instructions that, when executed by the processor, cause the processor to: analyzes the pet image to identify one or more characteristics of the pet image; And determines one or more characteristics of the pet based on the characteristics of the image. 14. The system of claim 12, wherein the storage device stores additional instructions that, when executed by the processor, cause the processor to: filters the pet food list based on at least one user preference to create a filtered pet food list. 15. The system of claim 14, wherein the pet characteristics include one or more characteristics selected from the group consisting of breed, breed size, pet size, physical condition, developmental stage, activity level, pet sex, pet sex status, and pet weight. 16. The system of claim 12, wherein the storage device stores additional instructions that, when executed by the processor, cause the processor to: calculates one or more variations selected from the group consisting of breed size variation, pet size variation, physical condition variation, developmental stage variation, activity degree variation, and pet weight variation. 17. The system of claim 12, wherein the storage device stores additional instructions that, when executed by the processor, cause the processor to: performs a calculation of the plurality of variations to generate a plurality of calculated variations; And evaluates a list of pet foods based on the sum of calculated variations to identify one or more recommended pet foods. 18. The system of claim 12, wherein the storage device stores additional instructions that, when executed by the processor, cause the processor to: provides the user with a pet food recommendation, wherein the pet food recommendation contains a plurality of pet food products; receives information from a user about a selected pet food product from a plurality of pet food products, and generates a pet food recommendation based on the selected pet food product. 19. The system of claim 12, wherein the storage device stores additional instructions that, when executed by the processor, cause the processor to: performs training of the pet image recognition model based on the behavioral data stored in the statistics database. 20. A computer-readable physical storage medium that stores pet food recommendation submission instructions that, when executed by the processor, cause the processor to: obtains an image of a pet, which shows a pet; analyzes, using an image recognition model, an image of a pet to determine one or more characteristics of the pet, the analysis including: training the pet image recognition model to detect image characteristics based on the percentage of pixels in the pet image that are in the pet; And identifying one or more characteristics of the pet using the trained pet image recognition model; accepts one or more pet-related input types; stores the behavioral data of the pet in the statistical database; performs training of the pet food recommendation model based on the behavioral data stored in the statistics database; And analyzes, using the pet food recommendation model, the characteristics of the pet and the raw data associated with the pet to form a pet food recommendation.

Images