JP2015028767A - Recipe recommendation device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a recipe recommendation device capable of recommending optimum recipe on the spot for acquired ingredients.SOLUTION: A recognition character string is obtained by performing a character recognition process (S20) with an imaged picture. Ingredient synonym memory means is referenced by the recognition character string, to specify an ingredient candidate (S30). NG word memory means is referenced by the recognition character string, to specify a recognition character string containing a NG word that includes the NG word (S40). An ingredient corresponding to the recognition character string, among recognition character strings, which corresponds to an ingredient candidate but does not contain the NG word is specified as the ingredient to be used (S50). A stock ingredient that has been obtained before is acquired (S55). Recipe memory means is searched by the specified ingredient and stock ingredient, to obtain a recipe (S60).

Description

  The present invention relates to a technique for recommending a recipe based on information read by an OCR (Optical Character Reader).

  Conventionally, when cooking, it has been practiced to purchase necessary ingredients according to the contents of the dish and to actually cook using the purchased ingredients. However, there is a problem that it is difficult to decide what kind of cooking to use when there are leftovers of purchased items. In order to solve this, a technique for presenting a menu to the user using information on the input food has been developed (see Patent Document 1).

JP 2002-251518 A

  However, the technique described in Patent Document 1 has a problem that it is difficult to know an optimum recipe on the spot for foods obtained by purchase or the like.

  Then, this invention makes it a subject to provide the recipe recommendation apparatus which can recommend an optimal recipe on the spot about the acquired foodstuff.

  In order to solve the above-mentioned problem, a recipe recommendation device according to the present invention includes an ingredient synonym storage means that stores information on ingredients and synonyms of ingredients in association with each other, and an NG word storage means that stores NG words to be excluded as ingredients. Recipe storage means for storing recipe and food information in association with each other, imaging means for imaging an object including a character string, character recognition processing is performed on a captured image obtained by the imaging means, and a recognized character string Character recognition means for obtaining the food synonym storage means by the recognized character string, food candidate candidate specifying means for specifying the recognized character string to be a food candidate, and NG word storage means by the recognition character string NG word processing means for identifying a recognized character string including an NG word, and corresponding to a recognized character string corresponding to a food candidate and not including an NG word among the recognized character strings Ingredient specifying means for specifying the ingredients as ingredients to be used, recipe storage means for searching the recipe storage means by the specified ingredients and acquiring the recipe, and recipe display means for displaying the acquired recipe It is characterized by having.

  According to the recipe recommendation device according to the present invention, a character recognition process is performed on a captured image to obtain a recognized character string, a food material synonym storage unit is referred to by the recognized character string, a food material candidate is specified, By referring to the NG word storage means by column, a recognized character string including the NG word having the NG word is specified, and among the recognized character strings, corresponding to the recognized character string corresponding to the food candidate and not including the NG word Since the ingredients are specified as the ingredients to be used, the recipe storage means is searched by the identified ingredients, and the recipe is acquired, so it becomes possible to recommend the optimum recipe on the spot for the acquired ingredients. .

  In addition, the recipe recommendation device according to the present invention further includes a food season information storage unit that stores information related to the season of the food, and the recipe acquisition unit stores the food season information according to the food specified by the food specifying unit. It is characterized in that, referring to the means, the seasonal level of each ingredient is acquired, and the recipe is acquired with priority given to the high seasonal ingredient.

  According to the recipe recommendation device according to the present invention, information on the season of ingredients is stored, the degree of the season of the specified ingredient is obtained, and the recipe is obtained with priority on ingredients having a high degree of season. Therefore, it is possible to propose an optimal dish according to the time.

  In addition, the recipe recommendation device according to the present invention updates information on the season of ingredients stored in the ingredient season information storage means using information on the season of ingredients determined based on the purchase information of ingredients in a specific area. It further has a food season information update means.

  According to the recipe recommendation device according to the present invention, the information on the season of the stored ingredients is updated using the information on the season of the ingredients determined based on the purchase information of the ingredients in the specific area. In particular areas, it is possible to reflect information on foods that are especially seasonal.

  Further, in the recipe recommendation device according to the present invention, the food specifying unit performs processing for storing information on the specified food in the stock food storage unit as stock food, and the recipe acquisition unit is specified by the food specifying unit. The recipe storage means is searched using the stock ingredients stored in the stock food storage means in addition to the food ingredients.

  According to the recipe recommendation device according to the present invention, the storage means stores the information on the specified foodstuff as stock food, and the stock food stored in the storage means in addition to the specified food at the time of the recipe acquisition. Since the recipe is searched using the above-mentioned recipe, it is possible to acquire the recipe using the previously acquired ingredients on the spot by using the previously specified ingredients as the inventory ingredients.

  Moreover, the recipe recommendation device according to the present invention includes an existing recipe storage unit that stores an existing recipe, and a recipe that is duplicated with an existing recipe stored in the existing recipe storage unit among recipes acquired by the recipe acquisition unit. And an already-excluded recipe excluding unit to be excluded, wherein the recipe display unit displays a recipe not excluded by the already-existing recipe excluding unit.

  According to the recipe recommendation device according to the present invention, the existing recipes are stored, and among the acquired recipes, those that overlap with the stored existing recipes are excluded, thereby preventing nutritional bias and rutification. It becomes possible to do.

  In addition, the recipe recommendation device according to the present invention further includes a search log transmission unit that transmits a search log for the recognized character string obtained in the process of acquiring the recipe to a management server. .

  According to the recipe recommendation device according to the present invention, since the search log for the recognized character string obtained in the process of obtaining the recommended recipe is transmitted to the management server, the search log is accumulated in the management server, By analyzing, it becomes possible to improve the accuracy of recommendation of dishes.

  In the recipe recommendation device according to the present invention, the search log transmission unit transmits the search log as error information to the management server when an instruction is given to the recipe displayed by the recipe display unit. It is characterized by.

  According to the recipe recommendation device according to the present invention, when an instruction is given to a recipe displayed as a recommended recipe, the search log is transmitted to the management server as error information. The error information corresponding to the instruction from the person can be grasped on the management server side, and the accuracy of dish recommendation can be improved.

  ADVANTAGE OF THE INVENTION According to this invention, it has the effect that it becomes possible to recommend an optimal recipe on the spot about the acquired foodstuff.

It is a figure which shows the utilization environment of the recipe recommendation apparatus which concerns on one Embodiment of this invention. It is a figure which shows the hardware constitutions of the recipe recommendation apparatus. 3 is a functional block diagram of the recipe recommendation device 10. FIG. It is a figure which shows an example of the information memorize | stored in foodstuff synonym memory | storage means D1. It is a figure which shows an example of the information memorize | stored in NG word memory | storage means D2. It is a figure which shows an example of the information memorize | stored in the consumption period memory | storage means D6. It is a figure which shows an example of the information memorize | stored in stock food storage means D7. It is a flowchart which shows the processing operation of the recipe recommendation apparatus which concerns on one Embodiment of this invention. It is a figure which shows an example of the search log obtained in the process of a recipe recommendation apparatus. It is a figure which shows the foodstuff of a season point high rank among the specified foodstuff. It is a figure shown about the process of the priority order determination of a recipe. 3 is a diagram illustrating an example of a search log stored in a management server 30. FIG. It is a figure which shows the display screen of the recipe recommendation apparatus. It is a figure which shows an example of the information memorize | stored in foodstuff purchase information storage means D8.

<1. Device configuration>
DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments of the invention will be described in detail with reference to the drawings. FIG. 1 is a diagram showing a usage environment of a recipe recommendation device according to an embodiment of the present invention. In FIG. 1, 10 is a recipe recommendation device, 30 is a management server, and 40 is a network. The recipe recommendation device 10 and the management server 30 are connected to a public network 40 such as the Internet, and can exchange data with each other.

  The recipe recommendation device 10 is a portable terminal device having an information processing function such as a smartphone and a tablet terminal, a network communication function, and a display function. The recipe recommendation device 10 inputs information on ingredients, and recommends based on the inputted information on ingredients. It functions as a device that determines and displays a recipe to be displayed. The management server 30 is a server computer that manages food information, data such as recipes, records information from the recipe recommendation device 10, and supplies necessary information to the recipe recommendation device 10.

  FIG. 2 is a hardware configuration diagram of the recipe recommendation device 10. As described above, the recipe recommendation device 10 can be realized by a smartphone, a tablet terminal, or the like. In the present embodiment, the recipe recommendation device 10 is realized by incorporating a dedicated program into the tablet terminal. A tablet terminal is a thin plate-like computer that can operate a screen with a touch panel. As shown in FIG. 2, a CPU (Central Processing Unit) 10a, a main memory RAM (Random Access Memory) 10b, a hard disk for storing programs and data executed by the CPU 10a, a flash memory, etc. A storage device 10c, an instruction input unit 10d such as a keyboard and a touch panel, a data input / output I / F (interface) 10e for data communication with an external device such as a data storage medium, and a display device such as a liquid crystal display A unit 10f, a communication unit 10g for performing network communication with other computers via a network, and a camera h are connected to each other via a bus. Actually, the tablet terminal has a microphone, a speaker, and the like in addition to those shown in FIG. 2, but is omitted because it is irrelevant to the present embodiment.

  FIG. 3 is a functional block diagram of the recipe recommendation device 10. In the hardware configuration shown in FIG. 2, the CPU 10 a reads the program stored in the storage device 10 c into the RAM 10 b and executes it, so that the recipe recommendation device 10 can function each unit shown in FIG. 3. Become. As shown in FIG. 3, the recipe recommendation device 10 includes an imaging unit 11, a character recognition unit 12, an ingredient candidate identification unit 13, an NG word processing unit 14, an ingredient identification unit 15, a recipe acquisition unit 16, an existing recipe exclusion unit 17, Search log transmission means 18, recipe display means 19, inventory ingredient acquisition means 20, ingredient season information update means 21, ingredient synonym storage means D1, NG word storage means D2, recipe storage means D3, ingredient season information storage means D4, existing recipe It has storage means D5, consumption period storage means D6, inventory food storage means D7, and food purchase information storage means D8.

  The imaging unit 11 is a unit that captures an image according to an instruction and obtains a captured image, and is realized when the CPU 10a activates the camera 10h. The character recognition unit 12 is a unit that performs character recognition processing on the captured image to obtain a recognized character string, and is realized by the CPU 10a executing a known OCR processing program. The food material candidate specifying means 13 is a means for specifying a recognized character string serving as a food material candidate by referring to the food material synonym storage means D1 using the recognized character string, and is realized by executing a dedicated program.

  The NG word processing means 14 is means for specifying the recognized character string containing the NG word by referring to the NG word storage means D2 by the recognized character string, and is realized by the CPU 10a executing a dedicated program. . The food material specifying means 15 is a means for specifying, as a food material to be used, a food material corresponding to a recognized character string that corresponds to a food material candidate and does not contain an NG word among the recognized character strings. It is realized by executing. The recipe acquisition unit 16 is a unit that searches the recipe storage unit D3 with the specified food and acquires a recipe, and is realized by the CPU 10a executing a dedicated program.

  The existing recipe exclusion means 17 is a means for excluding those recipes acquired by the recipe acquisition means 16 that overlap with the existing recipe stored in the existing recipe storage means D5, and the CPU 10a executes a dedicated program. Is realized. The search log transmission means 18 is means for transmitting information relating to processing of the recognized character string obtained in the process of obtaining a recommended recipe as a search log to the management server 30. The CPU 10a controls the RAM 10b and the communication unit 10g. It is realized by doing. The recipe display means 19 is a means for displaying recipes excluded by the existing recipe exclusion means 17, and is realized by the CPU 10a controlling the RAM 10b and the display unit 10f.

  The stock food acquisition means 20 is means for acquiring information on stock food stored in the stock food storage means D7, and is realized by the CPU 10a executing a dedicated program. The ingredient seasonal information update means 21 is means for updating the ingredient seasonal information stored in the ingredient seasonal information storage means D4 using the ingredient purchase information stored in the ingredient purchase information storage means D8, and the CPU 10a is a dedicated program. It is realized by executing.

  The foodstuff synonym storage means D1 is a storage means that stores the synonyms of foodstuffs. An example of information stored in the food synonym storage means D1 is shown in FIG. As shown in FIG. 4, the food material synonym storage means D1 stores a food material name and a synonym in association with a food material ID that is identification information for specifying the food material. There is a one-to-one correspondence between the ingredient ID and the ingredient name, and a plurality of synonyms are recorded in the ingredient name. The food synonym storage means D1 is realized by securing a predetermined area for storing the synonyms of food in the storage device 10c.

  The NG word storage unit D2 is a storage unit that stores NG words. An NG word is a word that should not be included in an ingredient name or a synonym. An example of the NG word stored in the NG word storage means D2 is shown in FIG. As shown in FIG. 5, in the NG word storage means D2, words such as “flavor”, “soup”, and “jelly” are often used together with the names of ingredients but are not suitable as ingredients used for cooking. ing. The character string including the word stored in the NG word storage unit D2 is excluded from the food candidate. The NG word storage means D2 is realized by securing a predetermined area for storing NG words in the storage device 10c.

  The recipe storage unit D3 is a storage unit that stores a recipe. The recipe storage means D3 includes a recipe ID that is identification information for specifying a recipe, a recipe name that is the name of the recipe, an ingredient ID of the ingredient used for the recipe, and a recipe that is presented to the user by displaying the recipe ID. Information is stored in association with each other. The recipe storage means D3 is realized by securing a predetermined area for storing the recipe in the storage device 10c.

  The food season information storage means D4 is a storage means for storing the food season information. The food season information storage means D4 stores information about the season of food. The information regarding the season of the food is information indicating the level of the season of the food. In this embodiment, the level of the season of the food is recorded by points. Specifically, seasonal points are recorded in association with the food ID in units of a predetermined period of one year. The predetermined period unit can be set as appropriate, but in the present embodiment, the unit is set to one week. Therefore, the points of each week of 52 weeks a year are stored for each food material. The seasonal point is a numerical value indicating that the larger the value, the higher the seasonal level. The food season information storage means D4 is realized by securing a predetermined area for storing the food season information in the storage device 10c.

  The existing recipe storage means D5 is a storage means for storing the existing recipe. In the existing recipe storage means D5, the recipe previously recommended by the recipe recommendation device 10 is stored as an existing recipe. Specifically, the recipe ID of the already-stored recipe is stored in association with the recommended date and time. Recently, this is to prevent the recommendation from being recommended again. Therefore, when the predetermined period has elapsed, it is deleted from the already-stored recipe storage means D5 and only those within the predetermined period are stored. The predetermined period can be set as appropriate, but is set to one week in the present embodiment. Therefore, the recipe selected more than a week ago may be displayed again. The existing recipe storage means D5 is realized by securing a predetermined area for storing the existing recipe in the storage device 10c.

  The consumption period memory | storage means D6 is a memory | storage means which memorize | stored the consumption period of each foodstuff. The consumption period is a period during which food can be consumed. An example of the consumption period stored in the consumption period storage means D6 is shown in FIG. As shown in FIG. 6, in the consumption period storage means D6, for each food, the consumption period is recorded on a daily basis in association with the food ID and the food name. The consumption period storage means D6 is realized by securing a predetermined area for storing the consumption period in the storage device 10c. In this embodiment, the consumption period is managed in units of days, but the consumption period is recorded in units of hours, minutes, seconds, etc., and is managed in units of hours, minutes, seconds. Is also possible.

  The stock food storage means D7 is a storage means for storing information on stock food. An inventory food material is a food material that has been acquired within a predetermined period in the past and is considered to be in stock. Specifically, the food specified by the food specifying means 15 is stored. Although it can set suitably as a predetermined period, it can be set as one week, for example. An example of the information on the stock food stored in the stock food storage means D7 is shown in FIG. As shown in FIG. 7, the foodstuff ID and the foodstuff name are recorded in the stock food storage means D7 in association with the acquisition date which is the acquired date. The stock food storage means D7 is realized by securing a predetermined area for storing the consumption period in the storage device 10c.

  The food purchase information storage unit D8 is a storage unit that stores food purchase information. The food purchase information is information including the average price of each food, the purchase frequency, etc. in a specific area. Details of the food purchase information storage means D8 will be described later. The food purchase information storage means D8 is realized by securing a predetermined area for storing the food purchase information in the storage device 10c.

<2. Processing action>
Next, the processing operation of the recipe recommendation device according to the present embodiment will be described using the flowchart of FIG. Here, a case will be described in which when a user purchases food at a store and obtains a receipt, a recipe suitable for the purchased food and the food existing as stock is recommended. First, the user activates the recipe recommendation device 10. Specifically, an application program for recommending recipes is started. Then, the imaging unit 11 is activated. In this state, when the camera 10h is pointed at the receipt on which the name of the food is described and the photographing operation is performed, the imaging unit 11 images the receipt (step S10).

  When the captured image of the receipt is obtained by imaging, the character recognition unit 12 performs character string recognition on the captured image (step S20). Specifically, a character string that is a noun is extracted as a recognized character string by recognizing a character by a known OCR process and further performing a natural language process.

  Next, the food material candidate specifying means 13 specifies the food material candidate using the extracted recognition character string (step S30). Specifically, the food synonym storage means D1 shown in FIG. 4 is referred to by the recognized character string, and a synonym that matches a part or all of the recognized character string is searched. If there is a matching synonym, the matching synonym is recorded in the search log in association with the recognized character string. The search log is recorded in the RAM 10b.

  At this time, a search log as shown in FIG. 9A is obtained. As shown in FIG. 9, for example, when the recognized character string is “carrot one bag”, the synonym storage means D1 is referred to, and the synonym “carrot” is obtained, and the recognized character string is “cooked food”. In this case, the synonym “鮭” is obtained.

  Next, the NG word processing means 14 specifies a recognized character string including the NG word (step S40). Specifically, referring to the NG word storage means D2, it is determined whether or not the NG word is included in the recognized character string obtained in step S20. If the NG word is included, the fact that the NG word is included is searched. Record in the log. Specifically, when an NG word is included in the recognized character string, the NG word is recorded in the NG word column of the search log, and when the NG word is not included, NULL is set.

  At this time, a search log as shown in FIG. 9B is obtained. As shown in FIG. 9B, for example, the recognition character string “pumpkin soup” includes the NG word “soup”, and the recognition character string “chicken soup flavor” includes the NG word “flavor”.

  Next, the food specifying unit 15 specifies the food using the recognized character string stored in the search log (step S50). More specifically, the search log is analyzed, and the food is specified for the object whose synonym is hit and whose NG word is NULL. In the example of FIG. 9C, “carrots”, “salmons”, “shimeji”, and “Kaiida Daikon” whose synonyms are hit and whose NG word is NULL are specified as food ingredients. In the example of FIG. 9, for convenience of explanation, the food name is shown, but in reality, the food ID is recorded in the search log. If the food material can be specified, the food material name may be recorded as shown in FIG. 9 instead of the food material ID.

  Although not in the example of FIG. 9, there is a case where a certain word is recorded in the food synonym storage means D1 as a synonym of two or more foods. In that case, the food with the highest priority is selected from a plurality of candidate foods. The priority determination method can be set as appropriate, but in the present embodiment, the higher-order food ID (those with a smaller number) is selected as having a higher priority. For example, when “green onion” (indicating long onion) and “green onion” (indicating onion) are hit, “long onion” and “onion” are hit as ingredients. When the ingredient ID of “long onion” is “10201” and the ingredient ID of “onion” is “10202”, the ingredient ID of “long onion” is higher in the ascending order than the ingredient ID of “onion” (number is smaller) Therefore, “long onion” is selected and recorded in the item of food name in the search log.

  Subsequently, the stock food acquisition means 20 acquires information on stock food that is already purchased and in stock (step S55). Specifically, first, the consumption period of the food specified by the food ID is acquired by referring to the consumption period storage D6 with the food ID of the acquired food stored in the stock food storage D7. And the inventory food acquisition means 20 determines whether today is in a consumption period using the acquired consumption period and the acquisition date memorize | stored in the inventory food storage means D7. Specifically, the consumption period is added to the acquisition date, and if the date obtained by addition is after today, it is determined that it is within the consumption period, and the date obtained by addition is before today (before the previous day). If there is, it is determined that it is outside the consumption period.

  For example, it is assumed that today is June 14 and will be described with reference to the examples of FIGS. In this case, “chicken thigh” is obtained by adding the consumption period “1 day” shown in FIG. 6 to “June 12” which is the acquisition date shown in FIG. Since it is before “June 14” of today, it is determined that it is out of the consumption period. When “pumpkin” is added to the acquisition date “June 12” shown in FIG. 7 and the consumption period “3 days” shown in FIG. 6 is added, the last consumable date is “June 15”. Since “June 14th” and after, it is determined that it is within the consumption period. The stock food acquisition means 20 adds only the stock food determined to be within the consumption period to the food specified in step S50 as a recipe acquisition target. In the examples of FIGS. 6 and 7, “pumpkin” and “egg” are determined to be within the consumption period and are added as recipe acquisition targets.

  Next, the recipe acquisition means 16 performs a recipe acquisition process using the foodstuff specified by step S50 and the stock foodstuff added by step S55 (step S60). Specifically, first, the season information storage means D4 is referred to by the ingredient ID of the identified ingredient, and the seasonal point of each ingredient is obtained. In this embodiment, the seasonal point corresponding to the week including the date of the day is obtained. And the predetermined number of foodstuffs with the high seasonal point obtained are specified from the upper rank. Although it is possible to set the predetermined number as appropriate, in the present embodiment, the top five ingredients with the highest seasonal points are specified. At the time of selection, if two or more ingredients exist on the winning selection, the top one (the one with the smallest number) is selected in ascending order of the ingredient ID.

  FIG. 10 is a diagram showing the food item specified by the food material specifying means 15 that has the highest seasonal point. In the example of FIG. 10, there are six ingredients with the highest seasonal point, and there are two seasonal points, “sweet potato” and “tamago”, as the fifth ingredient from the highest. . In the example of FIG. 10, the ingredient ID of “sweet potato” is “10305”, the ingredient ID of “tamago” is “10326”, and the ingredient ID of “sweet potato” is higher in the ascending order than the ingredient ID of “tamago” ( Since the number is small), “Sweet potato” is selected as the fifth ingredient from the top. Therefore, in the example of FIG. 10, five asparagus, cabbage, chicken thigh, pumpkin, and sweet potato are selected.

  In this embodiment, when comparing foods according to seasonal points, a difference is made between the foods obtained by the current imaging and the stock foods. Specifically, for the stock ingredients, the recipe acquisition means 16 multiplies the seasonal point obtained by referring to the ingredient seasonal information storage means D4 by a predetermined value (for example, 0.7) less than 1 to obtain a new season. Use this seasonal point as a point. About the foodstuff obtained by this imaging, since the seasonal point obtained with reference to the foodstuff seasonal information storage means D4 is used as it is, it becomes easy to give priority to a stock foodstuff.

  In addition, among stock foods obtained by imaging receipts before the previous time, those with the same date as the current imaging time are used as they are without multiplying the seasonal point by a predetermined value less than 1. You can also By doing in this way, when it goes to two or more stores and the receipt is divided into two or more, it can judge by connecting the information of the foodstuff recorded on two or more receipts.

  When a food item having a high season point is identified, the recipe acquisition unit 16 refers to the recipe storage unit D3 by the food material ID of each identified food material, and searches for a recipe using each food material. If the recipe using each foodstuff is obtained, the total value of the season point of foodstuff will be calculated for every recipe about all the obtained recipes. When the seasonal point total value for each recipe is obtained, ranking is performed so that the recipe having the maximum seasonal point total value is ranked higher. When there are a plurality of recipes having the same seasonal point total value, the recipe including the food with the maximum seasonal point is set higher. Furthermore, when there are also a plurality of recipes containing ingredients with the maximum seasonal point, the recipe ID is set higher in order of ascending order (those with smaller numbers).

  FIG. 11 is a diagram illustrating the process of determining the priority order of recipes. FIG. 11A shows the relationship between the ingredients identified as the top five seasonal points and the recipes acquired from the recipe storage means D3 by any of the ingredients. FIG. 11 shows a case where a total of six recipes are obtained from any of the five ingredients. In the example of FIG. 11, recipes 1 and 2 are obtained from the ingredients “asparagus”, recipes 2, recipes 3, and recipes 5 are obtained from the ingredients “cabbage”, and recipes 4 and 5 are obtained from the ingredients “chicken thigh meat”. In this example, the recipe 5 is acquired by the ingredients “pumpkin”, and the recipe 1, recipe 3, recipe 4, and recipe 6 are acquired by the ingredients “sweet potato”.

  In this case, for each acquired recipe, the seasonal points of the ingredients from which the recipe is acquired are added together. And ranking is done from the highest seasonal point total value. When there are a plurality of recipes having the same seasonal point total value, priority is given to a recipe including ingredients with high seasonal points. For example, in the example of FIG. 11, both the recipe 2 and the recipe 5 have a total point value of 40 points, but the maximum seasonal point of the ingredients included in the recipe 2 is 25 points of “asparagus” Since the maximum seasonal point of the ingredients included in 5 is 15 points of “cabbage”, the recipe 2 including “asparagus” is given priority. As a result, the ranking as shown in FIG. As shown in FIG. 11B, ranking is performed in the order of recipe 2, recipe 5, recipe 1, recipe 3, recipe 4, and recipe 6.

  As described above, recipe selection and ranking are performed. You may make it recommend to a user by displaying the recipe obtained at this time. However, in this embodiment, the existing recipe is excluded from the selected recipe. Specifically, the existing recipe excluding means 17 excludes the existing recipe from the selected dish (step S70). The already-existing recipe excluding means 17 refers to the already-existing recipe storage means D5 and excludes a recipe having the same recipe ID as the stored recipe ID from the recommendation target. Thereby, the recipe which should be recommended finally is determined.

  Next, the recipe display means 19 displays the determined recipe to be recommended on the screen (step S80). As a result, the user can know the optimum recipe using the ingredients just purchased on the spot.

  On the other hand, when a recipe to be recommended is determined in step S70, the search log transmission unit 18 transmits a search log as a processing result of the recognized character string to the management server 20 in parallel with the screen display in step S80 (step S70). S90). Specifically, a search log from the recognized character string to the search flag as shown in FIG. “TRUE” in the search flag indicates that the recipe can be acquired from the recipe storage means D3 after specifying the synonym and the food ingredient ID using the recognized character string. “FALSE” in the search flag indicates the recognized character string. Is used to indicate that the recipe could not be finally obtained from the recipe storage means D3. That is, in the example of FIG. 9D, it is shown that the recipe can be acquired from the recognized character string “Boiled Food”, and the recipe can be acquired from the recognized character string “Bunage Meiji”.

  The transmitted search log is received and registered by the management server 20. At this time, the management server 20 assigns and manages an error flag item. FIG. 12 is a search log recorded in the management server 20. The search log in the recipe recommendation device 10 shown in FIG. 9D is recorded in the management server 20 as a search log in the state shown in FIG. FIG. 12A is exactly the same as FIG. 9D except that an error flag item exists.

  The display in step S80 will be described. FIG. 13 is a diagram showing a recipe display screen. As shown in FIG. 13, the recipe display screen displays a recipe name, a food photo, and other information. In the rectangular area shown by shading in FIG. 13, a dish photo is actually displayed. Further, a detail display button, a cancel button, a recipe memo addition button, and an error report button are displayed, and an instruction can be given by performing a predetermined operation.

  When the error report button is operated by the user, it is determined that there is an error report instruction (step S100), and the search log transmission unit 18 transmits error information to the management server 20 (step S110). As the error information, information indicating that the displayed recipe is an error can be transmitted. In the present embodiment, the same information as that transmitted in step S90, that is, as shown in FIG. The search log that is the processing result of the recognized character string is transmitted.

  The management server 20 compares the received search log with the already recorded search log, and if they match, records “TRUE” in the error flag item. Thereby, this series of search logs records that the user has determined an error.

  By checking the search log in which “TRUE” is recorded in the error flag item, the administrator can infer that there is a problem in the process from the recognized character string to the recipe acquisition. For example, in FIG. 13, the recipe “Tsubaki-chan-chan” is recommended. This is obtained based on the ingredients “１２” and “shimeji” whose search flag is “TRUE” as shown in the search log of FIG. 12B, but corresponds to the ingredient “鮭”. The recognition character string is “Food feed”.

  From the search log, the administrator purchased the “boiled rice”, but did not purchase “boiled rice”, but recommended a recipe using “boiled rice” as an ingredient. You can guess that the button was pressed. Based on this estimation, it is possible to improve the recipe recommendation device 10 so that the ingredient “鮭” is not identified from the recognized character string “鮭 of food”.

<3. Update of seasonal points>
In the above example, the seasonal points stored in the ingredient seasonal information storage means D4 are given according to each week of 52 weeks of the year, and other than that, there is no influence, but various information You may make it update by. Next, the case where a seasonal point is updated according to the purchase information in a certain area is demonstrated.

  The recipe recommendation device 10 periodically accesses the management server 30 via the communication unit 10g, acquires the food purchase information, and stores it in the food purchase information storage unit D8. FIG. 14 is a diagram illustrating an example of the ingredient purchase information stored in the ingredient purchase information storage unit D8. As shown in FIG. 14, the ingredient purchase information is recorded in the items of rank, ingredient name, average price, purchase frequency, score, and added season point. Although omitted in FIG. 14, in fact, the food item purchase information also includes a food material ID that identifies the food material. The average price indicates the ratio of the average price of the food in a specific area to the national average price. For example, if the average price of spinach in a specific area is 142 yen and the national average price of spinach is 211 yen, -32% is recorded because spinach in a specific area is 32% cheaper than the national average.

  The purchase frequency indicates the ratio of the purchase frequency of the food in a specific area to the national average purchase frequency. For example, if the purchase frequency of spinach in a specific area is 6.2% and the national average purchase frequency of spinach is 4.5%, the record is + 37% because spinach in a specific area is 37% more than the national average. Is done. Various methods can be used as the method for calculating the purchase frequency, but in the present embodiment, the method is calculated as a ratio to the total shopping unit (receipt unit).

  The score is a value indicating the degree of the season, and is obtained by inverting the sign of the average price value and adding it to the purchase frequency value. Therefore, the lower the average price compared to the national average and the higher the purchase frequency compared to the national average, the higher the score. The fact that the average price is lower than the national average indicates that there is a particularly large supply in a specific area, so it is likely that it is more seasonal. In addition, the fact that the purchase frequency is higher than the national average indicates that many customers are making frequent purchases, so it is likely that it is more seasonal.

  The added season point is a point determined based on a predetermined rule based on the calculated score. Although it is possible to use the calculated score as an addition season point as it is, it is converted based on a predetermined rule in consideration of the influence on the season point stored in the food season information storage means D4. In the example of FIG. 14, a rule is adopted in which the added season point is set to 0 until a predetermined score, and the added season point is increased stepwise as the score increases.

  The rank is a rank assigned in descending order of scores. If the score is the same, the item with the highest purchase frequency is ranked higher. Among the items shown in FIG. 14, the average price and the purchase frequency are calculated based on information collected from each recipe recommendation device 10. The score, the addition point, and the rank may be calculated by the management server 30, but may be calculated by the recipe recommendation device 10. When calculating in the recipe recommendation device 10, the ingredient season information update unit 21 calculates a score and an additional season point using the average price and purchase frequency of the ingredient purchase information received from the management server 30.

  In the case where an additional season point is included in the ingredient purchase information received from the management server 30, the ingredient season information update unit 21 calculates the additional season point, and if the additional season point is newly obtained, The seasonal information update unit 21 updates the seasonal point of the ingredient seasonal information storage unit D4 by adding the addition seasonal point of each ingredient to the seasonal point stored in the ingredient seasonal information storage unit D4.

  Thus, since the seasonal point of the ingredients seasonal information storage means D4 is periodically updated, the seasonal point of each ingredient to be referred to reflects the season in the specific area in the recipe acquisition process of step S60.

  Since the management server 30 calculates the average price and the purchase frequency, it is necessary to acquire an individual purchase price for each purchase in the specific area. The individual purchase price can be acquired from a POS or the like, but is acquired from each recipe recommendation device 10 in this embodiment. Specifically, after the imaging unit 11 captures a receipt in step S10 and the character recognition unit 12 performs character string recognition in step S20, the purchase price of the purchased ingredients is calculated in parallel with the processing after step S30. get. Then, the recipe recommendation device 10 transmits the purchase price information included in one receipt obtained by one imaging to the management server 30 via the communication unit 10g in association with the food ID and the food name. . At this time, information indicating a specific area set in advance is also transmitted.

  When receiving the purchase price information corresponding to one receipt from the recipe recommendation device 10, the management server 30 records the purchase price in association with the received specific area. Each time it is received, it is added as the number of purchases (number of receipts), which is the denominator when calculating the purchase frequency. As described above, since information indicating a specific area set in advance is transmitted from each recipe recommendation device 10, the average price and the purchase frequency can be calculated for each area, The purchase frequency can also be calculated.

<4. Modified example>
The preferred embodiments of the present invention have been described above. However, the present invention is not limited to the above embodiments, and various modifications can be made. For example, in the above-described embodiment, the identification process of the food candidate (step S30) is performed on the recognized character string first, and the identification process of the character string including the NG word (step S40) is performed later. The order may be reversed. When it is expected that there are many recognized character strings including NG words, the order is reversed, the character strings including NG words are specified, and then the candidate foods are specified with the recognized character strings excluded. By performing the processing, the processing speed can be increased.

  Moreover, in the said embodiment, although the receipt on which the foodstuff name was printed was image | photographed and the character string for specifying a foodstuff was recognized, if the character string regarding a foodstuff was described, it will be other than a receipt. It is also possible to use. For example, a memo in which the name of the food is written by hand may be photographed and recognized by OCR to obtain a character string.

  Moreover, in the said embodiment, although the stock food acquisition means 20 acquired the information of stock food and made it the object of a recipe acquisition in step S55, you may make it abbreviate | omit the process of this step S55. By using information on stock ingredients, it is possible to acquire recipes that also use previously acquired ingredients on the spot, but even if you do not use information on inventory ingredients, only the ingredients that you currently purchase You can get the recipe you used.

10 ... Recipe recommendation device 10a ... CPU
10b ... RAM
10c: Storage device 10d: Instruction input unit 10e: Data input / output I / F
10f ... Display unit 10g ... Communication unit 10h ... Camera 11 ... Imaging means 12 ... Character recognition means 13 ... Food material candidate specifying means 14 ... NG word processing means 15 ... Ingredient identification means 16 ... Recipe acquisition means 17 ... Existing recipe exclusion means 18 ... Search log transmission means 19 ... Recipe display means 20 ... Inventory food acquisition means 21 ... Ingredient seasonal information update means 30 ... Management server 40 ... Network D1 ... Food synonym storage means D2 ... NG word storage means D3 ... Recipe storage means D4 ... Ingredient seasonal information storage means D5 ... Existing recipe storage Means D6 ... Consumption period storage means D7 ... Inventory food storage means D8 ... Food purchase information storage means

Claims (8)

An ingredient synonym storage means for storing the ingredient information and the synonym of the ingredient in association with each other;
NG word storage means storing NG words to be excluded as food ingredients;
Recipe storage means for storing recipes and food information in association with each other;
An imaging means for imaging an object including a character string;
Character recognition means for performing character recognition processing on the captured image obtained by the imaging means to obtain a recognized character string;
With reference to the food synonym storage means by the recognized character string, a food candidate candidate specifying means for specifying a recognized character string as a food candidate;
NG word processing means for identifying a recognized character string including an NG word by referring to the NG word storage means by the recognized character string;
Among the recognized character strings, a food material specifying means for specifying a food material corresponding to a recognized character string corresponding to a food material candidate and not including an NG word as a food material to be used;
Recipe acquisition means for searching the recipe storage means by the specified ingredients and acquiring a recipe;
Recipe display means for displaying the acquired recipe;
A recipe recommendation device comprising: In addition, the food season information storage means for storing information on the season of the food,
The recipe acquisition means refers to the ingredient season information storage means by the ingredients specified by the ingredient identification means, acquires the season level of each ingredient, and acquires a recipe giving priority to ingredients having a higher season grade. The recipe recommendation device according to claim 1.   It further includes season information update means for updating information about the season of the ingredients stored in the ingredient season information storage means using information about the season of ingredients determined based on the purchase information of the ingredients in the specific area. The recipe recommendation device according to claim 2. The said foodstuff identification means performs the process which memorize | stores the information of the specified foodstuff in an inventory food storage means as an inventory foodstuff,
The said recipe acquisition means searches the said recipe memory | storage means using the stock foodstuff memorize | stored in the said stock foodstuff memory means in addition to the foodstuff specified by the said foodstuff identification means. Item 4. The recipe recommendation device according to any one of items 3 to 4. An existing recipe storage means for storing an existing recipe;
Out of the recipes acquired by the recipe acquisition means, the recipe recipe excluding means for excluding those that overlap with the existing recipe stored in the existing recipe storage means,
The recipe recommendation device according to any one of claims 1 to 4, wherein the recipe display unit displays a recipe that has not been excluded by the existing recipe exclusion unit.   6. The apparatus according to claim 1, further comprising a search log transmission unit that transmits a search log for the recognized character string obtained in the process of acquiring the recipe to a management server. Recipe recommendation device given in the paragraph.   The recipe according to claim 6, wherein the search log transmission unit transmits the search log as error information to the management server when an instruction is given to the recipe displayed by the recipe display unit. Recommendation device.   The program for functioning a computer as a recipe recommendation apparatus as described in any one of Claims 1-7.

Images