Merchandise handling device

Abstract

The present invention provides a product processing apparatus. The product processing apparatus (1) includes a control unit (10) which retrieves product information from a product master table, either through a product call operation or by an image recognition device, and displays it on a display unit (5). When a product call operation is performed without specifying a product, the control unit (10) activates the image recognition device (30).

Description

Technical Field

[0001] One aspect of the present invention relates to a commodity processing apparatus. Background Technology

[0002] Product processing devices, such as a measuring label issuing device, are known. For example, Patent Document 1 (Japanese Patent Application Publication No. 2024-41228) discloses a label issuing device that identifies a product by photographing it and reading feature values ​​from the photographed image, retrieves information about the identified product from a product master table, and prints it on a label. Furthermore, recently, the use of AI (Artificial Intelligence) based image recognition devices for product identification has become common. Summary of the Invention

[0003] When using the aforementioned image recognition device to retrieve product information from the product master table and display it on the display unit, there are cases where the product can be accurately identified as the target and cases where it cannot. For example, if the product has a fixed shape, the image recognition device can accurately identify the product. However, for products that do not have this characteristic (e.g., cooked food whose ingredients vary depending on the season or whose surface ingredients change each time), it cannot accurately identify the product, and therefore it takes a considerable amount of time to lock in the product to be priced.

[0004] Furthermore, when the image recognition device specifies multiple products, it displays a list of these products and selects the matching product. However, this method has the following problems: repeating this operation each time is cumbersome, and the product pricing process is interrupted. Moreover, once conventional image recognition devices are introduced into the work area, they cannot improve the accuracy of product recognition. Therefore, even with the introduction of image recognition devices, the pricing process cannot be made more efficient.

[0005] One aspect of the present invention is to provide a product processing apparatus that can efficiently retrieve product information stored in a product master table even when an image recognition device is incorporated.

[0006] Methods for solving problems (1) A product processing apparatus according to one aspect of the present invention includes: a control unit that obtains product information of a product specified by a product call operation or a product specified by an image recognition device from a product master table and displays it on a display unit; when a product call operation is performed without specifying a product, the control unit activates the image recognition device.

[0007] In this product processing device, the product retrieval operation is performed when it is faster to retrieve product information by entering a retrieval number for the specified product. For example, after entering the product's unique retrieval number, the retrieval key is pressed. The control unit then retrieves the product information corresponding to the retrieval number from the product master table and displays it on the display unit. However, if the operator does not remember or is unsure of the retrieval number, retrieving the retrieval number is time-consuming, so a retrieval operation without specifying a product is performed. For example, a number not associated with a product, such as 0, is pressed, followed by the retrieval key. Or, only the retrieval key is pressed. The control unit then determines this as a retrieval operation without specifying a product and activates the image recognition device. The image recognition device then specifies the product based on its image data, retrieves the specified product information from the product master table, and displays it on the display unit. Thus, when the product retrieval operation is faster and product information can be displayed, the operator can perform the usual retrieval operation. Otherwise, by activating the image recognition device to display the specified product information, the pricing operation can be made more efficient. In addition, even for items whose recall number is remembered, the image recognition device can be activated simply by pressing the recall key, without entering the recall number. In this case, it can be confirmed whether the image recognition device can specify the correct item name.

[0008] (2) Alternatively, the product processing device described in (1) above may further include: a voice output device that, when the display unit displays product information related to the product, at least announces the product name contained in the product information. In this configuration, when product information is displayed through a product call operation, the displayed product name is announced, thus enabling the operator to confirm whether the call operation is correct. Furthermore, when a product is specified by an image recognition device and its product information is displayed, the operator can also confirm whether the product specification performed by the image recognition device is correct.

[0009] (3) Alternatively, the product processing apparatus described in (2) above, wherein when multiple products are specified by the image recognition device, the control unit causes the display unit to display the multiple specified products, and when a product is selected from the multiple displayed products, the display unit causes the display unit to display the product information of the selected product. In this structure, when multiple products are specified by the image recognition device, a matching product can be selected from the multiple products. Therefore, even if the recognition accuracy of the image recognition device is low, it can be compensated for by the product selection operation.

[0010] (4) Alternatively, any one of the commodity processing apparatuses described in (1) to (3) above may further include: a measuring unit that measures the commodity; and a label issuing unit that prints the weight of the commodity measured by the measuring unit and commodity information obtained from the commodity master table onto a label and issues it. In this structure, a price tag with the weight of the commodity printed on it can be issued, thus enabling the price tag to be affixed to the commodity for pricing.

[0011] (5) Alternatively, any one of the commodity processing devices described in (1) to (4) above may further include: a camera unit that takes pictures of commodities, and a control unit that provides the image data of the commodities taken by the camera unit, as well as the commodity name of the commodities specified by the commodity call operation or specified by the image recognition device, as training data to the image recognition device and enables it to perform machine learning. According to this structure, the image recognition device performs machine learning based on the training data, thereby improving the recognition accuracy of the image recognition device. Therefore, even if the recognition accuracy of the image recognition device is low, the recognition accuracy can be improved by repeatedly performing this machine learning, making the pricing operation more efficient.

[0012] According to one aspect of the present invention, even with the introduction of an image recognition device, it is possible to efficiently retrieve product information stored in the product master table. Attached Figure Description

[0013] Figure 1 This is a perspective view showing one embodiment of a metering tag issuing device.

[0014] Figure 2 yes Figure 1 A block diagram of a metering label issuing device.

[0015] Figure 3 This is an example of the initial display of product information on the display screen.

[0016] Figure 4 This is an example of a product identification dialog box displayed on the display panel.

[0017] Figure 5A This is an example of a product information screen displayed on the display panel. Figure 5B It is by Figure 1 An example of a product label issued by a metering label issuing device.

[0018] Figure 6A This is an example of a product selection dialog box displayed on the screen. Figure 6B This is an example of a product information screen displayed on the display panel.

[0019] Figure 7 This is an example of the operation process of a metering label issuing device equipped with an image recognition unit.

[0020] Figure 8 This is an example of a product information screen displayed on the display panel.

[0021] Figure 9 This is an example of a product main display screen shown on the display panel.

[0022] Figure 10A This is a perspective view showing a metering label issuing device and a metering packaging pricing device, including variations. Figure 10B This is a perspective view showing a modified example of a face-to-face metering label issuing device. Detailed Implementation

[0023] Hereinafter, a measuring label issuing device 1 according to one embodiment of a merchandise handling apparatus will be described with reference to the accompanying drawings. This measuring label issuing device 1 is a device that integrates a measuring device and a label issuing device, but it can also be a label issuing device without a measuring device. In this case, it has a structure with an attached product placement section and a product imaging section for imaging the product on the placement section. In the following description, the same reference numerals are used to denote the same elements, and repeated descriptions are omitted. Furthermore, in this specification, "A or B" may include either A or B, but does not exclude the inclusion of both A and B.

[0024] The weighing label issuing device 1 measures the weight of goods and issues a product label LC (refer to) to be affixed to the weighed goods. Figure 5B The label issuing device 1 of this embodiment is configured to retrieve product information of a product specified by a product recall operation or image recognition by the image recognition unit (image recognition device) 30 from the storage unit 12 and display it on the display unit 5. Figure 1 and Figure 2 As shown, the metering label issuing device 1 includes a main body shell 3, an operation unit 4 including a fixed key, a display unit 5, a speaker 6, a metering unit 7, a shooting unit 8, a printing unit 9, a storage unit 12, and a controller (control unit) 10.

[0025] The metering label issuing device 1 has a built-in image recognition unit 30, but it can also be structured as follows: the image recognition unit 30 is separated from the metering label issuing device 1 and constituted by a computer equipped with AI, and the computer is connected to the metering label issuing device 1. In this case, only communication between the two devices is added. Therefore, the following will describe the case where the image recognition unit 30 and the metering label issuing device 1 are integrated and do not require this communication operation.

[0026] An operation unit 4, a display unit 5, a speaker 6, a printing unit 9, and a controller 10 are provided on the main body casing 3. The main body casing 3 is formed into a generally rectangular parallelepiped shape. An opening and closing door is provided on the front surface of the main body casing 3. By opening the opening and closing door, the printing unit 9, which is disposed inside the main body casing 3, is exposed.

[0027] The operation unit 4 is located on the main body casing 3. The operation unit 4 is the part that receives various operations from the operator. The operation unit 4 includes the "unit price" key, "fixed amount" key, "tare" key, "print" key, "recall" key, and number keys, which are necessary for a price-computing scale.

[0028] Display unit 5 displays various information such as the status and operation of the measuring label issuing device 1, and information related to the goods being measured. Display unit 5 may also be configured to include a touch panel. In this case, similar to operation unit 4, it functions to receive various operations from the operator.

[0029] The speaker 6 is located near the display unit 5. Alternatively, the speaker 6 can be located in other positions, or it can be an external speaker separate from the main housing 3. The speaker 6 outputs (broadcasts) product information, including at least a portion of the product name. The timing and content of the voice output are controlled by the controller 10. The output content refers to, for example, the product name, etc., emitted from the speaker 6. In this embodiment, the speaker 6 and the controller 10 constitute a voice output device 15.

[0030] The measuring unit 7 is separately constructed from the main housing 3. Therefore, it can also be a simple label issuing device (product processing device) without the measuring unit 7. The measuring unit 7 mainly includes a measuring pan 7A, a weighing sensor (not shown), a signal processing circuit, and a transmitting module. The weighing sensor is located at the bottom of the measuring pan 7A and supports the measuring pan 7A, converting the mechanical strain generated by placing the object to be measured on the measuring pan 7A into an electrical signal and outputting it. The signal processing circuit amplifies the electrical signal output from the weighing sensor and converts it into a digital signal. The transmitting module transmits the digital signal wirelessly or wiredly to the controller 10 inside the main housing 3.

[0031] The imaging unit 8 photographs the product. The imaging unit 8 is installed on the main housing 3 so as to photograph the entire measuring tray 7A. When a product is placed on the measuring tray 7A, the imaging unit 8 photographs the product and generates imaging information. The imaging unit 8 may be, for example, a CCD image sensor or a CMOS image sensor that acquires color images. The imaging unit 8 may include a stereo camera or an infrared camera that acquires temperature images of the product. The imaging unit 8 outputs the imaging information to the controller 10. Furthermore, in the case of a simple label issuing device without the measuring unit 7, a product placement unit is provided separately to replace the measuring unit 7.

[0032] The printing unit 9 is housed within the main body casing 3 and includes a box (not shown), a printing unit, and a cutting unit. The box supports the label roll in a removable manner. The label roll is formed by winding strip-shaped backless labels onto a paper tube. The backless label (hereinafter referred to as "label") has an adhesive coated on the back of the paper substrate, a heat-sensitive agent that develops color upon heating coated on the front, and a silicone resin as a release agent further coated on top. The label roll is wound onto the paper tube with the back (adhesive side) of the label as the inside. Alternatively, in this embodiment, a label roll with backed labels can be used instead of a backless label. The printing unit 9 prints product information on the label and issues it as a product label LC.

[0033] The printing unit includes a print head and an impression roller, which prints product information onto the front of the label. The print head is a thermal print head. The print head is controlled by the controller 10, which will be described in detail later. The impression roller presses against the back of the label, pushing the front of the label against the print head. The cutting unit cuts the label from the printing unit into product labels LC of a specified length (see reference). Figure 5B That is, the cutting unit issues the portion printed by the printhead as the product label LC. Furthermore, the printing method of the printing unit 9 is an example; various printing methods such as thermal transfer printing, inkjet printing, etc., can be used.

[0034] Storage unit 12 stores various information. Storage unit 12 may be composed of, for example, RAM, SSD (Solid State Drive), HDD (Hard Disk Drive), etc. In this embodiment, storage unit 12 stores product information for each product in a product master table. Product information includes the product's call number (call No.), the product name of the specified product, the name of the raw materials, net content, processing / manufacturing date, shelf life / best before date, storage temperature, processor / manufacturer, etc. In the product master table, besides... Figure 9 In addition to the product information shown on the main product display screen SC4, information about whether machine learning is performed on each product by the image recognition unit (image recognition device) 30, which will be described later.

[0035] The controller 10 is part of the control device 1 for issuing metering tags. It includes a CPU (Central Processing Unit), ROM (Read Only Memory), RAM (Random Access Memory), etc., which are interconnected via buses such as address buses and data buses. The controller 10 can be configured as software that loads a program stored in ROM onto RAM and executes it via the CPU. The controller 10 can also be configured as hardware based on electronic circuits, etc. The controller 10 has an image recognition unit 30 and an action control unit 40 as functional components, including an input unit 31, a designation unit 32, a model generation and storage unit 33, and an output unit 34. The image recognition unit 30 and the action control unit 40 can be configured as software that loads a program stored in ROM onto RAM and executes it via the CPU.

[0036] The input unit 31 inputs the image data output from the imaging unit 8 to the designation unit 32. The data input to the designation unit 32 may also be data extracted from the image data output from the imaging unit 8, specifically the area corresponding to the product (the product or the product and its surroundings). In addition, the input unit 31 accepts the learned model sent by a server device (not shown) or a computer equipped with AI and stores it in the model generation storage unit 33.

[0037] The model generation and storage unit 33 stores the learned models generated by external AI, and also stores newly generated learned models based on the training data input from the motion control unit 40 through machine learning.

[0038] The designation unit 32 uses the learned model obtained from the model generation and storage unit 33 to designate the product name based on the product image data captured by the imaging unit 8. The designation unit 32 detects that a product is placed on the measuring plate 7A based on the image data input from the input unit 31. Specifically, the designation unit 32 detects that a product is placed based on the difference (background difference) between the image data input from the input unit 31 and a pre-stored base image (background image).

[0039] When a product is detected, the designation unit 32 begins designating the product based on a measurement stabilization signal from the motion control unit 40. In this embodiment, the designation unit 32 uses a learned model from the model generation and storage unit 33 to designate the product in the image data. The learned model designates the product based on the features of the product represented by the image data. In this case, clustering is used when the product is of different types such as fresh fish, cooked food, and fresh vegetables. The learned model may include a neural network (CNN) or a Transformer for clustering. The learned model may also further include a multi-layer (e.g., 8 or more layers) neural network. That is, the learned model may also be generated through deep learning. Furthermore, the learned model may be prepared for each product to be processed. For example, it may be a learned model categorized as lean meat, fresh fish, fresh vegetables, cooked food, etc.

[0040] The designation unit 32 uses the learned model from the model generation and storage unit 33 to designate the product name of the image data input from the imaging unit 8. In response to inputting image data into the neural network of the learned model, the designation unit 32 obtains the estimation result output from the neural network. The estimation result may include cases where multiple similar products are included or cases where only one product is included. Additionally, it may sometimes include information indicating that a product cannot be designated. Based on such estimation results, the designation unit 32 outputs the product name of the product placed in the measurement unit 7 to the output unit 34. The output unit 34 outputs the product name input from the designation unit 32 to the motion control unit 40.

[0041] The motion control unit 40 controls various actions of the metering tag issuing device 1. For example, the motion control unit 40 controls... Figure 3 The guidance message shown is displayed on the product information screen SC1. Additionally, when a product placed on the measuring plate 7A is measured by the measuring unit 7, and the "call" key is pressed after inputting a call number specifying the product from the operation unit 4, the motion control unit 40 displays the product information of the product specified by the call number as shown below. Figure 5A The product information screen SC1 is shown. Figure 5A The display shows that the item placed on the measuring section 7 is a fixed-amount item, and the entered call number is "0012". Next, when the "Print" button on the operation section 4 is pressed, the item is issued as follows: Figure 5B The product label shown is LC.

[0042] The product retrieval operation (regular retrieval operation) involves, for example, entering the product's unique retrieval number and then pressing the "Retrieve" button. The retrieval number is a unique number registered for each product, thus allowing the product to be specified using the retrieval number. When a product is specified by entering the retrieval number, the motion control unit 40 retrieves the product information of the specified product from the product master table in the storage unit 12 and causes the display unit 5 to display the information as shown below. Figure 5A The product information screen shown is SC1.

[0043] On the other hand, if the operator does not remember the call number, a call operation without specifying a product is performed in order for the image recognition unit 30 to assign a product (a temporary call operation different from the regular call operation). For example, "0" is entered as an unregistered number, and then the "Call" key is pressed. Or, only the "Call" key is pressed. Then, the motion control unit 40 determines that it is a call operation in which the product has not yet been assigned, and causes the image recognition unit 30 to assign the product name of the product placed on the measuring unit 7.

[0044] When the image recognition unit 30 specifies the product name of the item placed on the measuring plate 7A, the product name is input to the motion control unit 40. Based on the input product name, the motion control unit 40 retrieves the product information from the main product table and displays it on the display unit 5. At this time, as... Figure 4 As shown, the motion control unit 40 causes the display unit 5 to display a product recognition dialog box SC2 indicating that the product has been recognized. The product recognition dialog box SC2 displays the call number from the product information of the product recognized by the image recognition unit 30 (in... Figure 4 The example shown is number 0012) and the product name (in Figure 4 The example shown is salmon nigiri sushi. Additionally, the product recognition dialog box SC2 includes an OK button B21 for the operator to press when the product name specified by the image recognition unit 30 is deemed correct, and a Cancel button B22 for the operator to press when the product name is deemed incorrect.

[0045] When in Figure 4 When the OK button B21 is pressed in the product identification dialog box SC2, the motion control unit 40 determines that the product has been specified and causes the display unit 5 to display the following: Figure 5A The product information screen SC1 is shown. Next, the motion control unit 40 controls the speaker 6 to broadcast the product information, such as the product name and amount, displayed on the product information screen SC1.

[0046] In addition, Figure 4 In the product recognition dialog box SC2, clicking the OK button B21 will redirect to... Figure 5A In the case of the product information screen SC1, the motion control unit 40 also outputs via voice the product name displayed on the product information screen SC1, which is the product name specified by the image recognition unit 30 (e.g., "Salmon nigiri sushi recognized by the image recognition unit"). In this state, when the operator operates, for example, the "print" button included in the operation unit 4, the printing unit 9 is controlled to print, such as... Figure 5B The product label shown is LC.

[0047] On the other hand, when the designation unit 32 extracts multiple product candidates and reports the results to the motion control unit 40, the motion control unit 40, for example, causes the display unit 5 to display the product selection dialog box SC3 as shown in FIG6. At this time, the motion control unit 40 can, for example, display multiple products sequentially starting from those with high recognition rates output by the image recognition unit 30 (learned model).

[0048] exist Figure 6A In the example of the product selection dialog box SC3 shown, "Product No. 0010: Sardine Nigiri Sushi", "Product No. 0011: Saury Nigiri Sushi", "Product No. 0012: Salmon Nigiri Sushi", and "Product No. 0013: Tuna Nigiri Sushi" are displayed. These displays are configured so that the operator can select items in the product selection dialog box SC3. The product name selected by the operator is displayed in reverse black and white, making it easy to identify that it has been selected. The motion control unit 40 then retrieves the product information of the selected product from the product master table in the storage unit 12 and displays it on the display unit 5 as shown below. Figure 6B The product information screen SC1 is shown. Additionally, if the image recognition unit 30 cannot specify any product, the motion control unit 40 causes the display unit 5 to display a message such as "Product not recognized. Please enter a call number".

[0049] exist Figure 6B In the product information screen SC1 shown, when the learning setting unit B1 used for machine learning is touched, the display switches from "performing" learning to "not performing" learning. However, the learning setting unit B1 is set to "performing" image recognition as an initial value. When "performing" image recognition is set, the motion control unit 40 outputs the image data of the product on the measuring plate 7A and the product name "salmon nigiri sushi" as training data to the input unit 31 of the image recognition unit 30. The model generation and storage unit 33 performs machine learning based on the training data received from the input unit 31, generates a new learned model, and updates the storage. However, the image recognition unit 30 can also send the machine learning to an external AI for execution, obtain its results, and store them in the model generation and storage unit 33.

[0050] By repeatedly performing this machine learning process, the recognition accuracy of the designated part 32 can be improved. Therefore, simply placing "salmon nigiri sushi" on the measuring plate 7A will display the result as shown below. Figure 6B The product information screen SC1 shown is displayed, without the need to display as shown in the image. Figure 6A The product list is shown. When this stage is reached, touch the learning setting unit B1 to switch the "Execute" display to the "Do Not Execute" display (see reference). Figure 8(Learning setting unit B1). Thus, the machine learning in the model generation and storage unit 33 ends, and afterwards, even if the appearance of "salmon nigiri sushi" changes, the designation unit 32 can still designate the product in the image data as "salmon nigiri sushi" and display it. Figure 6B The product information screen shown is SC1.

[0051] On the other hand, when a product is retrieved from the operation unit 4, for example, by entering "0012" as the retrieval number and then pressing the "retrieve" key, the motion control unit 40 retrieves the product information of the product name "Salmon Nigiri Sushi" corresponding to the retrieval number "0012" from the storage unit 12, and displays it on the display unit 5. Figure 5A The product information screen SC1 is shown. Next, the motion control unit 40 controls the speaker 6 to broadcast the product information, such as the product name and amount, displayed on the product information screen SC1.

[0052] Next, when it is displayed Figure 5A or Figure 6B When the "Print" button is pressed while the product information screen SC1 is active, the motion control unit 40 controls the printing unit 9 to print the product label LC for "Salmon Nigiri Sushi". As a result, the label issuing device 1 issues the product label LC for "Salmon Nigiri Sushi".

[0053] Next, the operation of controller 10 will be explained. Figure 7 An example illustrating the operation flow of controller 10 is given, assuming the following scenario: when an item is placed... Figure 1 When the measuring unit 7 determines that the measuring is stable on the measuring plate 7A, the image data of the product acquired by the imaging unit 8 is output to the controller 10.

[0054] exist Figure 7 In step S1, a product retrieval operation is performed. For example, "salmon nigiri sushi" is placed on the measuring plate 7A, and "0012" is entered as the retrieval number from the operation unit 4. Next, when the "retrieve" button is pressed, in step S2, the motion control unit 40 determines whether the retrieval number is a 4-digit regular number (a registered number). If it is a regular number (a registered number), the process proceeds to step S3, where the product information corresponding to the retrieval number "0012" is retrieved from the product master table in the storage unit 12 and displayed on the display unit 5. Figure 5A and Figure 6B This is an example of how product information is displayed on screen SC1.

[0055] On the other hand, in step S1, if the operator does not remember the call number of the item placed on the measuring plate 7A, the operator may, for example, press the number key "0" and then press the "Call" key. Alternatively, the operator may simply press the "Call" key. Then, in step S2, the motion control unit 40 determines that the input call number is an irregular number (unregistered number) and proceeds to step S4, whereby the motion control unit 40 activates the image recognition unit 30. That is, the designation unit 32 of the image recognition unit 30 designates the item on the measuring plate 7A based on the image data input from the imaging unit 8 and reports the result to the motion control unit 40.

[0056] In step S5, if the report indicates that a product cannot be specified, the motion control unit 40 proceeds to step S6, and the display unit 5 displays a message such as "No matching product. Please enter a call number." If a product can be specified, the process proceeds to step S7 to determine whether there are multiple or a single specified product. If there are multiple specified products, the process proceeds to step S8, where a message such as "No matching product. Please enter a call number" is displayed. Figure 6A The system displays multiple product names and allows users to select one product. When a product is selected, the motion control unit 40 proceeds to step S9, retrieves the product information of the selected product from the product master table, and displays it on the display unit 5. In step S7, if only one product is specified, the system also proceeds to step S9, retrieves the product information of the specified product from the product master table, and displays it on the display unit 5. Figure 6B Here is an example of how the product information screen SC1 is displayed.

[0057] Thus, when Figure 6B When the product information screen SC1 is displayed, the motion control unit 40 moves to step S10 and controls the speaker 6 to announce the product information displayed on the product information screen SC1, such as the product name and price. The operator checks whether the product name in the product information screen SC1 is correct based on the product information screen SC1 and the announced product name. If there is no error, the operator presses the "Print" button on the operation unit 4 in step S11. Then, in step S12, the motion control unit 40 controls the printing unit 9 to issue the product label LC for "Salmon Nigiri Sushi".

[0058] When the product tag LC is issued, in step S13, the motion control unit 40 determines whether the "salmon nigiri sushi" product requires machine learning. Specifically, in Figure 6B In the product information screen SC1, if "Execute" learning is displayed in the learning settings section B1, it is determined that machine learning is required; if "Do not execute" learning is displayed, it is determined that machine learning is not required.

[0059] The learning setting unit B1 that makes the judgment in step S13 is included in both the product information screen SC1 when the product is specified by the product call operation and the product information screen SC1 when the product is specified by the image recognition unit 30. Therefore, in the case of the product call operation, the recognition accuracy of the image recognition unit 30 is also improved by machine learning.

[0060] In the learning setting section B1 of the product information screen SC1, "Execute" learning is set as the initial value. This is based on the understanding that when the image recognition unit 30, as an image recognition device, is first introduced, the recognition accuracy will not improve without repeated machine learning. Therefore, when learning from... Figure 6A If an item is selected in the selection screen shown, and the image recognition unit 30 is deemed to have low recognition accuracy, the motion control unit 40 proceeds to step S14, whereby the image recognition unit 30 performs machine learning. That is, the motion control unit 40 provides the image recognition unit 30 with the image data of the item on the measuring plate 7A and the selected item name "salmon nigiri sushi" as training data.

[0061] The same applies when specifying the "salmon nigiri sushi" item by inputting the call number "0012". Therefore, the image recognition unit 30 causes the model generation and storage unit 33 to perform machine learning and store the updated learned model (steps S15 and S16). If this machine learning is repeated every time product information is obtained, the recognition accuracy of the image recognition unit 30 can be improved. Furthermore, when the image recognition unit 30 inputs image data of "salmon nigiri sushi" placed on the measuring plate 7A, if only "salmon nigiri sushi" is displayed instead of multiple product candidates, the touch... Figure 6B The product information screen SC1 is located in the learning settings section B1. Therefore, the "Execute" learning option is displayed in the learning settings section B1 as follows: Figure 8 The display shows a switch to "Do not perform" learning; after that, machine learning will no longer be performed. Figure 9 Example of a portion of the stored content of the product master table, showing the following: For some products, such as "nigiri sushi" and "sushi platter", after repeated machine learning, as products that do not require machine learning, the "execute" learning at the learning setting section B1 has been switched to "do not execute" learning.

[0062] In this way, if the number of products for which "non-learning" is not performed at the learning setting unit B1 increases, the image recognition unit 30 can immediately assign a matching product name based on the image data of the product on the measuring plate 7A. Therefore, in the measuring label issuing device 1 of this embodiment, the recognition accuracy of the image recognition unit (image recognition device) 30 can be improved while performing the pricing operation.

[0063] In the metering label issuing device 1 of the above embodiment, the voice output device 15 will announce the product name displayed on the display unit 5, regardless of whether the product is specified by the image recognition unit 30 or called by the operator. This allows the operator to confirm whether the specified product is correct.

[0064] In the metering label issuing device 1 of the above embodiment, when the metering value of the product placed on the metering unit 7 is stable, the imaging unit 8 acquires image data of the product. Therefore, the image recognition unit 30 can designate a product based on the image data acquired when the metering value is stable, i.e., when the product placed on the metering plate 7A of the metering unit 7 is in a stable state. This improves the recognition accuracy of the image recognition unit 30 when designating a product.

[0065] The above description describes one embodiment, but one aspect of the present invention is not limited to the above embodiment. Various modifications can be made without departing from the spirit of the invention.

[0066] In the above embodiments, one aspect of the present invention is applied to, for example... Figure 1 The example of metering tag issuing device 1 shown is illustrated, but it can also be applied to devices such as... Figure 10A The metering, packaging, and pricing device (product handling device) 101 shown measures and packages the goods, then issues and affixes a product label LC containing printed product information to the goods. The metering, packaging, and pricing device 101 includes a metering unit 102, a packaging unit 103, a printing unit 104, an affixing unit 105, a display and operation unit 106 consisting of a touch panel, a speaker 107, a controller (control unit) 110, and a storage unit 112. In a modified example, the speaker 107 and the controller 110 constitute a voice output device 115.

[0067] Furthermore, one aspect of the present invention can also be applied to, for example... Figure 10B The illustrated electronic scale 201 is a face-to-face type with a printer, featuring a staff-side operating unit 205A and a customer-side operating unit 205B. The electronic scale 201 includes a main housing 203, staff-side operating units 205A and 205B (both composed of touch panels), a speaker 206, a measuring unit 207 with a measuring disc 207A, a label issuing unit 209 for issuing product labels LC, a controller (control unit) 210, and a storage unit 212. In a modified example, the speaker 206 and controller 210 constitute a voice output device 215. Alternatively, it can be configured as a product handling device without a label issuing device 207. Furthermore, in this electronic scale 201 and label issuing device, the structure is such that the measuring disc 207A serves as a product placement unit, with a camera positioned above it.

Claims

1. A commodity processing device, characterized in that, have: The control unit retrieves product information from the master product table, either through a product recall operation or by an image recognition device, and displays it on the display unit. When a call operation is performed without specifying the product, the control unit activates the image recognition device.

2. The commodity processing apparatus according to claim 1, characterized in that, It also has: A voice output device that, when the display unit displays product information related to the product, at least announces the product name contained in the product information.

3. The commodity processing apparatus according to claim 1 or 2, characterized in that, When there are multiple products specified by the image recognition device, the control unit causes the display unit to display the multiple specified products. When a product is selected from the multiple displayed products, the display unit displays the product information of the selected product.

4. The commodity processing apparatus according to claim 1 or 2, characterized in that, It also has: The measurement department measures the goods; and The label issuing department prints the weight of the goods measured by the measurement department and the goods information obtained from the goods master table on the labels and issues them.

5. The commodity processing apparatus according to claim 1 or 2, characterized in that, It also has: The photography department, which photographs the products mentioned, The control unit provides the image data of the product captured by the camera unit, as well as the product name of the product specified by the product's calling operation or specified by the image recognition device, as training data to the image recognition device and enables it to perform machine learning.

Images