Game management system

Abstract

The present invention aims at improving the prior art to manage games in a casino. A game management system (100) is provided with: a camera (11) and a camera (12) that shoot a table top (72) of a game table for a specific card game from an oblique upper side and generate images; and a card detection device (301) that analyzes the images using artificial intelligence and detects game cards by detecting numbers printed on the game cards placed at any position on the table top (72) to indicate the value of the cards, and the angle of view and size are not fixed.

Description

[0001] Divisional application This application is a divisional application of patent application number 202610587144.4, filed on April 29, 2026, entitled "Game Management System". Technical Field

[0002] This invention relates to a game management system for managing games in casinos. Background Technology

[0003] In casinos, game tables, game cards, game coins, and other casino-related items are used to simulate games. From the perspectives of casino operation analysis, detecting cheating, ensuring smooth game operation, and providing convenience for players, digitizing game information using these physical gaming items is effective.

[0004] For example, for a specific card game played on a game table using game cards and coins, by photographing the game cards and coins on the game table and analyzing the images, the placement amount and hand of each player can be identified, thereby enabling player customer rating or detecting cheating behavior by players or dealers (e.g., Patent Document 1).

[0005] Existing technical documents Patent documents Patent Document 1: U.S. Patent No. 11660541. Summary of the Invention

[0006] The problem that the invention aims to solve The purpose of this invention is to improve upon existing technology for managing games in casinos.

[0007] means for solving problems One type of game management system has the following structure: the game management system includes: a camera that captures an image of the game table from an obliquely upward position and generates an image; and a card detection device that uses artificial intelligence to analyze the image and detects cards by detecting numbers, which are numbers representing the face value of cards printed on cards placed at any position on the table, and the angle and size of the numbers being observed are not fixed.

[0008] In the aforementioned game management system, the game table can be a game table for a specific card game, and the cards can be cards that constitute the hand of a player or the dealer.

[0009] In the aforementioned game management system, the card detection device can detect the card by detecting the pairs of numbers.

[0010] In the aforementioned game management system, the card detection device can detect the card by detecting the combination of the numbers in a specified positional relationship with the markers indicating the suit.

[0011] The aforementioned game management system may also include: a card face value recognition device, which uses artificial intelligence to analyze the image to identify the face value of the cards detected by the card detection device; and a hand recognition device, which identifies a deck of cards consisting of multiple cards detected by the card detection device, and identifies the hand of each deck based on the face value identified by the card face value recognition device.

[0012] The aforementioned game management system may also include an orientation determination device, which determines the orientation of the card based on the orientation of the card face value and / or suit detected by the card detection device, or the arrangement direction of the card face value and suit.

[0013] The aforementioned game management system may also include a position determination device, which determines the position of the card based on the card face value and suit position detected by the card detection device.

[0014] Another aspect of the game management system of the present invention has the following structure: the game management system includes: a camera that captures an image of the tabletop of a specific card game table from an oblique angle and generates an image; a hand card recognition device that uses artificial intelligence to analyze the image to identify the hand cards; a placement recognition device that uses artificial intelligence to analyze the image to identify the placement of game coins; and a play recognition device that, based on the hand cards identified by the hand card recognition device and the placements identified by the placement recognition device, identifies play content according to the game rules. The play recognition device is capable of recognizing splitting cards and doubling placements as player actions. The play recognition device is configured to: firstly determine whether a player's hand card has been split into two hands based on multiple images captured in a time sequence; secondly determine whether a player's placement has increased the amount based on the multiple images captured in a time sequence; and, based on the results of the first and second determinations, identify whether splitting cards and doubling placements have been performed.

[0015] In the aforementioned game management system, when the first determination determines that a player's hand of cards has been divided into two hands, and the second determination determines that the player's placement has increased the amount, the play recognition device can identify that the card division has been performed.

[0016] In the aforementioned game management system, when the first determination determines that a player's hand of cards has not been divided into two hands, and the second determination determines that the player's placement has increased in value, the play recognition device identifies that the double placement has been performed.

[0017] The aforementioned game management system may also include a gesture recognition device, which uses artificial intelligence to analyze the image to recognize gestures. The play recognition device may also identify whether the card splitting has been performed based on the gestures detected by the gesture recognition device.

[0018] The aforementioned game management system may also include a gesture recognition device, which uses artificial intelligence to analyze the image to recognize gestures. The play recognition device also identifies whether the double placement has been performed based on the gestures recognized by the gesture recognition device.

[0019] In the aforementioned game management system, the play recognition device can identify that the two hands belong to the same player when the cards have been dealt.

[0020] In the aforementioned game management system, the game table for the specific card game can have multiple game positions, and the hand card recognition device corresponds the two hand cards to the same game position.

[0021] The aforementioned game management system may also include a player identification device that identifies players participating in the game. The hand card identification device corresponds the hand cards to the players identified by the player identification device, and the hand card identification device corresponds the two hand cards to the same player.

[0022] In the aforementioned game management system, the hand recognition device can determine whether a hand can be divided according to the rules.

[0023] In the aforementioned game management system, the play recognition device can determine whether the player's placement has been multiplied, as the second determination.

[0024] Another aspect of the game management system of the present invention has the following structure: the game management system includes: a camera that captures an image of the tabletop of a game table for a specific card game; a hand card recognition device that analyzes the image to detect the face value and suit of the cards, thereby identifying the hand formed by the cards dealt to the tabletop; a placement recognition device that identifies the placement of game coins on the tabletop; and a play recognition device that determines whether the specific card game being played on the tabletop is being conducted according to the rules based on the hand card recognized by the hand card recognition device and the placement recognized by the placement recognition device. The play recognition device is configured such that, when the hand card recognition device identifies that the dealer's face-up card is an Ace, it determines that the hand card identified by the hand card recognition device has been placed safely based on the placement identified by the placement recognition device in sequence; when it is determined that a safe placement has been placed, and the hand card recognition device identifies that the dealer's hand meets a specific point value, the safe placement is determined to be successful.

[0025] Another aspect of the game management system of the present invention has the following structure: the game management system includes: a camera that captures an image of the tabletop of a game table for a specific card game; a hand card recognition device that analyzes the image to detect the face value and suit of the cards, thereby identifying the hand formed by the cards dealt to the tabletop; a placement recognition device that identifies the placement of game coins on the tabletop; and a play recognition device that, based on the hand card recognized by the hand card recognition device and the placement recognized by the placement recognition device, determines whether the specific card game being played on the tabletop is being conducted according to the rules. The play recognition device is configured to, based on the placements identified by the placement recognition device in a time sequence, determine that the placements in the game have been reduced; when it is determined that the reduction has occurred, determine whether it is a situation where surrender is possible based on the rules, and, based on the placements identified by the placement recognition device in a time sequence, determine whether the reduction is reduced to half of the original placement value, thereby determining that the reduction is a surrender according to the rules.

[0026] The aforementioned game management system may also include a gesture recognition device, which uses artificial intelligence to analyze the image to recognize gestures. The game recognition device recognizes the surrender based on the gestures recognized by the gesture recognition device. Attached Figure Description

[0027] Figure 1 This is an appearance diagram of the game management system according to an embodiment of the present invention.

[0028] Figure 2 This is a block diagram illustrating the structure of the game management system according to an embodiment of the present invention.

[0029] Figure 3This is a block diagram illustrating the structure of the game management system according to an embodiment of the present invention.

[0030] Figure 4 This is a diagram illustrating an example of an image captured by a camera according to an embodiment of the present invention.

[0031] Figure 5 This is a diagram illustrating an example of a game card used in an embodiment of the present invention.

[0032] Figure 6 This is a diagram illustrating the image analysis performed by the card detection device according to an embodiment of the present invention.

[0033] Figure 7 This is a diagram illustrating an example of the game card configuration according to an embodiment of the present invention.

[0034] Figure 8 This is a diagram illustrating an example of the game card configuration according to an embodiment of the present invention.

[0035] Figure 9 This is a diagram illustrating an example of a gesture (asking for a sign) according to an embodiment of the present invention.

[0036] Figure 10 This is a diagram illustrating an example of a gesture (asking for a sign) according to an embodiment of the present invention.

[0037] Figure 11 This is a diagram illustrating an example of a gesture (stop sign) according to an embodiment of the present invention.

[0038] Figure 12 This is a diagram illustrating an example of a gesture (card) according to an embodiment of the present invention.

[0039] Figure 13 This is a diagram illustrating an example of a gesture (double placement) according to an embodiment of the present invention.

[0040] Figure 14 This is a diagram illustrating an example of a gesture (surrender) according to an embodiment of the present invention.

[0041] Figure 15 This is a diagram illustrating an example of an image captured by a camera according to an embodiment of the present invention.

[0042] Figure 16 This indicates that the player identification device in this embodiment of the invention... Figure 15 The image shows the result of player identification after the image is processed.

[0043] Figure 17 This is a diagram illustrating an example of the action (card dealing) in the BJ card game according to an embodiment of the present invention.

[0044] Figure 18This is a diagram illustrating an example of the action (card dealing) in the BJ card game according to an embodiment of the present invention.

[0045] Figure 19 This is a diagram illustrating an example of an action (double placement) in the BJ card game according to an embodiment of the present invention.

[0046] Figure 20 This is a diagram illustrating an example of an action (double placement) in the BJ card game according to an embodiment of the present invention.

[0047] Figure 21 This is a diagram illustrating an example of an action (double placement) in the BJ card game according to an embodiment of the present invention.

[0048] Figure 22 This is a diagram illustrating an example of an action (insurance) in the BJ card game according to an embodiment of the present invention.

[0049] Figure 23 This is a diagram illustrating an example of an action (equal payout) in the BJ card game according to an embodiment of the present invention.

[0050] Figure 24 This is a diagram illustrating an example of an action (equal payout) in the BJ card game according to an embodiment of the present invention.

[0051] Figure 25 This is a diagram illustrating an example of the actions in the BJ card game according to an embodiment of the present invention. Detailed Implementation

[0052] The following describes a method for implementing the invention; however, the embodiments described below are merely examples, and the invention is not limited to the embodiments described below. Furthermore, the elements of the various embodiments can be combined arbitrarily.

[0053] The following describes an example of applying the game management system of the present invention to the management of a specific card game such as BJ Card Game, but the present invention can be applied to other card games and other casino games.

[0054] Figure 1 This is an external view of the game management system according to an embodiment of the present invention. The game management system 100 is formed on a game table 7 for BJ card games. A game coin tray 71 is provided on the tabletop 72 of the game table 7. The game coin tray 71 is used to replenish the dealer's game coins. The dealer sits in front of the game coin tray 71 to deal cards in the game.

[0055] In Game Table 7, the side opposite the dealer's position is curved, allowing multiple players to sit along the edge of this curve to play BJ Cards. Game Table 7 provides multiple playing positions for multiple players.

[0056] Cameras 11 and 12 are positioned on either side of the dealer at game table 7. These cameras 11 and 12 are used to film the tabletop of game table 7 from an angle above. Cameras 11 and 12 film not only the game cards and coins on the table but also the players. It should be noted that cameras 11 and 12 can each have multiple cameras with different shooting directions, angles, and resolutions. In this case, each of these multiple cameras can be configured to be suitable for filming game cards, suitable for filming game coins, and suitable for filming players.

[0057] Figure 2 This diagram illustrates an example of a desktop layout according to an embodiment of the present invention. The desktop 72 is decorated with multiple placement areas 77 corresponding to multiple player positions for playing a BJ card game. Figure 2 In the example, seven placement areas 77 are provided corresponding to the seven player positions. Each placement area 77 is provided with a side placement area 78. For example, the side placement area can be an area for placing pairs. In addition, a safety placement area 76 for placing safety is drawn further to the dealer's side than these placement areas 77 and the side placement areas 78. The safety placement area 76 is set as a long and narrow placement area that spans multiple player positions.

[0058] A coin tray 71 is centrally located at the dealer's position. An electronic shoe 51 is placed on the tabletop 72 to the right of the dealer's position (left side when viewed from the dealer's perspective). A game card storage box 74 is located on the tabletop 72 to the left of the dealer's position (right side when viewed from the dealer's perspective), used to pre-store game cards to be used in the game. A display panel 61 is located next to the coin tray 71, displaying various information to assist the dealer's operations.

[0059] Figure 3 This is a block diagram illustrating the structure of a game management system 100 according to an embodiment of the present invention. The game management system 100 includes cameras 11 and 12, RFID readers 21 and 22, a game management device 30, a player database 41, a game coin database 42, a game history database 43, an electronic shoe 51, and a display panel 61.

[0060] In this embodiment, an RFID tag is embedded in the game coin. The RFID tag stores a game coin ID, which is used to uniquely identify the game coin.

[0061] RFID reader 21 reads the RFID tags of game coins stored in the game coin tray 71. RFID reader 22 reads the RFID tags of game coins placed in the placement area. RFID reader 21 and RFID reader 22 each include: one or more RFID antennas; and a reader that drives these antennas and acquires the reading signals from the antennas.

[0062] The RFID reader 22 can have an RFID antenna in each of the seven placement areas 77, and can read the game tokens placed in each placement area 77. Interference antennas can be provided corresponding to each RFID antenna to limit the RFID tags of game tokens in each placement area from being read by RFID antennas in adjacent placement areas. Alternatively, interference antennas can be omitted, and the location of the game tokens can be determined based on the reading strength of multiple RFID antennas, assuming that the game tokens in a certain placement area are read by multiple adjacent RFID antennas.

[0063] The RFID reader 22 can sequentially drive multiple RFID antennas and synchronously drive interference antennas around the placement area 77 of the RFID antenna being read, thereby reading the game coins in each placement area 77 in a time-difference manner.

[0064] Cameras 11 and 12 capture images of the tabletop 72 of the game table 7 from an oblique angle and generate images. Furthermore, as described above, cameras 11 and 12 also capture images of the players playing games at the game table 7.

[0065] Figure 4 This diagram illustrates an example of an image captured by a camera according to an embodiment of the present invention. The image captures game coins placed in a designated area for playing the BJ card game, game cards distributed to each player, and the hands of the players and the dealer. Cameras 11 and 12 generate images by capturing images in response to predetermined trigger conditions and output them to the game management device 30. Alternatively, cameras 11 and 12 may also generate images by capturing images at predetermined intervals and output them to the game management device 30 as needed.

[0066] return Figure 3 The electronic shoe 51 stores multiple decks of game cards. The dealer draws game cards one by one from the electronic shoe 51, starting with the frontmost card. The electronic shoe 51 reads the drawn game cards to identify their face value and suit.

[0067] Figure 5This diagram illustrates an example of a game card according to an embodiment of the present invention. The numbers (1-10, J, Q, K) representing the card's face value and the symbols representing the suit (spades, clubs, hearts, diamonds) are arranged vertically in the upper left and lower right corners of the game card. It should be noted that while the face value of the game card is represented by the numbers 1-10 and the letters J, Q, and K, in this specification, J, Q, and K, representing the face value, are also referred to as "numbers" along with the numbers 1-10.

[0068] In addition, the short edges of the game cards are printed with codes indicating the card's face value and suit. These codes are printed with UV ink, visible only under ultraviolet light. A number corresponding to the face value is printed in the center of the game card for each suit. Face cards (J, Q, K) are represented by corresponding symbols.

[0069] The electronic shoe 51 is equipped with an ultraviolet lamp and a sensor. By irradiating ultraviolet light onto the edge of the short side of the drawn game card and detecting the code by the sensor, it identifies the face value and suit of the drawn game card and outputs the identification result to the game management device 30. The sensor outputs an open / close signal corresponding to the code that moves in front of the sensor as the game card is drawn. Based on the open / close signal output by the sensor, the electronic shoe 51 identifies the face value and suit represented by the code and outputs the identification result to the game management device 30.

[0070] Furthermore, if the electronic shoe 51 detects that a game card has been drawn, it will output a signal indicating this to the game management device 30. If the game management device 30 receives this signal, it will instruct the cameras 11 and 12 to take pictures.

[0071] The game management device 30 includes multiple devices for managing games. The game management device 30 is composed of a computer equipped with a processing unit, a memory, a storage device, etc. The devices included in the game management device 30 can be configured such that the processing unit reads a computer program stored in the storage device into the memory and executes the program.

[0072] The card detection device 301 uses artificial intelligence to analyze images and detects cards by detecting the numbers representing the card values ​​printed on game cards placed at any position on the table 72. Depending on the position of the game card on the table 72, the angle at which the numbers printed on the game card are viewed is not fixed, and their size within the image is also not fixed.

[0073] Figure 6This diagram illustrates the image analysis performed by the card detection device according to an embodiment of the present invention. Game cards closer to cameras 11 and 12 have a larger elevation angle when viewed from the cameras and are thus displayed larger in the image. Conversely, game cards farther from cameras 11 and 12 have a smaller elevation angle when viewed from the cameras and are thus displayed smaller in the image. The card detection device 301 identifies numbers by performing image analysis on the images. Such number recognition can, for example, utilize an inference model using a neural network, or a machine learning model such as SVM.

[0074] As mentioned above, the numbers printed on game cards are presented in an unpredictable manner due to the positional relationship between the camera and the game card. Therefore, it is difficult to detect the numbers through simple template matching, and more advanced image recognition processing, such as feature extraction and matching, is required. In this application, the technology for performing such advanced image recognition is referred to as artificial intelligence. The card detection device 301 uses this artificial intelligence to analyze the image and detect the numbers on game cards that are arranged in any position on the table 72 with any orientation.

[0075] As existing technology, there are known techniques for identifying a player's hand by detecting the game cards themselves. However, the card detection device 301 in this embodiment does not detect the game cards themselves, but rather detects the numbers of face value and the symbols of suit printed on them. Thus, as Figure 4 As shown, even if the game cards are stacked on the table 72 according to the BJ card game rules, or are obscured by the dealer's hand and cannot be observed, the card detection device 301 of this embodiment can detect the game cards simply by finding the combination of card value and suit printed in the upper left and lower right corners.

[0076] It should be noted that in this embodiment, the card detection device 301 detects game cards by searching for combinations of numbers (face value) and symbols (suits) printed in the upper left and lower right corners of the game cards. However, alternatively, the card detection device 301 can also detect game cards by detecting only the numbers (face value) printed in the upper left and lower right corners of the game cards, or by detecting only the symbols (suits) printed in the upper left and lower right corners of the game cards.

[0077] Furthermore, the card detection device 301 can also detect game cards by detecting pairs of numbers representing the card's face value and symbols representing the suit. As described above, the upper left and lower right corners of the game cards are printed with numbers representing the card's face value and symbols representing the suit, respectively. Therefore, when the card detection device 301 detects these combinations, it can determine that a game card is present. In this case, the card detection device 301 can detect game cards by detecting pairs of numbers representing the card's face value, or it can detect game cards by detecting pairs of symbols representing the suit.

[0078] Furthermore, when the detected numbers and symbols are in a predetermined positional relationship—that is, when they are arranged vertically with equal size—the card detection device 301 can determine that the numbers and symbols represent the face value and suit of a game card, and thus determine that a game card is present. Therefore, even if a number or symbol (spades, clubs, hearts, diamonds) that is not used to represent the face value of a game card is detected in the image, for example, if numbers or symbols are detected on a player's clothing, the card detection device 301 will not determine the presence of a game card as long as they are not in the predetermined positional relationship.

[0079] The face value recognition device 302 uses artificial intelligence to analyze images and identify the face value of the cards detected by the card detection device 301. The face value recognition device 302 uses the image recognition results of the card detection device 301 to identify the face value. If the card detection device 301 detects a number, the face value recognition device 302 performs image recognition again on the detected area to identify the content of that number, i.e., the face value. It should be noted that the face value recognition device 302 can also similarly identify suits.

[0080] The position determination device 303 determines the position of a card based on the combination of numbers (card face value) and symbols (suits) detected by the card detection device 301. As described above, cameras 11 and 12 receive instructions from the game management device 30 to take pictures when a game card is drawn from the electronic shoe 51. Once a game card is drawn from the electronic shoe 51, it is transported to the distribution position by the dealer. This movement is completed when the next game card is drawn from the electronic shoe 51. Therefore, by taking pictures at the moment a game card is drawn from the electronic shoe 51, the movement of the previous game card has been completed, and cameras 11 and 12 can capture pictures of the game card after the movement is completed by photographing the tabletop 72 in this state.

[0081] The position determination device 303 identifies newly distributed game cards on the table 72 and determines their positions by comparing the latest image with its previous image. The position of the game card is represented in table coordinates by converting the coordinates within the image to coordinates on the table 72. Thus, even if cameras 11 and 12 detect the same game card, these game cards will be integrated using the coordinates on the table 72.

[0082] The position determination device 303 can determine the position of a game card based on pairs of numbers and symbols detected by the card detection device 301. That is, if two pairs of the same number and symbol are detected, it can be determined that there are game cards that are diagonally opposite to them.

[0083] The orientation determination device 304 determines the orientation of the game cards dealt to the table 72 based on the orientation of the numbers (face value) and symbols (suits) detected by the card detection device 301. Alternatively, the orientation determination device 304 can also determine the orientation of the game cards based on the arrangement direction between the numbers (face value) and symbols (suits) detected by the card detection device 301. Figure 5 As shown, in the top left corner of the game card, numbers and symbols are arranged vertically with the numbers (card value) on top and the symbols (suit) on the bottom. Therefore, by determining the arrangement of the numbers and symbols, the orientation of the card can be determined.

[0084] Alternatively, the orientation determination device 304 can determine the orientation of the game card based on the position of the pairs of numbers and symbols detected by the card detection device 301. That is, when the card detection device 301 detects two pairs of numbers and symbols, as described above, the diagonal of the game card can be determined, and therefore, the orientation determination device 304 determines the orientation of the game card based on the direction of that diagonal.

[0085] It should be noted that the position determination device 303 can also determine the card's position based on the orientation determination device 304's determination of the orientation. That is, as... Figure 5 As shown, the game cards have combinations of numbers and symbols printed in the upper left and lower right corners. Therefore, in addition to the position of the combinations of numbers and symbols, the orientation of the game cards can also be considered to determine where the game cards are located on the table 72.

[0086] The hand recognition device 305 recognizes a deck of cards composed of multiple game cards detected by the card detection device 301, and identifies the hand of each deck based on the face value of the multiple game cards identified by the face value recognition device 302. That is, the hand of a player or dealer is usually determined by the combination of the face values ​​of the multiple game cards dealt to that person.

[0087] The hand card recognition device 305 groups multiple game cards based on the position and orientation of the game cards identified by the position determination device 303 and the orientation determination device 304. As described above, whenever a new game card is drawn from the electronic shoe 51, the cameras 11 and 12 capture images and generate images. Therefore, in the images obtained in this way, the game cards distributed to the table 72 increase one by one.

[0088] The hand recognition device 305 identifies the hand cards in relation to the player's position. That is, the hand recognition device 305 identifies which player's position the identified hand cards belong to. Alternatively, the hand recognition device 305 may also identify the hand cards in relation to the placement area 77 set for each player's position. Or, the hand recognition device 305 may also identify the hand cards in relation to the placement identified by the placement recognition device described later and / or the player identified by the player recognition device described later.

[0089] Figure 7 and Figure 8 This is a diagram illustrating an example of the configuration of game cards according to an embodiment of the present invention. Figure 7 and Figure 8 This represents a player's hand. The first card dealt is the 3 of Diamonds, followed by the Ace of Hearts. The second Ace of Hearts is positioned so that its lower left portion covers the upper right portion of the first 3 of Diamonds, indicating that these two cards constitute one hand. Furthermore, the entire hand containing the later-drawn Ace of Hearts is observed by cameras 11 and 12. In the BJ card game, the dealer manipulates the cards; players do not touch them, hence this arrangement is implemented by the dealer.

[0090] like Figure 8 As shown, the 9 of Spades is added to the hand as the third game card. The third game card is positioned such that its lower left portion overlaps the upper right portion of the second game card. Thus, a hand consists of a deck of three game cards: the first game card, the second game card overlapping the first game card, and the third game card overlapping the second game card.

[0091] Each game card displays its face value and suit combination in the top left and bottom right corners. Therefore, as... Figure 8 As shown, if the game cards are arranged sequentially by overlapping the lower left part of the new game card with the upper right part of the previous game card, then for any game card, the combination of the face value and suit in the upper left and lower right corners can be observed by cameras 11 and 12.

[0092] In this way, the hand card recognition device 305 can know the order in which the cards are dealt. Furthermore, even if the later drawn game card is stacked on top of the earlier drawn game card and dealt to the table, and a portion of the earlier drawn game card is obscured by the later drawn game card placed on the table 72, the face value, suit, position, and orientation of the earlier drawn game card can still be recognized in the image before the next game card is drawn.

[0093] The hand recognition device 305 determines whether multiple game cards that partially overlap each other constitute a hand by grouping them together. In the BJ card game, according to the operating rules, the multiple game cards that constitute a hand are arranged on the table 72 by the dealer in a manner that partially overlaps each other. Therefore, it can be determined that multiple game cards that partially overlap each other constitute a hand.

[0094] It should be noted that in other card games, the configuration of multiple game cards that make up a hand is determined by the game's operating rules. Therefore, the hand recognition device 305 identifies each hand based on the face value, suit, position, and orientation of the game cards according to these operating rules.

[0095] The placement recognition device 306 identifies game coins placed by the player in placement area 77 based on images from cameras 11 and 12. The placement recognition device 306 identifies which placement area 77 contains how many game coins in total. The placement recognition device 306 also identifies, based on images, game coins added during gameplay (such as splitting, doubling, insurance, etc., described later) and game coins paid from the dealer's game coin tray to the winning placement.

[0096] The placement recognition device 306 uses artificial intelligence to analyze images from cameras 11 and 12 to identify stacks of placed game coins and each individual coin within a stack. The same method applies to adding or paying in the game, identifying stacks of game coins and the individual coins within them. Of the multiple game coins in a stack, only the side view is visible, except for the top coin. The type (value) and even denomination of a game coin can be determined by the side pattern. Therefore, the placement recognition device 306 identifies the type of game coin by analyzing the side patterns of the multiple game coins in a stack.

[0097] The gesture recognition device 307 uses artificial intelligence to analyze the images from camera 11 and camera 12 to recognize the gestures made by the player or dealer. Figures 9-14 This is a diagram illustrating an example of a gesture according to an embodiment of the present invention. Figures 9-14 The gesture is an example of a gesture from the BJ card game.

[0098] Figure 9 and Figure 10 The gesture indicating "wanting a card" is shown. (For example...) Figure 9 The gesture shown is tapping the table repeatedly with an extended five-fingered hand, and as shown in the image. Figure 10 The gesture shown is to extend the index and middle fingers together and tap the table repeatedly 72 to indicate that you want a card. Figure 11 The gesture indicating "suspend trading" is shown. (For example...) Figure 11 As shown, a hand gesture of waving the five fingers out on the table 72 indicates a stop.

[0099] Figure 12 The gesture indicating "splitting" is shown. (For example...) Figure 12 As shown, extending the index and middle fingers together to point at the cards in the hand, and then separating the index and middle fingers, indicates a split. It should be noted that the gesture recognition device 307 can also determine a split solely by separating the index and middle fingers. Furthermore, the gesture recognition device 307 can also recognize any two fingers, such as the index and little fingers, extended, as a split gesture.

[0100] Figure 13 The gesture indicating "double placement" is shown. (For example...) Figure 13 As shown, pointing with the index finger to the newly placed game coin indicates doubling. Figure 14 The gesture indicating "surrender" is shown. (For example...) Figure 14 As shown, extending the index finger and placing it on the table 72 and sliding it horizontally indicates surrender.

[0101] Cameras 11 and 12 acquire multiple images in a time sequence through continuous shooting. The gesture recognition device 307 identifies the shape of the hand, changes in the shape of the hand, and hand movements from these multiple images, thereby recognizing the gesture.

[0102] Player identification device 308 identifies players playing games at game table 7. Player identification device 308 identifies players at multiple levels. It identifies players by analyzing images captured by cameras 11 and 12 using artificial intelligence. Player identification includes facial recognition and pose recognition. Facial recognition includes various levels of functions such as detection, tracking, analysis, verification, association, and identification. By identifying and detecting players through facial recognition, it is possible to determine the location and number of players.

[0103] By utilizing facial recognition technology from artificial intelligence to track players, it's possible to analyze a player's placement activities across multiple games, even if individual players cannot be identified. Furthermore, by analyzing players' facial expressions using AI, it's possible to infer their mood. Combining this mood estimation with game preferences (e.g., placement amount, card splitting, doubling down, and side-placing strategies) can be beneficial for casino operations and marketing.

[0104] Furthermore, by utilizing artificial intelligence for facial authentication, comparing a player's face with a registered face, it is possible to verify whether the player is authorized to be present. Further, by associating the recognized face with the player's skeletal model, it is possible to link the recognized player's face with that player's activities (e.g., placement activities). Moreover, by pre-registering information used to identify players in association with their facial images, it is possible to identify which registered player the player is.

[0105] Even when prioritizing player privacy, the aforementioned objectives can be achieved without identifying players through facial recognition. For example, for players not participating in a loyalty program, the recognized face can be associated with and stored in conjunction with temporarily assigned loyalty points. Later, when the player joins the loyalty program, the loyalty points previously associated with the face can be activated.

[0106] Furthermore, even if the player is uncertain, by pre-storing the player's facial image obtained from the image in the player database 41 or the game currency database 42 in association with the game currency ID held by that player, when the game currency ID is read from the RFID tag of the game currency at the game table 7 or counter, the facial image of the game currency owner can also be obtained. It can then be determined whether the combination of the read game currency ID and the game currency owner's facial image matches the combination stored in the player database 41 or the game currency database 42, thereby determining whether the person is the correct owner of the game currency. This enables measures to prevent improper fund flows.

[0107] If the owner associated with the game currency does not match the player who actually placed it, the player identification device 308 can issue a warning, prompting the dealer to take corrective action. This warning can be displayed on the display panel 61. Corrective action could include requiring identification documents or prohibiting further participation in the game.

[0108] Furthermore, the player identification device 308 can record the time and place of discovering such inconsistencies, as well as information about the players or game currency involved, in the player database 41, the game currency database 42, and / or the game history database 43. Additionally, in the event of the aforementioned inconsistency, the player identification device 308 can consider it as a transfer of game currency, associate the player recorded as the owner with the player who actually placed the currency, and record the transfer between these players in the game history database 43.

[0109] Based on this record, the game management device 30 further analyzes the game history or activities outside of game facilities (e.g., hotel accommodations, dining) of related players using artificial intelligence to conduct marketing to related players (e.g., providing services to family members when related players are family members), or to detect cheating behaviors such as improper fund flows.

[0110] The aforementioned facial images can be stored as the images themselves and used for comparison and other processing, or they can be stored as facial feature quantities obtained through image recognition and used for comparison and other processing. By processing the facial feature quantities extracted from the facial images without processing the facial images themselves, player privacy can be protected.

[0111] Figure 15 This diagram illustrates an example of an image captured by a camera according to an embodiment of the present invention. The player recognition device 308 recognizes the facial image and the player's posture. In this image, in addition to the tabletop 72 and the game tokens placed on it, the player's upper body is also captured. The player recognition device 308 uses a neural network to perform facial recognition and posture recognition on the player within the image.

[0112] Figure 16 This indicates that the player identification device in this embodiment of the invention... Figure 15 The image shows the result of player recognition. The player recognition device 308 identifies key points of the player's body, including fingers, wrists, elbows, shoulders, nose, and eyes, and extracts the facial region, associating the extracted facial region with the identified key points as information of the same person.

[0113] The game coin operation recognition device 309 first searches for game coins in each of the time-series images continuously captured by cameras 11 and 12 at predetermined time intervals. By comparing the game coins recognized in past frames with the game coins recognized in the latest frame, the game coin operation recognition device 309 identifies that the game coins have changed, i.e., someone has operated the game coins.

[0114] For example, if a different game coin is detected in the latest frame compared to those detected in previous frames, the game coin operation recognition device 309 identifies it as a newly placed game coin on the table 72. Furthermore, if a game coin previously identified in a previous frame is not detected in the latest frame, the game coin operation recognition device 309 identifies it as having been removed from the table 72. Moreover, if a game coin previously identified in a previous frame is not detected in the same location in the latest frame, but is newly detected in a different location, the game coin operation recognition device 309 identifies it as having moved from its original location to the new location.

[0115] The game coin operation recognition device 309 also infers which finger operated which game coin stack based on the positional relationship between the game coin stack recognized by the placement recognition device 306 and the finger recognized by the player recognition device 308. Specifically, the placement recognition device 306 sequentially analyzes multiple continuously captured images to search for game coins. If a game coin stack is identified from an image, the player recognition device 308 performs player recognition on that image.

[0116] The coin handling recognition device 309 associates a finger identified by the player recognition device 308 with the coin stack at the closest position to the stack identified by the placement recognition device 306. Through the player recognition device 308's identification of the player, the area from the finger to the face is associated and identified as one player. Therefore, the coin handling recognition device 309 can ultimately associate the coin stack with the facial image. This achieves the aforementioned tracking, analysis, verification, and association functions.

[0117] The play recognition device 310 identifies the play content according to the game rules based on the hand cards recognized by the hand card recognition device 305 and the placements recognized by the placement recognition device 306. To this end, the play recognition device first matches the hand cards with the placements. That is, the play recognition device 310 determines which placement is matched between the hand cards recognized by the hand card recognition device 305 and the placements recognized by the placement recognition device 306.

[0118] The play recognition device 310 aligns the placement with the placement area 77 and the hand cards with the placement area 77, thereby matching the hand cards corresponding to the same placement area with the placement. Alternatively, the play recognition device 310 can also match the hand card closest to the placement with the placement based on the position of the hand card and the position of the placement.

[0119] When there are few other players at game table 7 and adjacent placement areas are open, a player may sometimes place more than two cards in a single game. In this case, the play recognition device 310 matches the player with their hand based on the correspondence between the placement and the player, and the correspondence between the placement and the hand. Thus, even if a player plays with multiple hands, the play content can be correctly matched with the player.

[0120] The play recognition device 310 determines the game result for each player based on the recognition results of each player's hand and the dealer's hand. In addition, the play recognition device 310 determines the winnings or losses of each player based on the placement amount of each player identified by the placement recognition device 306 and the determined game result.

[0121] The play detection device 310 stores game results and win / loss amounts for each game in the game history database 43. The game history records, linked to the game results, include the game's date and time, table number, and dealer information. Furthermore, the play detection device 310 stores each player's betting amounts and win / loss amounts in the player database 41. By utilizing artificial intelligence to statistically analyze the game history records, the play detection device 310 can obtain information effective for marketing and detect cheating behavior. For example, the play detection device 310 can identify players suspected of cheating if their win rate is close to the theoretical value, but their game history records showing small betting amounts when losing and large betting amounts when winning are statistically significant.

[0122] As described above, the play recognition device 310 can identify the distribution order of distributed game cards by sequentially analyzing multiple images captured in a time sequence. That is, by comparing the game card recognition results in past frames with the game card recognition results in the latest frame, it can determine the game cards added in the latest frame. By repeating this process, the distribution order of distributed game cards can be identified. However, due to occlusion or other reasons, game cards may not be recognized in a certain frame. As a result, more than two new game cards may be recognized in the latest frame compared to previously recognized game cards.

[0123] In this situation, the play recognition device 310 makes a judgment according to the game rules and even the operation rules, thereby determining the order in which the two or more newly recognized cards should be distributed. For example, in the BJ card game, the operation is carried out according to the following rules: the first card is distributed face up to the player on the left and right from the dealer's perspective, and the dealer's first card is distributed face down. The second card is distributed face up to the player on the left and right from the dealer's perspective, and the dealer's second card is distributed face down. The dealer's first card is then placed face up, and the player waits for actions such as drawing cards, stopping cards, and distributing cards.

[0124] When the play recognition device 310 recognizes two or more new game cards at once, it determines the distribution order of these newly recognized game cards based on the operation rules.

[0125] The play recognition device 310 identifies player actions such as splitting cards, doubling down, activating insurance, and surrendering based on the hand cards identified by the hand card recognition device 305 and the placement identified by the placement recognition device 306. Specifically, the play recognition device 310 determines whether a player's hand has been split into two hands based on multiple images captured sequentially by cameras 11 and 12.

[0126] Figures 17-25This is a diagram illustrating an example of the actions in the BJ card game according to an embodiment of the present invention. Figures 17-25 The actions shown are identified by the play recognition device 310 by analyzing multiple images captured in sequence by cameras 11 and 12.

[0127] Figure 17 This is an example diagram showing a player indicating the scoring cards. Figure 18 It means succession Figure 17 The diagram shows the state of the hand after the cards have been split. The play recognition device 310 recognizes that the cards have been split based on the following determinations: (1) it determines that the face value of the two game cards that make up the initial hand (the hand consisting of the first two game cards) is the same; (2) it determines that the gesture recognition device 307 has been instructed to split the cards by gesture; (3) it determines that the two stacked game cards have been divided into two groups; and (4) it determines that the placement amount has been doubled.

[0128] In the above example, the play recognition device 310 recognizes that a split has been made based on the determinations of (1) to (4), but it is not necessary to recognize that a split has been made based on all of these determinations of (1) to (4). For example, the play recognition device 310 may recognize that a split has been made based only on (3) and (4), or it may recognize that a split has been made based only on (1) and (3), or it may recognize that a split has been made based only on (2) or only on (3). In addition, the determination of (4) may be to determine that the amount placed has increased, or to determine that a stack of game coins has been added, instead of determining that the amount placed has doubled.

[0129] When cards have been dealt, the play recognition device 310 identifies two hands as belonging to the same player and assigns the player corresponding to the original initial hand to the two new hands. Furthermore, if the original initial hand corresponds to a player's position or a placement area set for each player's position, the two new hands are assigned to the player's position corresponding to the original initial hand or a placement area set for each player's position.

[0130] Furthermore, after the cards have been split, the play recognition device 310 associates the two hand cards with the original stack of game coins and the newly added stack of game coins, respectively.

[0131] It should be noted that the play recognition device 310 makes an error or even a mistake in its judgment operation when (1) is negative and (3) is positive. The situation where (1) is negative refers to a situation where splitting cards is not allowed according to the rules. In this case, the situation where (3) is positive is not allowed according to the rules. Therefore, the play recognition device 310 makes an error or even a mistake in its judgment operation.

[0132] Figure 19 This is an example of a soft hand. Figure 20 Indicates succession Figure 19 The state after the player doubles the placement. The play recognition device 310 recognizes that double placement has been performed based on the following judgments: (3) it is determined that the two stacked game cards are not divided into two groups, (5) it is determined that the gesture recognition device 307 is instructed to double placement through gesture, and (6) it is determined that the placement amount is doubled.

[0133] The play recognition device 310 can recognize double placement based on the situation that the judgment of (3) is negative and the judgment of (6) is positive, without relying on the judgment of (5) gesture.

[0134] Figure 21 It means succession Figure 20 The diagram shows the state after the third game card is dealt. The third game card is arranged to overlap the upper right of the second card, but is placed horizontally. Therefore, the play recognition device 310 can determine that the third game card is placed horizontally by (7), thereby recognizing that double placement has occurred. In this case, the play recognition device 310 can recognize double placement by adding the determination of (7) to the above (3), (5) and / or (6), or it can recognize double placement by relying solely on the determination of (7).

[0135] Figure 22 This is a diagram illustrating an example of safe placement. When (8) the hand card recognition device 305 recognizes the dealer's face-up card as A, and (9) the safe placement area 76 is identified as having been added based on the image sequentially recognized by the placement recognition device 306, the play recognition device 310 identifies the hand card recognized by the hand card recognition device 305 as having been safe placed. It should be noted that the play recognition device 310 may identify safe placement solely based on the determination in (9).

[0136] When the play recognition device 310 determines that a safe placement has been made, and the hand recognition device 305 identifies that the dealer's hand has a specific point value, i.e., BJ, the safe placement is considered successful. If the dealer's face-up card is an Ace, the dealer confirms whether the hand value reaches twenty-one, i.e., BJ, by checking the second game card. This confirmation can be done by the dealer lifting the edge of the second card so that the player cannot see its value, or by embedding a dedicated camera in the tabletop 72, allowing the camera to read the card value while the second game card is face down.

[0137] The confirmed result is that if the dealer's hand is BJ, the dealer immediately flips the second card face up. Therefore, the game card is photographed by cameras 11 and 12, and the image is analyzed by the hand recognition device 305. The play recognition device 310 can confirm whether the dealer's hand is BJ (whether the insurance placement was successful) based on the image analysis result.

[0138] Figure 23 This image shows an example where the player's hand is BJ and the dealer's face-up card is an Ace. In this situation, the player can declare equal payout. Because the player verbally declares equal payout, it cannot be detected through image recognition gestures, game cards, or placement.

[0139] like Figure 24 As shown, the play recognition device 310 identifies a payout as equal to the placement amount by recognizing an image before the dealer's second game card is placed face up.

[0140] Figure 25 This is a diagram illustrating an example of surrender. The play recognition device 310 determines, in (10) based on the placement identified sequentially by the placement recognition device 306, that a placement in the game has been deducted, and (11) when it is determined that the gesture recognition device 307 has indicated surrender via gesture, it recognizes that surrender has been made. It should be noted that the play recognition device 310 may recognize surrender based solely on either (10) or (11).

[0141] When the play recognition device 310 determines that the amount placed has been reduced by surrendering, it determines whether the situation is one where surrendering is possible based on the rules, and determines whether the reduction is to half of the original amount based on the placements identified by the placement recognition device 306 in a timely manner.

[0142] return Figure 3 The game state recognition device 311 recognizes the game state. Game state recognition includes the recognition of game phases. Game phases include the placement phase, the card distribution phase after placement, and the settlement phase where lost game coins are collected and won game coins are paid out after the game result is determined. Game phases may also include a pause phase where the game is paused due to no one participating, and a shuffling phase where the cards are shuffled in response to the cutting of the deck.

[0143] These game phases are determined by using image analysis to recognize the distribution of game cards on the table 72, the flipping of game cards from face down to face up, changes in game coins, gestures, etc. That is, the game state recognition device 311 identifies the game state based on the hand cards recognized by the hand card recognition device 305 and the placement recognized by the placement recognition device 306.

[0144] Furthermore, the game status recognition device 311 also identifies the exchange (color change) of game coins based on images. The game status recognition device 311 can determine that a color change has occurred when, during the non-clearing phase, some game coins disappear from the game coin tray and some are added to it, based on the reading results of the RFID reader 21 used to read the RFID tags of game coins in the game coin tray. The game status recognition device 311 can then determine whether the total amount has changed due to the color change based on the reading results of the RFID reader 21, thereby determining whether the color change was performed correctly.

[0145] In the game management system 100 of this embodiment, RFID antennas are provided between the game coin tray 71 and the placement area 77 to read the RFID tags of the game coins. However, in order to support the dealer's operation of buying and changing colors, a dealer area can also be set between the game coin tray 71, which serves as the game coin storage area, and the placement area 77, and a third RFID antenna different from the RFID antenna of the game coin tray 71 and the RFID antenna of the placement area 77 can be set in this area.

[0146] In game phases other than the settlement and card distribution phases, i.e., when purchases are possible, the game status recognition device 311 determines that a purchase is to be made when the RFID tag of game coins is read by the third RFID antenna, and activates the purchase processing function. During the purchase process, the game status recognition device 311 displays on the display panel 61 that a purchase will be made, and also displays the total amount of game coins read by the third RFID antenna on the display panel 61.

[0147] Subsequently, when the game status recognition device 311 reads game coins of a different type but the same amount as those previously read by the third RFID antenna, it determines that the transaction is not a purchase but a color change, and displays on the display panel 61 that a color change will be performed. The total amount of game coins read by the third RFID antenna before and the total amount of game coins read by the third RFID antenna after are also displayed on the display panel 61.

[0148] As described above, the game status recognition device 311 automatically determines that a color change will be performed and displays the amount related to the color change on the display panel 61. Therefore, the dealer can perform the color change process while checking the display panel 61.

[0149] It should be noted that color exchange is an equivalent exchange transaction, but if the amount related to the exchange is not equivalent in these transactions, the game status recognition device 311 displays a warning on the display panel 61 indicating that the equivalent exchange was not performed correctly.

[0150] Display panel 61 displays the results of the various processes performed by the game management device 30. For example, it can display the recognition results of each player's hand cards as recognized by the hand card recognition device 305, the recognition results of each player's placement as recognized by the placement recognition device 306, the recognition results of gestures as recognized by the gesture recognition device 307, the recognition results of players as recognized by the player recognition device 308, the actions as recognized by the play recognition device 310, and any other results recognized by the game management device 30.

[0151] In particular, the display panel 61 can display warnings when necessary, such as when the player identification device 308 identifies that the owner of the game coin is inconsistent with the player who placed the game coin, when the play identification device 310 identifies that various actions are not performed correctly, or when the settlement is not performed correctly.

[0152] (Variation Example 1) To improve game efficiency, the game management system 100 in Modification 1 automatically detects game stages. In the game management system 100, the game state recognition device 311 can identify game stages based on the detection of a target object by analyzing images captured by cameras 11 and 12. For example, the game state recognition device 311 can recognize the start of the placement stage based on the detection of game coins by the placement recognition device 306. Furthermore, the game state recognition device 311 can determine that the card distribution stage has begun based on the detection of game cards by the card detection device 301.

[0153] Player identification device 308 identifies the player's body, face, and hands from the image; placement identification device 306 identifies game coins; and card detection device 301 detects game cards. Game status identification device 311 recognizes the start of the placement phase when the initial image frame of the game coins is identified by placement identification device 306. Furthermore, game status identification device 311 recognizes the end of the placement phase and the start of the card distribution phase when the initial image frame of the game cards is detected by card detection device 301. Further, game status identification device 311 recognizes the end of the card distribution phase and the start of the settlement phase when the initial image frame of the distributed game cards is detected as a predetermined number.

[0154] During the card distribution phase, if the position of the game coin changes, or if the position, type, or order of the game cards differs from the prescribed rules, the play recognition device 310 determines that cheating has occurred. Furthermore, the play recognition device 310 determines the payout based on the game result obtained from the hand identified by the hand recognition device 305, the player's placement amount identified by the placement recognition device 306, and the rules.

[0155] Furthermore, player identification device 308 can match players to each other based on the distance between their positions. Placement identification device 306 can match players to placements based on the correspondence between players and player positions obtained by player identification device 308. Placement identification device 306 can also match players to their hand based on the distance between the player's hand identified by player identification device 308 and the hand cards identified by hand card identification device 305.

[0156] (Variation Example 2) To reduce the burden on the dealer and improve the accuracy of monitoring warnings, the game management system 100 in Modified Example 2 automatically detects changes in game currency and issues warnings. The play recognition device 310 compares the placement identified by the placement recognition device 306 when the game status recognition device 311 identifies it as being in the placement phase with the placement identified by the placement recognition device 306 when the game status recognition device 311 identifies it as being in the card dealing phase, and outputs a warning when there is a change between the two.

[0157] Cameras 11 and 12 continuously capture images of the tabletop 72 at predetermined time intervals during the card distribution phase, thereby generating multiple images in a time sequence and outputting them to the game management device 30. The placement recognition device 306 identifies placements from each of the multiple images in the time sequence and determines the placement amount. The placement recognition device 306 uses the placement recognition from the time the card distribution phase switches to the initial placement.

[0158] After switching to the card dealing stage, the placement recognition device 306 continues to recognize placements, and the play recognition device 310 compares the recognized current placement with the initial placement. The play recognition device 310 outputs a warning if there is a change between the current and initial placements. It should be noted that the play recognition device 310 can output a warning if the changed image frame is only one frame, or if the changed image continues for a specified number of frames. Furthermore, the play recognition device 310 can withdraw the warning if no further changes are detected in frames following the frame where the placement change was detected.

[0159] (Variation Example 3) To reduce losses for the casino, the game management system 100 in Modification 3 can detect changes in player placement during the liquidation phase. When the game status recognition device 311 determines that the game is in the liquidation phase and the placement recognition device 306 detects a change in placement, the play recognition device 310 outputs a warning.

[0160] Cameras 11 and 12 capture images of the tabletop 72 at predetermined time intervals, outputting continuous images to the game management device 30. Card detection device 301 detects game cards from the images, and game status recognition device 311 identifies the game stage as being in the liquidation phase based on the detected game cards. When game status recognition device 311 identifies the game stage as being in the liquidation phase, and when game coin operation recognition device 309 identifies that a player has added or removed placed game coins based on placement recognition device 306, play recognition device 310 outputs a warning.

[0161] At this time, the game coin operation recognition device 309 recognizes that the player has added or removed placed game coins based on the first frame image (excluding the hand) before the game coin change was detected, the second frame image (the hand was detected), and the third frame image (the game coin change was detected).

[0162] Since the dealer and the player are in different positions, the play recognition device 310 can determine that the player has moved the game tokens based on the hand movements towards a pre-set direction and the changes in the game tokens. Furthermore, the player recognition device 308 can determine whether the person manipulating the game tokens is the player or the dealer by identifying the player associated with the hand shown in the image.

[0163] (Variation Example 4) In Variation 4, the game management system 100 analyzes images of a state where an error was detected in the configuration of game coins during payment and recycling (a stopped state), and images showing placement changes at the end of game coin placement. The play recognition device 310 determines the placement change based on the images of the stopped state and the end of game coin placement, and outputs a warning. The play recognition device 310 also outputs a processing warning in response to the dealer's hand appearing in or near the placement area.

[0164] The game management device 30 acquires an image of the moment when the game coin placement ends. The game management device 30 also acquires an image of the desktop 72 in a stopped state. The play recognition device 310 compares the game coin information (number or amount) in the image of the stopped state with the image of the moment when the game coin placement ends. Furthermore, in response to a situation where there is a change in the game coins between the image of the stopped state and the image of the moment when the game coin placement ends, the play recognition device 310 determines whether the area where the game coins changed is any of the placement areas 77.

[0165] If the area where the game currency changes is any of the placement areas 77, the play recognition device 310 outputs a warning. Furthermore, based on the image of the stopped state, the play recognition device 310 recognizes that a hand associated with the dealer is present in or near the placement area 77 and outputs a warning. The dealer changing the game currency indicates that payment is in progress. The play recognition device 310 recognizes the hand in the image as the dealer's hand based on facial recognition or the direction the hand is extending.

[0166] (Variation Example 5) Modification 5 of the game management system 100 has a structure that improves the accuracy of game currency and player recognition, as well as the accuracy of payment recognition. The player recognition device 308 associates placement during the placement phase with the player's image (first association). Based on the first association, the player recognition device 308 associates game currency during the settlement phase with the player's image (second association). The player recognition device 308 identifies the player based on the player's image and determines the payment based on the placement associated with the player's image in the first association and the game result.

[0167] When the game state recognition device 311 determines that the game is in the placement stage, the placement recognition device 306 acquires images from the camera 11 and the camera 12, and identifies the placement by analyzing the images. The player recognition device 308 identifies the player (first placement recognition), and the play recognition device 310 matches the identified placement with the identified player.

[0168] When the game status recognition device 311 determines that the game is in the settlement phase, before payment is collected, the placement recognition device 306 acquires images from cameras 11 and 12 and identifies placement (second placement recognition) by analyzing the images. Based on the above correspondence, if it is determined that the second placement recognition is normally associated with the player, the play recognition device 310 queries the personal identification information corresponding to the player's image. Furthermore, the play recognition device 310 determines the payment corresponding to the personal identification information based on the second placement recognition and the game result.

[0169] (Variation Example 6) To save labor costs and improve game efficiency, the game management system 100 in Modified Example 6 automatically switches the game state. The card detection device 301 analyzes the image to determine the quantity of game items (number of game cards) on the game table 7, and the card value recognition device 302 analyzes the image to identify the information (suit and card value). When the quantity of game items reaches a predetermined value, the play recognition device 310 determines the game result based on the quantity of game items, the identification information, and the predetermined game result generation conditions (the player whose total card points are closer to 9 wins, etc.). Based on the game result determination by the play recognition device 310, the game state recognition device 311 switches the game stage to the result processing stage.

[0170] Cameras 11 and 12 continuously capture images of the tabletop 72 at predetermined time intervals, thereby generating multiple sequential images, which are then output to the game management device 30. The card detection device 301 analyzes the images to detect the game cards on the tabletop 72, thereby determining the number of game cards. Furthermore, the card value recognition device 302 analyzes the images to identify the card value and suit of the game cards on the tabletop 72.

[0171] When the number of game cards reaches a predetermined value, the play recognition device 310 determines the game result based on the number of game cards, their suits and face values, and the game result generation conditions derived from the game rules. For example, if four cards are dealt, the play recognition device 310 determines that the player who receives the card with the total points closest to 9 wins, based on the game result generation conditions derived from the game rules of a specific card game.

[0172] Before determining the game result, the play recognition device 310 can determine whether the game card is facing forward or backward, and use facing forward as a condition to determine the game result. Furthermore, the play recognition device 310 can determine the game result based on whether the distance between the game card and the hand exceeds a predetermined distance. Additionally, the play recognition device 310 can determine the game result based on the recognition result of the position determination device 303, using the condition that the position of the game card does not change within a predetermined time.

[0173] If the play recognition device 310 determines the game outcome, the game state recognition device 311 recognizes that the game phase has switched from the card dealing phase to the settlement phase. When there are four game cards, if the game outcome cannot be determined, the dealer deals a fifth game card. The play recognition device 310 determines whether the game outcome can be determined based on the increase in the number of game cards to five. Furthermore, if the play recognition device 310 determines that the dealing of game cards did not follow the game rules or operating rules, it outputs a warning indicating an anomaly and stops the game.

[0174] (Variation Example 7) Modification 7 of the game management system 100 includes a structure for reducing casino losses due to player cheating. The placement recognition device 306 determines whether there has been a change in placement within the placement area 77 based on an image from a moment when the game state recognition device 311 determines the game is in the liquidation phase. If a change has occurred, the placement recognition device 306 identifies the placement area 77 involved in the change. The game coin operation recognition device 309 determines whether the person associated with the change in placement is a player. The play recognition device 310 outputs a warning if the person associated with the change in placement is a player.

[0175] The placement recognition device 306 acquires images captured by cameras 11 and 12 when the game state recognition device 311 determines that the game stage is in the liquidation phase. Based on the images, the placement recognition device 306 identifies the placement area 77 where game coins are recycled or paid out as the target area. The play recognition device 310 compares the images from the placement phase with those from the liquidation phase. If a change in placement occurs within the target area, based on the recognition result of the game coin operation recognition device 309, it determines the person corresponding to the added or removed game coins and outputs a warning if the determined person is a player.

[0176] As described above, according to embodiments and variations of the present invention, an improved technique for managing games in casinos is provided. The various features of the game management system described as embodiments and variations thereof can be arbitrarily combined and applied.

[0177] (Variation Example 8) In Variation 8, the game management system 100 makes predictions related to casino operations based on data stored in the game history database 43. The game history database 43 stores game history records, for example, spanning more than a year. The game management device 30 refers to the game history database 43 and uses artificial intelligence to predict factors such as the number of players, the number of game sessions, and transaction volume. This artificial intelligence can use predictive models designed for general demand forecasting. The game management device 30 can use predictive models such as linear regression models, classification models, decision tree models, random forest models, and neural network models.

[0178] The game management device 30 predicts, based on demand forecasting, how many players will be present during a specific period, date, and time slot, how many game tables should be open at that time, and the predicted transaction volume and even sales volume. Based on this forecast, the casino configures the dealer, opens game tables, prepares the required amount of game currency for the game tables, or takes necessary measures.

[0179] Furthermore, the game management device 30, on a per-game-table basis, predicts when game coins need to be replenished and when game coins need to be deposited, based on the game history records of each game table in the game history database 43. By displaying the prediction results on the display panel 61, the dealer can instruct the replenishment of game coins before the game coins in the game coin tray 71 are insufficient, and instruct the deposit of game coins before they overflow from the game coin tray 71.

[0180] Symbol explanation: 7 Game Tables 71 Game Coin Tray 72 Desktop Cameras 11 and 12 21, 22 RFID reading devices 30 Game management devices 301 Card Detection Device 302 Card Value Recognition Device 303 Position Determination Device 304 Orientation Determination Device 305 Hand Card Recognition Device 306 Placement of identification device 307 Gesture Recognition Device 308 Player Identification Device 309 Game Coin Operation Recognition Device 310 Play Recognition Device 311 Game Status Recognition Device 41 Player Database 42 Game Currency Database 43 Game History Database 51 Electronic Boots 61 Display panel.

Claims

1. A game management system, wherein, The game management system has the following features: A camera that captures an image of the game table's surface from an angle above and generates a visual representation. and The card detection device uses artificial intelligence to analyze the image and detects the card by detecting a number, which is a number representing the face value of the card printed on the card placed at any position on the table, and the angle and size of the number being observed are not fixed.

2. The game management system according to claim 1, wherein, The game table mentioned is a game table specifically designed for card games. The cards mentioned are those that make up the player's or dealer's hand.

3. The game management system according to claim 1, wherein, The card detection device detects the card by detecting the pairs of numbers.

4. The game management system according to claim 1, wherein, The card detection device detects the card by detecting the combination of the numbers in a predetermined position with the markings indicating the suit.

5. The game management system according to claim 1, wherein, The game management system also has: A card denomination recognition device that uses artificial intelligence to analyze the image to identify the card denomination detected by the card detection device; and The hand card recognition device identifies a deck of cards consisting of multiple cards detected by the card detection device, and identifies the hand cards in each deck based on the face value of the cards identified by the face value recognition device.

6. The game management system according to claim 1, wherein, The game management system also includes an orientation determination device, which determines the orientation of the card based on the orientation of the card face value and / or suit detected by the card detection device, or the arrangement direction of the card face value and suit.

7. The game management system according to claim 1, wherein, The game management system also includes a position determination device, which determines the position of the card based on the card face value and suit position detected by the card detection device.

8. A game management system, wherein, The game management system has the following features: A camera that captures an image of the tabletop of a specific card game from an oblique angle above and generates an image; A hand card recognition device that uses artificial intelligence to analyze the image to identify the hand cards; A placement recognition device that uses artificial intelligence to analyze the image to identify the placement of game coins; as well as The play recognition device identifies the play content based on the hand cards identified by the hand card recognition device and the placement identified by the placement recognition device, according to the game rules. The play recognition device can identify actions such as dealing cards and doubling the placement as player actions. The play recognition device is configured as follows: Based on multiple images taken in chronological order, a first determination is made to determine whether a player's hand of cards has been divided into two hands; Based on multiple images taken in chronological order, a second determination is made to determine whether the player's placement increases the amount; Based on the results of the first and second determinations, it is determined whether the card splitting and doubling of placement have been performed.

9. The game management system according to claim 8, wherein, When the first determination determines that a player's hand of cards has been divided into two hands, and when the second determination determines that the player's placement has increased the amount, the play recognition device recognizes that the card division has been performed.

10. The game management system according to claim 8, wherein, When the first determination determines that a player's hand of cards has not been divided into two hands, and when the second determination determines that the player's placement has increased in value, the play recognition device recognizes that the double placement has been performed.

11. The game management system according to claim 9, wherein, The game management system also includes a gesture recognition device that uses artificial intelligence to analyze the images and identify gestures. The play recognition device also identifies whether the card splitting has been performed based on the gesture detected by the gesture recognition device.

12. The game management system according to claim 10, wherein, The game management system also includes a gesture recognition device that uses artificial intelligence to analyze the images and identify gestures. The play recognition device also identifies whether the double placement has been performed based on the gestures recognized by the gesture recognition device.

13. The game management system according to claim 9, wherein, When the play recognition device detects that the cards have been dealt, it identifies the two hands as belonging to the same player.

14. The game management system according to claim 9, wherein, The game table for this specific card game has multiple game positions. The hand card recognition device corresponds the two hand cards to the same game position.

15. The game management system according to claim 9, wherein, The game management system also includes a player identification device that identifies players participating in the game. The hand recognition device associates the hand cards with the players identified by the player recognition device. The hand recognition device associates the two hands with the same player.

16. The game management system according to claim 9, wherein, The hand recognition device determines whether a hand can be divided into cards based on rules.

17. The game management system according to claim 8, wherein, The play recognition device determines whether the player's placement is multiplied, as the second determination.

18. A game management system, wherein, The game management system has the following features: A camera that captures images of the tabletop of a game table used for a specific card game; A hand recognition device analyzes the image to detect the face value and suit of the cards, thereby identifying the hand consisting of the cards dealt to the table. A placement recognition device that identifies the placement of game coins on the table; as well as The play recognition device determines whether a specific card game being played on the table is being conducted according to the rules, based on the cards identified by the hand recognition device and the placement identified by the placement recognition device. The play recognition device is configured as follows: When the hand card recognition device identifies the dealer's face-up card as A, based on the placement identified by the placement recognition device in a time sequence, it is determined that the hand card identified by the hand card recognition device has been placed safely. When it is determined that insurance placement has been carried out, and the hand card recognition device identifies that the dealer's hand card meets a specific point value, the insurance placement is deemed successful.

19. A game management system, wherein, The game management system has the following features: A camera that captures images of the tabletop of a game table used for a specific card game; A hand recognition device analyzes the image to detect the face value and suit of the cards, thereby identifying the hand consisting of the cards dealt to the table. A placement recognition device is provided to identify the placement of game coins on the table. as well as The play recognition device determines whether a specific card game being played on the table is being conducted according to the rules, based on the cards identified by the hand recognition device and the placement identified by the placement recognition device. The play recognition device is configured as follows: Based on the placement identified by the placement recognition device in a time sequence, it is determined that the placement in the game has been reduced in value; When it is determined that the aforementioned reduction has been carried out, through Based on the aforementioned rules, it is determined whether the situation warrants surrender, and... Based on the placement identified sequentially by the placement identification device, it is determined whether the reduction is to half of the original placement amount. Therefore, the reduction in amount is determined to be a surrender in accordance with the rules.

20. The game management system according to claim 19, wherein, The game management system also includes a gesture recognition device that uses artificial intelligence to analyze the images to identify gestures. The play recognition device identifies the surrender based on the gesture detected by the gesture recognition device.

Images