Management System

Abstract

This system provides a novel mechanism for detecting fraudulent activity in games at amusement facilities, as well as errors and fraudulent activity during betting and settlement of chips. [Solution] A fraud detection system for detecting fraudulent activity in a game in which chips are collected and redeemed according to the win / loss result comprises a camera 2 that photographs the chips 9 stored in the chip tray 17 of the dealer 5, an image analysis device 12 that analyzes the images taken by the camera 2 to detect the amount of chips 9 stored in the chip tray 17, a card distribution device 13 that determines the win / loss result of the game, and a control device 14 that detects fraudulent activity by comparing the win / loss result of the game with the amount of chips 9 stored in the chip tray 17 before and after the collection and redemption of chips.

Description

Technical Field

[0001] The present invention relates to a system for detecting fraud in games at a gaming venue, or mistakes and fraud during chip betting or settlement.

Background Art

[0002] In gaming venues such as casinos, attempts are being made to prevent various types of fraud. Gaming venues have surveillance cameras for monitoring fraud, and prevent fraud by determining fraud in games or fraud due to chip collection or reimbursement different from the win / loss results from images obtained from the surveillance cameras.

[0003] On the other hand, it has been proposed to attach a wireless IC (RFID) tag to each chip to grasp the amount of the chip in order to grasp the number and total amount of the bet chips.

[0004] In the card game monitoring system described in Patent Document 1, it is determined whether the chips placed on the gaming table are collected or reimbursed as per the win / loss results by analyzing the movement of the chips through image analysis, and fraud monitoring is performed.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0006] An object of the present invention is to provide a novel system for detecting fraud in games at a gaming venue, or mistakes and fraud during chip betting or settlement.

Means for Solving the Problems

[0007] A fraud detection system according to one aspect of the present invention is a fraud detection system for a game hall having a plurality of game tables, comprising: a game recording device that records the state of games played at the game tables as video using a camera; an image analysis device that performs image analysis on the recorded video of the state of games; a win / loss result determination device that determines the win / loss result of each game at the game table; and a control device that detects fraudulent activity at the game tables using the image analysis results from the image analysis device and the win / loss result determined by the win / loss result determination device, wherein the control device detects the position and type of chips bet by each player via the image analysis device and The system determines the number of chips and the total amount of chips in the dealer's chip tray at the gaming table. From the total amount of chips in the chip tray before settlement for each game, it adds or subtracts the increase or decrease in chips in that game, calculated from the position, type, and number of chips bet by all players in that game and the win / loss result of that game obtained by the win / loss result determination device. It then compares the expected total amount of chips in the chip tray after settlement at the end of the game with the actual total amount of chips in the chip tray at the end of the game, obtained via the image analysis device, and determines whether there is a difference between the expected total amount and the actual total amount.

[0008] In the fraud detection system described above, the control device may determine whether there is a difference between the expected total amount of chips and the actual total amount of chips in the chip tray. This can be done by using the image analysis device to determine the position, type, and number of chips each player places, and then comparing this with the actual total amount of chips in the chip tray. The control device may determine whether there is a difference between the expected total amount and the actual total amount.

[0009] In the fraud detection system described above, the control device may compare the total amount of chips in the chip tray before settlement of each game with the actual total amount of chips in the chip tray, which is calculated by adding the increase in the chip tray for that game based on the position, type, and number of chips bet by the losing player, and determine that there is no difference between the expected total amount and the actual total amount. Furthermore, if the control device determines that there is a difference between the expected total amount and the actual total amount, it may determine that there is a payment error and generate a payment error signal to indicate the payment error.

[0010] In the fraud detection system described above, the chip tray is provided with a collection chip tray for collecting and temporarily storing chips bet by losing players. The image analysis device and the control device may compare the expected amount of chips in the collection chip tray, calculated from the position, type, and number of chips bet by the losing players, with the actual total amount of chips in the collection chip tray, and determine whether there is a difference between the expected total amount and the actual total amount in the collection chip tray.

[0011] In the above fraud detection system, obtaining the actual total amount of chips in the chip tray after settlement at the end of the game via the image analysis device is: 1) When the redemption of winning chips is complete, 2) When the cards used in the game are collected and discarded in the discard area of ​​the table, 3) When a predetermined button attached to the win / loss result determination device is pressed, 4) When the markers indicating the win or loss are returned to their original positions.

[0012] In the above fraud detection system, when the control device determines that the actual total amount of chips known in the dealer's chip tray at the gaming table does not correspond to the increase or decrease in chips calculated from the total amount of chips bet by all players and the outcome of the game, the game recording device may be configured to either assign an index or timestamp to the acquired video, or to identify and play back chip collection scenes or payment scenes, so that the record of the game in which the discrepancy occurred can be analyzed.

[0013] In the fraud detection system described above, the image analysis device or control device may be structured to be able to obtain information on the type, number, and location of the chips bet, even if some or all of the chips placed on the game table are hidden by the camera's blind spot.

[0014] In the above fraud detection system, the control device is 1) The position, type, and number of chips bet at each playing position on the gaming table are recorded, and the win / loss history and amount of chips obtained from the win / loss results of each game are compared with statistical data from past games to extract any unusual situations, or 2) At the playing position on the gaming table, situations where the amount of chips bet when losing is less than the amount of chips bet when winning are identified as unique situations compared to statistical data from past games. It may be a structure that allows for this.

[0015] In the fraud detection system described above, the control device may be able to compare and determine whether the amount of chips held in the dealer's chip tray at the gaming table has increased or decreased after the exchange of bills for chips, in accordance with the amount of chips paid for the exchanged bills, or the amount of bills paid for the exchanged chips.

[0016] In the fraud detection system described above, the control device may further include a database that records the history of the exchange of bills and chips, and may refer to the database at regular intervals or on a daily basis to compare and determine whether the amount of chips recorded in the dealer's chip tray at the gaming table has increased or decreased in accordance with the amount of chips paid corresponding to the exchanged bills, or the total amount of bills paid corresponding to the exchanged chips.

[0017] In the fraud detection system described above, the control device may be able to identify the player at the play position extracted as a difference or unusual situation via the image analysis device.

[0018] In the fraud detection system described above, the control device may have a warning function that notifies the other gaming table of the presence of the identified player when the identified player leaves their seat and sits at another gaming table.

[0019] In the above fraud detection system, the control device further 1) In each game, check for any movement of chips between the start of card drawing or the dealer's game start operation and the display of the game's outcome by the card distribution device. 2) After each game has finished, while the dealer is collecting the chips that the losers among the game participants had bet, check whether the losers have taken any chips. 3) After each game ends, while the dealer is collecting the chips bet by the losing players, check whether any additional chips were added. 4) After each game has finished, the dealer has checked whether he has paid out the chips that the winners among the game participants had placed. 5) After each game has finished, the winner among the game participants must determine whether they have taken the chips they wagered and the chips they have received. It may have a function to determine at least one of the following.

[0020] In the above-described fraud detection system, the win / loss result determination device may be a card distribution device that distributes cards on the gaming table, or a control device that determines the win / loss results of each game based on the information of the image analysis device that reads the cards distributed on the gaming table with a camera.

Advantages of the Invention

[0021] According to the fraud detection system of the present invention, fraud in the collection and repayment of chips according to the win / loss results of games can be detected.

[0022] Further, according to the system of the present invention, even if a card is bent due to squeezing by a player, which is often done in baccarat games or the like, the rank and suit of the card can be determined by image analysis, and the total amount of dead corners and overlapping chips can be grasped together with the position. Also, fraud during the exchange of bills and chips can be detected.

Brief Description of the Drawings

[0023] [Figure 1] FIG. 1 is a diagram showing an overall overview of a fraud detection system in a casino having a plurality of gaming tables according to a first embodiment of the present invention. [Figure 2A] FIG. 2A is a perspective view of chips showing examples of different stacked states of chips grasped in the first embodiment of the present invention. [Figure 2B] FIG. 2B is a perspective view of chips showing examples of different stacked states of chips grasped in the first embodiment of the present invention. [Figure 3A] FIG. 3A is a diagram showing details of a chip tray according to the first embodiment of the present invention. [Figure 3B] FIG. 3B is a diagram showing another example of a chip tray according to the first embodiment of the present invention. [Figure 4] FIG. 4 is an enlarged view of a mark explaining the dirt on a card grasped in the first embodiment of the present invention. [Figure 5A]Figure 5A is a plan view showing the table of markers according to the first embodiment of the present invention. [Figure 5B] Figure 5B is a plan view showing the back of the marker according to the first embodiment of the present invention. [Figure 6] Figure 6 is a simplified explanatory diagram illustrating the video of the exchange state between banknotes and chips as observed in the first embodiment of the present invention. [Figure 7] Figure 7 is a plan view showing an overall overview of the baccarat game fraud detection system according to a second embodiment of the present invention. [Figure 8] Figure 8 is a plan view showing an overview of the progress of a baccarat game in the fraud detection system of the second embodiment of the present invention. [Figure 9] Figure 9 is an explanatory diagram showing how a dealer collects the chips won by the casino in a baccarat game. [Figure 10] Figure 10 is an explanatory diagram showing how a dealer pays a winning customer (game participant) in a baccarat game according to a second embodiment of the present invention. [Figure 11] Figure 11 is an explanatory diagram showing how a winning customer (game participant) receives chips and payment in a baccarat game according to a second embodiment of the present invention. [Figure 12A] Figure 12A is an explanatory diagram illustrating the images that are subject to image analysis of the collection of chips won by the casino by the dealer in the fraud detection system of the second embodiment of the present invention. [Figure 12B] Figure 12B is an explanatory diagram illustrating the images subject to image analysis of the chips won by the casino by the dealer in the fraud detection system of the second embodiment of the present invention. [Figure 12C] Figure 12C is an explanatory diagram illustrating the images that are subject to image analysis of the chips won by the casino by the dealer in the fraud detection system of the second embodiment of the present invention. [Figure 13] Figure 13 is an explanatory diagram illustrating the images that are subject to image analysis of illegally recovered chips in the fraud detection system of the second embodiment of the present invention. [Figure 14A]Figure 14A is an explanatory diagram illustrating the images subject to image analysis of chips won by game participants in the fraud detection system of the second embodiment of the present invention. [Figure 14B] Figure 14B is an explanatory diagram illustrating the images subject to image analysis of chips won by game participants in the fraud detection system of the second embodiment of the present invention. [Figure 14C] Figure 14C is an explanatory diagram illustrating the images subject to image analysis of chips won by game participants in the fraud detection system of the second embodiment of the present invention. [Figure 14D] Figure 14D is an explanatory diagram illustrating the images subject to image analysis of chips won by game participants in the fraud detection system of the second embodiment of the present invention. [Figure 15] Figure 15 is a side cross-sectional view of a card distribution device according to a second embodiment of the present invention. [Figure 16] Figure 16 shows an example of a card according to a second embodiment of the present invention. [Figure 17] Figure 17 is a plan view of the main part of the card guide section of the card distribution device of the second embodiment of the present invention, with a portion of it cut away. [Figure 18A] Figure 18A is a side cross-sectional view of a key part showing a distribution restriction device that restricts the entry and exit of cards from the card storage section of a card distribution device according to a second embodiment of the present invention. [Figure 18B] Figure 18B is a side cross-sectional view of a key part showing a modified example of a distribution restriction device that restricts the entry and exit of cards from the card storage section of a card distribution device according to a second embodiment of the present invention. [Figure 19] Figure 19 shows the relationship between the output waveforms of sensors and the marks in a card distribution device according to a second embodiment of the present invention. [Modes for carrying out the invention]

[0024] (First Embodiment) In casinos and other gaming establishments, chips are piled high and placed on the gaming table. However, IC tag readers installed beneath the table have difficulty accurately reading the total amount. If the sensitivity of the reader is increased, chips placed in different locations (where the outcome of the game depends) are added together, making it impossible to determine the total amount of chips at each location. Furthermore, imaging from cameras presents challenges such as blind spots due to the camera's field of view and shadows caused by overlapping chips, making it impossible to determine the total amount of chips.

[0025] Furthermore, a challenge exists in baccarat games where players often squeeze the cards (bending face-down cards to gradually examine their rank and other characteristics), causing the cards to bend, making it impossible to determine the card's rank and suit through image analysis from a camera.

[0026] Furthermore, it has been recognized that fraud at gaming tables is becoming more sophisticated, and that advanced betting methods that cannot be detected by simply looking at the amount of winnings at a particular table cannot be detected by cameras or tracking winnings. In addition, fraudulent activities by dealers and players through collusion cannot be adequately prevented with conventional technology.

[0027] In order to solve the various problems described above, the fraud detection system in a gaming parlor having multiple gaming tables according to the first embodiment is: A fraud detection system for a gaming parlor having multiple gaming tables, A game recording device that records the progress of a game played on the aforementioned gaming table as video footage via a camera, including the dealer and the player. An image analysis device for image analysis of the recorded video of the game's progress, A card distributing device that determines and displays the win / loss results of each game on the aforementioned game table, The system includes a control device that detects fraudulent activity at the game table using the image analysis results from the image analysis device and the win / loss results determined by the card distribution device.

[0028] Furthermore, the fraud detection system includes a card distribution device with a structure capable of reading the rank of the distributed cards, and a control device capable of comparing the rank information obtained by the image analysis device from the video of each card distributed at the game table with the rank information of the cards read by the card distribution device to determine whether they match or not.

[0029] Furthermore, the fraud detection system, which includes an image analysis device or control device, is an artificial intelligence-based or deep learning structure capable of obtaining card rank information from cards distributed at the gaming table that have been bent or soiled by the player.

[0030] Furthermore, in the fraud detection system, the control device grasps the position, type, and number of chips bet by each player via the image analysis device, and determines whether the collection of losing chips and payment of winning chips bet by each player were carried out appropriately according to the outcome of the game by analyzing video footage of the game's progress via the image analysis device.

[0031] Furthermore, the fraud detection system, which includes an image analysis device or control device, is an artificial intelligence-based or deep learning-based system capable of obtaining information on the type, number, and location of the chips bet, even if some or all of the chips placed on the gaming table are hidden by the camera's blind spot.

[0032] Furthermore, the fraud detection system, with a control device that utilizes artificial intelligence or a deep learning structure, is capable of comparing and calculating, according to the game's outcome, whether the amount of chips tracked in the dealer's chip tray at the gaming table has increased or decreased after the game ends and settlement occurs, in accordance with the recovery of losing chips and the payment of winning chips to each player.

[0033] Furthermore, the fraud detection system, which includes a control device that tracks the position and amount of chips bet at each playing position on the gaming table, and uses artificial intelligence or a deep learning structure to extract unique situations by comparing each player's win / loss history and the amount of chips obtained from the win / loss results of each game with statistical data from past games.

[0034] Furthermore, the fraud detection system, in which the control device is an artificial intelligence-based or deep learning structure capable of extracting a situation where, at a particular playing position on a game table, the amount of chips wagered when losing is less than the amount of chips wagered when winning is a unique situation compared with statistical data from past games.

[0035] Furthermore, the fraud detection system has a structure that allows the control device to identify individual players at play locations where a specific situation is extracted via the image analysis device, or where a winning amount exceeding a predetermined amount has been achieved.

[0036] Furthermore, the fraud detection system includes a control device that has a warning function that notifies the other gaming table of the presence of the identified player when the identified player leaves their seat and sits at another gaming table.

[0037] In order to solve the various problems described above, the fraud detection system for a gaming parlor having multiple gaming tables according to the present invention is A game recording device that records the progress of a game played on the aforementioned gaming table as video footage via a camera, including the dealer and the player. A card distributing device that determines and displays the win / loss results of each game on the aforementioned game table, An image analysis device for image analysis of the recorded video of the game's progress, The game table is equipped with a control device that can detect banknotes and chips using the image analysis results from the aforementioned image analysis device, The aforementioned image analysis device or control device is capable of detecting, based on information obtained from the card distribution device or the dealer, when an exchange of banknotes for chips is taking place on the game table outside of card dealing, and is capable of recognizing the total value of genuine banknotes verified by a black light, and is capable of recognizing the total value of chips even if some or all of the chips placed on the game table as exchange targets are partially or entirely hidden by the camera's blind spot, and is an artificial intelligence-based or deep learning structure capable of comparing the total value of banknotes placed on the game table by the player with the total value of chips placed by the dealer and determining whether the amounts match.

[0038] Furthermore, the fraud detection system, which includes a control device, is an artificial intelligence-based or deep learning structure capable of comparing and calculating whether the amount of chips tracked in the dealer's chip tray at the gaming table has increased or decreased in accordance with the amount of chips paid for the exchanged bills after the exchange of bills for chips and settlement.

[0039] Furthermore, the fraud detection system, in which the control device is an artificial intelligence-based or deep learning structure capable of comparing and calculating the agreement or disagreement between the amount of banknotes deposited by the dealer based on the input and the total amount of banknotes based on the image analysis results of the image analysis device, after the exchange of banknotes and chips has taken place and settlement has been made. Moreover, the control device is an artificial intelligence-based or deep learning structure capable of comparing and calculating the agreement or disagreement between the total amount of banknotes deposited by the dealer based on the input and the total amount of banknotes based on the image analysis results of the image analysis device at the gaming table in charge of the dealer.

[0040] According to the fraud detection system of this embodiment, even if a card is bent due to a player squeezing it, which is a common practice in games such as baccarat, the rank and suit of the card can be determined through image analysis, and the total amount and location of chips in blind spots or overlapping chips can also be determined. Furthermore, fraud during the exchange of bills for chips can also be detected.

[0041] The overall outline of the fraud detection system for a gaming arcade having multiple gaming tables according to the first embodiment of the present invention will be described in more detail below. Figure 1 is a diagram showing the overall outline of the system, and the fraud detection system for a gaming arcade having multiple gaming tables 4 includes a game recording device 11 that records the progress of games played on the gaming tables 4 as video via multiple cameras 2, including players 6 and dealers 5; an image analysis device 12 that performs image analysis on the recorded video of the game progress; and a card distribution device 3 that determines and displays the win or loss result of each game on the gaming tables 4. The card distribution device 3 is a so-called electronic shoe already used by those skilled in the art, and is programmed with the rules of the game in advance. It is structured to read the information on the distributed cards C and determine the win or loss of the game. For example, in a baccarat game, a banker win, a player win, or a tie is basically determined by the rank of 2-3 cards, and the determination result (win or loss result) is displayed on a result display lamp 13.

[0042] This fraud detection system further includes a control device 14 that compares the actual card rank determined by the image analysis results of the image analysis device 12 with the win / loss result determined by the card distribution device 3 to detect fraudulent activity (such as a mismatch between the sum of the ranks of the distributed cards and the win / loss result) that occurs at the game table 4. The card distribution device 3 is structured to read the rank (A, 2-10, J, Q, K) and suit (hearts, spades, etc.) of cards C that are manually distributed by the dealer 5. The control device 14 is structured to compare the rank and suit information obtained by the image analysis device 12 (using artificial intelligence) from the video (captured using camera 2) of each card distributed at the game table 4 with the card and suit information read by the card distribution device 3 to determine whether they match or not. The image analysis device 12 and the control device 14 in this fraud detection system have a structure that combines a computer, program, and memory, either as an integrated unit or as a group of components.

[0043] The image analysis device 12 and control device 14 have an artificial intelligence-based or deep learning structure that can obtain information about the rank of a card even if the card C is distributed on the game table 4 and bent or soiled by the player 6. As shown in Figure 4, soiled card C can make it difficult to distinguish between clubs and spades. Even in such cases, the suit can be determined by analyzing and judging the image using an artificial intelligence-based computer or control system and deep learning (structure) technology. Furthermore, even if a card is bent by squeezing it, which is often done in games such as baccarat, the artificial intelligence-based computer or control system and deep learning (structure) technology can recognize the suit and rank of the card before deformation by utilizing self-learning from a large number of image deformations. Since artificial intelligence-based computers or control systems and deep learning (structure) technology are already known and available to those skilled in the art, a detailed explanation is omitted.

[0044] The control device 14, which utilizes artificial intelligence or has a deep learning structure, can determine, via the camera 2 and image analysis device 12, which position (player, banker, or pair) each player 6 has bet on in the betting area 8 with their chips 9, the type of chips 9 bet (each color of chip 9 is assigned a different value), and the number of chips bet. The chips 9 are stacked offset from each other, as shown in Figure 2A, rather than stacked in a vertical alignment. In this case, if the camera 2 is positioned in the direction of arrow X shown in Figure 2A (or if the orientation of the chips 9 is relatively in a blind spot), it is expected that the chips 9 will not be visible (enter the blind spot), as shown in Figure 2B. In artificial intelligence-utilizing computers or control systems, or deep learning (structure) technology, self-learning functions, etc., are used to recognize the hiding of chips 9 due to blind spots (when part of a chip is hidden, or when the entire chip is hidden), and accurately determine the number of chips, etc. In this way, it is possible to determine which position in the betting area 8 (player, banker, or pair) each chip 9 was bet on, the type of chip 9 bet (each color of chip 9 is assigned a different value), and the number of chips bet. Therefore, in each game, the control device 14 determines whether the collection of losing chips bet by each player 6 (indicated by arrow L) and the payment of winning chips (9W) to the winning player 6W were carried out appropriately according to the outcome of the game determined by the card distribution device 3, by analyzing the video of the game's progress via the image analysis device 12.

[0045] The control device 14 can analyze and understand the total amount of chips 9 in the dealer's chip tray 17 at the game table 4 using the image analysis device 12. After the game ends and settlement is complete, the control device 14 can compare and calculate whether the total amount of chips 9 in the chip tray 17 has increased or decreased according to the game's outcome, based on the amount of lost chips 9 collected from each player 6 and the amount of winning chips 9W paid to the winning player 6W. Even though the total amount of chips 9 in the chip tray 17 is always known by means such as RFID, the control device 14 determines whether the increase or decrease is correct by analyzing video of the game's progress via the image analysis device 12. These also utilize artificial intelligence or deep learning structures.

[0046] In this example, fraud and errors are detected based on the game's outcome, information on how many of each type of chip 9 were bet on each position in the betting area 8 (player, banker, or pair), and the increase or decrease in the chip 9 in the chip tray 17 after the collection of losing chips and redemption of winning chips 9. Therefore, fraud and errors can be detected without knowing the movement of the chip 9 after the game ends, i.e., whether the bet chips 9 moved to the player's side or the dealer's side.

[0047] Here, the outcome of a game can be determined according to the rules of baccarat, for example, by reading the rank of the card C dealt in the game using the card distributing device 3. Alternatively, the outcome of a game can also be determined by taking a picture of the game table 4 with camera 2, analyzing the image with image analysis device 12, and comparing the analysis results with the game rules using control device 14. In this case, the camera 2, image analysis device 12, and control device 14 constitute the win / loss determination device. Information on which players at each play position 7 placed how many chips of which type in each betting area 8 (player, banker, or pair) is obtained by taking a picture of the chips 9 placed in the betting area 8 with camera 2 and analyzing the image for each play position 7 with image analysis device 12.

[0048] Furthermore, the increase or decrease in the amount of chips 9 in the chip tray 17 before and after the recovery of losing chips 9 and the redemption of winning chips 9 can be calculated by comparing the total amount of chips 9 in the chip tray 17 before the recovery of losing chips 9 and the redemption of winning chips 9 with the total amount of chips 9 in the chip tray 17 after the recovery of losing chips 9 and the redemption of winning chips 9. The total amount of chips 9 in the chip tray 17 before the recovery of losing chips 9 and the redemption of winning chips 9, and the total amount of chips 9 in the chip tray 17 after the recovery of losing chips 9 and the redemption of winning chips 9, can be detected by photographing the chip tray 17 containing the chips 9 with the camera 2 and analyzing the image with the image analysis device 12. Alternatively, the total amount of chips 9 contained in the chip tray 17 may be detected by embedding an RFID indicating the amount in the chip 9 and providing an RFID reader in the chip tray 17.

[0049] For example, suppose the total value of chips 9 in chip tray 17 before the game starts is Bb, and the total value of chips 9 in chip tray 17 after the game ends, when the lost chips have been collected and the winning chips have been redeemed, is Ba. Also, suppose that in this game, the total value of chips 9 placed in all 7 playing positions in the player area is bp, the total value of chips 9 placed in all 7 playing positions in the banker area is bb, and the total value of chips 9 placed in all 7 playing positions in the tie area is bt. For example, if the outcome of this game is a banker win, then Ba - Bb = bp - bb + bt should hold true. Alternatively, the total value of chips 9 in chip tray 17 after the game ends, Ba, should be (Bb + bp - bb + bt). If this is not the case, it can be determined that there was fraud or error in the collection or redemption of chips.

[0050] Figure 3A shows details of the chip tray in this embodiment, and Figure 3B shows another example of the chip tray. The chip tray 17 is provided with a collection chip tray 171 for collecting and temporarily storing the chips 9L bet by the losing player 6L, and a redemption chip tray 172 for storing the redemption chips 9W. The image analysis device 12 and the control device 14 determine the location, type, and number of chips 9L bet by the losing player 6L, and calculate the increase in the chips 0L in the game (the amount of chips 9 that should be in the collection chip tray 171). Furthermore, the image analysis device 12 and the control device 14 determine the actual total amount of chips 9 in the chip tray 171 after collection, and compare the expected total amount with the actual total amount to determine if there is a difference.

[0051] Furthermore, the redemption of the chips 9W for the winning player 6W is performed using the chips 9 in the redemption chip tray 172, which allows sufficient time for the image analysis device 12 and the control device 14 to determine the actual total value of the chips 9 in the recovery chip tray 171 after collection.

[0052] The game table 4 is equipped with a discard area 41 and / or discard slots 42 for discarding cards C used in the game. After each game ends, the cards C used in that game are collected and discarded by placing them in the discard area 41 or discard slots 42 on the game table 4.

[0053] The game table 4 is further equipped with markers 43 to indicate the outcome of the game. Figure 5A is a plan view showing the front of the marker, and Figure 5B is a plan view showing the back of the marker. In the baccarat game, two types of markers are used: marker 43a indicating a player win and marker 43b indicating a banker win. When the game is decided, the dealer 5 flips over the marker of the player or banker who won. This makes the outcome of the game clearly indicated on the table. The flipped marker is returned to its original position by the dealer 5 after the collection and redemption of the chips 9 is complete. Returning the marker also means that the next game is ready to begin.

[0054] As described above, in this embodiment, the control device 14 calculates the chip balance for each game based on the amount of chips bet on the game table 4 and the outcome of the game, and verifies the increase in the chip balance in the chip tray 17 after the game. If the control device 14 detects a difference in this verification, it issues an alarm or adds a record to that effect to the video recording captured by the camera 2. The casino operator can investigate the cause of the difference by reviewing the video.

[0055] The fraud detection system of this embodiment adds or subtracts the increase or decrease in chips in the game, calculated from the total amount of chips 9 in the chip tray 17 before settlement of each game, based on the position, type, and number of chips 9 bet by all players 6 in that game and the win / loss result of the game obtained by the win / loss result determination device. It then compares the expected total amount of chips 9 in the chip tray 17 after settlement at the end of the game with the actual total amount of chips 9 in the chip tray 17 at the end of the game, obtained via the image analysis device 12, and determines whether there is a difference between the expected total amount and the actual total amount.

[0056] The control device 14, via the image analysis device 12, grasps the position, type, and number of chips bet by each player. When all of the losing chips bet by each player have been collected, it grasps the actual total value of the chips in the chip tray. It then compares the actual total value of the chips in the chip tray 17 with the total value of the chips in the chip tray 17, which is calculated by adding the increase in the chips bet by the losing players to the total value of the chips in the chip tray 17 before the settlement of each game. The control device 14 then determines whether there is a difference between the expected total value and the actual total value.

[0057] The control device 14 compares the total amount of chips 9 in the chip tray 17 before settlement of each game with the actual total amount of chips 9 in the chip tray 17, which is calculated by adding the increase in the chip tray 17 for that game based on the position, type, and number of chips 9 bet by the losing player, to the total amount of chips 9 in the chip tray 17 before settlement of each game. If it determines that there is no difference between the total amount and the actual total, and if it determines that there is a difference between the total amount and the actual total, it determines that there is a payment error and generates a payment error signal to indicate the payment error.

[0058] The chip tray 17 is equipped with a collection chip tray 171 for collecting and temporarily storing the chips 9 bet by losing players. The image analysis device 12 compares the expected total amount of chips 9 in the collection chip tray 171, which is calculated by adding the increase in the value of chips 9 in the game based on the position, type, and number of chips 9L bet by the losing players, with the actual total amount of chips 9 in the collection chip tray 171, and determines whether there is a difference between the expected total amount and the actual total amount.

[0059] When the control device 14 determines that the actual total amount of chips 9 tracked in the dealer's chip tray 17 of the gaming table 4 does not correspond to the increase or decrease in chips calculated from the total amount of chips bet by all players and the outcome of the game, the game recording device 11 can either assign an index or timestamp to the acquired video, or identify and play back the chip collection scene or payment scene, so that the game in which the above-mentioned difference occurred can be analyzed.

[0060] Thus, the control device 14 obtains the total amount of chips in the chip tray 17 after settlement at the end of the game via the image analysis device 12. In this case, the decision after settlement is made when any of the following 1) to 4) occurs. 1) When the redemption of the winning chips 9 is complete, 2) When card C used in the game is collected and discarded to the discard area 41 or discard slot 42 of the table, 3) When a predetermined button attached to the win / loss result determination device is pressed, 4) When the win / loss marker 43 is returned to its original position.

[0061] Furthermore, the control device 14 is an artificial intelligence-based or deep learning structure that can extract unique situations (set by the casino) when a player wins or loses a game, or when a player loses a game, by comparing the win / loss history and the amount of chips won (amount won) of each player 6 obtained from the win / loss results of each game with statistical data from a large number of past games (big data). Typically, it includes an artificial intelligence-based or deep learning structure that can extract unique situations when a player wins a game, such as when a player wins a game, and the amount of chips bet when they lose is small, and the amount of chips bet when they win is large, for several games in a row, by comparing this with statistical data from past games (big data, etc.).

[0062] Furthermore, the control device 14 of this fraud detection system (integrated with the image analysis device 12) is structured to identify individual players 6 at play positions 7 that have been extracted as unusual circumstances or have won more than a predetermined amount. Such identification of players 6 is done by the image analysis device 12, which obtains facial images by feature point extraction, etc., and assigns an identity number (ID, etc.) to them. The control device 14 then has a warning function that notifies the other game table of the presence of the identified player 6 when the player leaves their seat and takes a seat at another game table. Specifically, it notifies the pit manager or the person in charge of each table (which may also be a dealer) who manages each game table 4, in order to prevent further unusual phenomena.

[0063] The control device 14 further includes a database that records the history of the exchange of banknotes K and chips 9. At regular intervals or on a daily basis, it refers to the database and compares and determines whether the amount of chips 9 recorded in the chip tray 17 of the dealer 5 at the game table 4 has increased or decreased in accordance with the amount of chips 9 paid for corresponding to the exchanged banknotes K, or the total amount of banknotes K paid for corresponding to the exchanged chips 9.

[0064] In the above example, it is also possible to monitor the win / loss history and the amount of chips won (amount won) for each playing position 7 without identifying individual players 6. In this case, it would be impossible to track each player 6 when they leave their seat, but it would be possible to detect unusual situations such as a pattern where the amount of chips bet when losing is small and the amount of chips bet when winning is large for several games at a particular playing position 7 on a single gaming table 4. When such a playing position 7 is detected, there is suspicion that there was fraud or a mistake at that playing position 7. Then, by reviewing the video footage taken at that playing position 7, the fraud or mistake can be discovered.

[0065] Specifically, camera 2 is positioned to photograph at least the chips 9 placed in the betting area 8 of the gaming table 4. The image analysis device 12 analyzes the images captured by camera 2 to detect whether the chips were placed in the player, banker, or tie position in the betting area 8 for each player position 7, and the amount of chips placed. The card distribution device 3 also functions as a win / loss result determination device to determine the outcome of the game. The control device 14 records (monitors) the win / loss history and the amount of chips obtained (chip winnings) for each play position 7 based on the position in the betting area 8 where the chips 9 were placed (player, banker, or tie) and the outcome of the game. Note that only the win / loss history and / or chip winnings may be recorded. If the win / loss history and / or chip winnings history is an unusual situation (set by the casino) compared with statistical data from a large number of past games (big data), the control device 14 identifies this player position 7 as a play position suspected of fraudulent activity.

[0066] If cheating is suspected at a player's position 7, the cheating detection system may generate an alarm (light, sound, or vibration) that can be perceived by the dealer at least at that moment. This can prevent the cheating from continuing, at least by interrupting the game at that point. Additionally, information indicating that cheating is suspected may be added to the video footage captured and recorded by camera 2. This allows the cause of the suspected cheating to be investigated by reviewing the video.

[0067] The fraud detection system in this embodiment for a gaming establishment with a gaming table further includes a function to inspect the exchange of bills and chips, which often occurs at the gaming table 4. In amusement establishments such as casinos, before a game, player 6 exchanges bills (cash, etc.) for game chips at a designated chip exchange counter. However, if player 6 runs out of chips, they can continue the game by exchanging cash (bills) for chips 9 on the gaming table (baccarat table, etc.) without leaving their seat at the gaming table 4. However, this creates an opportunity for fraud to occur between the dealer 5 and the player. The exchange of cash (bills) for chips 9 on the gaming table (baccarat table, etc.) must be performed when the game is not in progress. The card distribution device 3 is capable of detecting the start and end of card dealing (the time when the winner is determined) in order to determine the winner of the game. Therefore, the card distribution device 3 detects situations other than card distribution (dealing), and the control device 14 detects when an exchange of banknotes and chips 9 is taking place at the game table 4 in a situation other than card dealing (as shown in Figure 6). Card dealing (or other situations) can be detected based on information obtained from the operation of the card distribution device 3 or the dealer 5.

[0068] The control device 14 can recognize the number and value of banknotes by performing image analysis on the surface of banknotes K. Furthermore, at the game table 4, whether the banknotes K to be exchanged for chips 9 are genuine is determined by shining a black light on them to detect the genuineness mark G. As shown in Figure 6, the control device 14 also verifies this genuineness mark G through image analysis, recognizes the total value of genuine banknotes, and can recognize the total value of chips even if multiple chips placed on the game table for exchange are hidden by the blind spot of camera 2. It is an artificial intelligence-based or deep learning structure that can compare the total value of banknotes K placed on the game table 4 by the player with the total value of chips 9 placed by the dealer 5 and determine whether the amounts match.

[0069] The control device 14 is an artificial intelligence-based or deep learning structure capable of comparing and calculating whether the total amount of chips 9 in the dealer's 5 chip tray 17 of the gaming table 4 has increased or decreased in accordance with the amount of chips paid for the exchanged bills after the exchange of bills for chips and settlement. It is also possible that the total amount of chips 9 in the dealer's 5 chip tray 17 is always known in advance by RFID or the like on the chips 9. Alternatively, the total amount of chips 9 contained in the chip tray 17 can be detected by photographing the chip tray 17 containing the chips 9 with the camera 2 and analyzing the image with the image analysis device 12.

[0070] Furthermore, the control device 14 verifies whether the increase or decrease in the amount of chips 9 in the chip tray 17 before and after the exchange of banknotes for chips matches the amount of chips exchanged as a result of image analysis on the game table 4. The amount of banknotes paid may be input to the control device 14 by the dealer 5 via key input or the like, or it may be determined by taking a picture of the game table 4 where the banknote payment is made with the camera 2 and analyzing the image with the image analysis device 12.

[0071] As described above, the control device 14 determines whether the reduction in chips 9 from the chip tray 17 due to the exchange of banknotes for chips matches the amount of banknotes paid by the player 6 to the dealer 5. Furthermore, the control device 14 is an intelligent control device that can compare and calculate the match or non-match between the amount of banknotes deposited by the dealer 5 (usually by key input, etc.) and the calculated amount of banknotes as a result of image analysis by the image analysis device 12 after the exchange of banknotes for chips and settlement. It may also be an artificial intelligence-based or deep learning-based control device.

[0072] Furthermore, the control device 14 is an artificial intelligence-based or deep learning structure capable of comparing and calculating the agreement or disagreement between the total amount of banknotes deposited by the dealer at the game table 4 in the dealer's jurisdiction and the total amount of banknotes determined by the image analysis results from the image analysis device 12.

[0073] The control device 14 compares and determines whether the amount of chips 9 held in the chip tray 17 of the dealer 5 at the game table 4 has increased or decreased after the exchange of bills for chips 9, in accordance with the amount of chips 9 paid for the exchanged bills, or the amount of bills paid for the exchanged chips 9.

[0074] (Second Embodiment) Many table games played in casinos and other amusement establishments include baccarat and blackjack. These games use a standard deck of 52 playing cards, and the cards are distributed onto the playing table from a card dispensing device equipped with multiple pre-shuffled decks (6 to 9 or 10 decks). The winner is determined based on the number (rank) of the cards dealt and the game rules.

[0075] The distribution of cards from the card dispensing machine and the settlement of bets to customers (game participants) are handled by the dealer or other staff member in charge of the game table. Casinos and other gaming establishments are making efforts to prevent errors and fraudulent activities in the settlement of bets to customers (game participants).

[0076] The card game monitoring system described in International Publication WO2015 / 107902 uses surveillance cameras to read the movement of chips and check whether or not the bets have been paid to the winners.

[0077] In games like baccarat and blackjack, there's a problem with settling bets placed by customers and those paid by dealers to customers (game participants): it's impossible to detect when these transactions are taking place or who placed or took the chips, making it difficult to determine if they are correct.

[0078] To solve the various problems described above, the fraud detection system of the second embodiment is a fraud detection system for a gaming parlor having a gaming table, A game monitoring device that uses cameras to monitor the progress of games played at gaming tables, An image analysis device that performs image analysis on the video obtained from the aforementioned camera, A card distributing device that determines and displays the win / loss results of each game on the aforementioned game table, The system includes a control device that, in each game, uses the analysis results of the image analysis device to identify the positions of the chips placed on the game table by the game participants, and further uses the win / loss results to determine the winners and losers among the participants in each game. The control device further 1) In each game, check for any movement of chips between the start of card drawing or the dealer's game start operation and the display of the game's outcome by the card distribution device. 2) After each game has ended, while the dealer is collecting the chips bet by the losing players, check to see if there have been any chip movements by anyone other than the dealer. 3) After each game ends, while the dealer is collecting the chips bet by the losers among the game participants, check whether any chips have been added. 4) After each game has finished, the dealer has checked whether he has paid out the chips that the winners among the game participants had placed. 5) After each game has finished, the winner among the game participants must determine whether they have taken the chips they wagered and the chips they have received. It includes a function to determine at least one of the following.

[0079] Furthermore, the control device may be configured to determine at least one of 1) to 5) above by using the analysis results of the image analysis device to detect the hand movements of the dealer and the game participants, the movement of the chips, or the hand movements and the movement of the chips.

[0080] Furthermore, the control device may be configured to determine whether the amount of chips paid to the winner by the dealer is correct based on the amount wagered by the winner among the game participants.

[0081] Furthermore, the game fraud detection system may further include a monitor or lamp that displays a warning or indicator based on the determination result.

[0082] According to the fraud detection system of this embodiment, in baccarat and blackjack, it is possible to detect the timing of bets placed by customers and the settlement of bets by dealers to customers (game participants), as well as who placed or took the chips. By detecting these mistakes and fraudulent activities, the system can issue warnings or displays them, leading to measures to prevent recurrence.

[0083] Before proceeding to a detailed description of this embodiment, we will explain the flow of the baccarat game played in amusement facilities such as casinos. In the second embodiment, components similar to those in the first embodiment will be described using the same numbers.

[0084] As shown in Figure 7, at the game table 4, the customer (game participant / player) 6 sits in the playing position (chair) 7 facing the dealer 5. The customer (game participant) 6 then places chips 9 in the betting area 8 in front of them to bet on whether the baccarat game will end with the player (PLAYER) or the banker (BANKER), or whether it will be a tie (TIE) (hereinafter referred to as "betting"). The dealer 5 then waits for the right moment to signal "No" to end the betting by the customer (game participant) 6. The dealer calls out "More Bet" and makes a gesture such as moving their hand sideways (as shown in Figure 7). In baccarat, once "No More Bet" is called and card drawing begins or the dealer 5 starts the game, and before the card distribution device 3 displays the result of the game, the player 6 cannot move chips, place additional bets, or retrieve chips that have already been bet.

[0085] Next, playing cards 1 are drawn one by one from the card distribution device 3 with their backs facing up onto the game table 4. Initially, four cards are drawn, and as shown in the circles 1-4 in Figure 8, the first card goes to the player, the second to the banker, the third to the player, and the fourth to the banker. These are then distributed and placed in the area 10 (player area 10P and banker area 10B) on the game table 4, as seen from the dealer 5. Based on the rank (number) of the first four cards 1 through 4 and the conditions in the detailed rules of the baccarat game, the dealer 5 draws a fifth card 1, and then a sixth card 1, which then go to either the player or the banker. The winner of the game is then determined based on the rank (number) of the first four cards 1 through 4 (and possibly the fifth and sixth cards as well) and the detailed rules of the baccarat game. Here, the card distributing device 3 has the game rules programmed into it in advance, and is structured to read the information of the distributed card 1 (rank (number) and suit) and determine the winner or loser of the game. The win / loss determination result (win / loss result) determined by the card distributing device 3 is then checked to see if it matches the win / loss result determined by the dealer or the like, as described above.

[0086] The following describes the overall outline of the fraud detection system for games in a gaming arcade according to an embodiment of the present invention. Figure 7 is a diagram showing the overall outline of the system, which includes a game recording device 11 that records the progress of games played at a gaming table 4 as video footage via a camera 2, including the customers (game participants) 6 and the dealer 5; an image analysis device 12 that performs image analysis on the recorded video footage of the game's progress; and a card distribution device 3 that has the function of determining and displaying the win / loss result of each game at the gaming table 4. The card distribution device 3 is a so-called electronic shoe, already used by those skilled in the art, and has the rules of the game pre-programmed into it. It is structured to detect the timing when cards 1 are distributed by the dealer 5 at the beginning of each game, and to read the information (rank (number) and suit) of each card 1 that is distributed, in order to determine the win or loss of the game. For example, in a baccarat game, the banker win, the player win, and a tie are basically determined by the rank of 2-3 cards, and the determination result (win / loss result) is displayed by an indicator lamp 13.

[0087] The control device 14 of this fraud detection system is equipped with a chip detection function that uses the analysis results of the image analysis device 12 to identify in each game whether the customer 6 (game participant) placed a chip 9 in the player's side or the banker's side betting area 8 on the game table 4. The position and total amount of the chips 9 (whether the chips 9 were placed in the player's side or the banker's side betting area 8) are not normally readable, for example, if the chips 9 are shifted or overlapping, or if they are in a blind spot from the camera 2's position. The control device 14 is configured to accurately determine the position and number of chips by recognizing the blind spots and other concealment of the chips 9 (cases where part of a chip is hidden, or where the entire chip is hidden) using existing artificial intelligence-based computers or control systems, or self-learning functions based on deep learning (structure) technology. Furthermore, the structure for detecting the position and type of chips 9 in the betting area 8 is not limited to this, and for example, it may be configured to detect by reading the ID embedded in the chip.

[0088] As described above, the control device 14 can determine the position (position of bets on player, banker, or pair), type (each color of chip 9 is assigned a different value), and number of chips placed by each player 6 via the camera 2 and image analysis device 12. It can also detect which player 6 placed a bet on player (or, if multiple players 6 placed a bet on player, which player 6 placed the highest bet), and which player 6 placed a bet on banker (or, if multiple players 6 placed a bet on banker, which player 6 placed the highest bet). The image analysis device 12 and control device 14 in this fraud detection system have a structure that combines a computer, program, and memory, either as a single unit or as a set of multiple components.

[0089] The control device 14 is structured to determine whether the rank and suit information obtained by the image analysis device 12 from the video (using camera 2) of each card 1 distributed at the game table 4 matches the rank and suit information read by the card distribution device 3. In each game, the control device 14 determines whether the collection of losing chips 9 bet by the customer (game participant) 6 and the payment of winning chips to the winning customer (game participant) 6 were carried out appropriately according to the game's outcome determined by the card distribution device 3, by analyzing the video of the game's progress via the image analysis device 12.

[0090] The control device 14, as a characteristic function of the present invention, has the following functions 1) to 5) in accordance with the rules of the baccarat game and determines whether or not any cheating in violation of the rules has occurred. That is, 1) In each game, from the moment the card distribution device 3 signals the start of card dispensing, or from the moment the dealer 5 presses the start button 4s to initiate the game, until the card distribution device 3 displays the game's win / loss result, the camera 2 is used to monitor for any movement of the chips 9 using information obtained by the image analysis device 12 (as shown in Figure 8). 2) After each game ends, while the dealer 5 collects the chips 9 that the loser among the game participants 6 had bet (as shown in Figure 9), the camera 2 is used to monitor whether the loser 6 has taken the chips 9 illegally, using information obtained by the image analysis device 12. 3) After each game has ended, while Dealer 5 is collecting the chips 9 that the losers among the game participants had bet, the camera 2 is used to monitor, using information obtained by the image analysis device 12, whether anyone other than Dealer 5 (winners or losers) has added winning chips 9W or placed new chips 9 on the winning side that they had not bet on. 4) After each game ends, the image analysis device 12 uses camera 2 to monitor whether the dealer 5 has correctly placed the payment chip 9W in the position where the winning player 6 had placed their bet (as shown in Figure 10). 5) After each game has finished (the dealer 5 operates the card distribution device 3 to display the win / loss result on the display lamp 13), the image analysis device 12 uses camera 2 to monitor whether the winner 6W among the game participants 6 has taken the chips they bet 9 and the chips 9W that were paid out (as shown in Figure 11).

[0091] The control device 14 uses the camera 2 to analyze the information obtained by the image analysis device 12 as follows. Specifically, it uses the analysis results of the image analysis device 12 to monitor items 1) to 5) above by detecting the hand movements of the dealer 5 and the game participants 6, the movement of the chips, or the hand movements and chip movements. However, in the basic analysis, it is necessary to know who took the chips 9. The method of this analysis will be explained below using Figures 12 to 14.

[0092] Analysis of how Dealer 5 took the 9 chips bet by player 6L (Figure 12). The chips 9 bet by participant 6L, who lost the game, are collected by dealer 5. Whether or not these chips have been collected is monitored by analyzing the information obtained by the image analysis device 12 using camera 2. First, the change from the state in which the bet chips 9 are present (Figure 12A) to the state in which they are absent (Figure 12C) is detected by image analysis. Then, the image between the state in which chips 9 are present and the state in which they are absent (Figure 12B) is analyzed. In the image between the state in which chips 9 are present and the state in which they are absent (Figure 12B), the direction from which hand 5h is extending (from above in Figure 12 or elsewhere) is analyzed, and if it is extending from above (or if the hand appears from above and then exits upwards), it is determined that hand 5h belongs to dealer 5, and if the hand extends from any other direction, it is determined to be fraudulent. Fraud is detected based on this rule.

[0093] While the dealer 5 is collecting the 9 chips bet by participant 6L who lost the game, the system monitors whether anyone else is taking the lost chips (Figures 12 and 11). In the images between the state where chips 9 are present and the state where they are absent, the analysis of whether the loser 6L, etc., among the participants 6 took the chips, as shown in Figure 13, detects through image analysis that hand 6h extends or moves from below (originally from above) in Figure 13. This is determined to mean that a hand other than dealer 5, such as 6h, took the chips 9, and this is judged to be fraud.

[0094] Analysis of a situation where dealer 5 correctly pays (places) chip 9W for a winning chip of 9, and the winner 6W among the game participants 6 takes it. First, as shown in Figure 14A, chip 9W is redeemed for the winning chip, as shown in Figure 14B according to the game rules. The change from the state shown in Figure 14A to the state shown in Figure 14B is detected, and at the same time, image analysis is used to determine whether the hand is dealer 5's hand 5h. After this, as shown in Figure 14C, the winner 6W's hand 6h extends (moves) to the same betting area, and then the control device 14 checks, based on the image analysis results, whether all chips 9 are gone (state shown in Figure 14D), according to the game rules, and determines whether or not there was any fraud.

[0095] Furthermore, the control device 14 is configured to determine whether the amount of chips paid to the winner by the dealer 5 is correct based on the amount bet by the winner 6W among the game participants 6. A specific example is given below. The position and total amount of chips 9 (whether chips 9 were bet on the player's side or the banker's side betting area 8) are not normally readable, for example, if the chips 9 are shifted and overlapping, or if they are in a blind spot from the camera 2's position. The control device 14 is configured to accurately determine the position and number of chips by recognizing the blind spots of chips 9 (cases where part of a chip is hidden, or where the entire chip is hidden) using existing artificial intelligence-based computers or control systems, or self-learning functions based on deep learning (structure) technology. Furthermore, the structure for detecting the position 8 and type of chips 9 in the betting area 8 is not limited to this, and for example, it may be configured to detect by reading the ID embedded in the chip.

[0096] As described above, the control device 14 can determine the position 8 of the chips 9 bet by each player 6 (the position where they bet on player, banker, or pair), the type (each color of chip 9 is assigned a different value), and the number of chips via the camera 2 and the image analysis device 12. It can also detect which customer 6 bet on player (if there are multiple customers 6 who bet on player, which customer 6 bet the highest amount), and which customer 6 bet on banker (if there are multiple customers 6 who bet on banker, which customer 6 bet the highest amount).

[0097] Furthermore, the control device 14 of the game's fraud detection system analyzes and monitors the information obtained by the image analysis device 12 using the camera 2 in accordance with the rules of the baccarat game. It performs the monitoring described in 1) to 5) above and determines whether or not fraud is being committed in violation of the rules. When fraud is detected, the card distribution detection device 14C lights up the abnormality indicator lamps 16 installed on both the card distribution device 3 and the game table 4, respectively, and outputs the detection of fraud wirelessly or via wired connection 15 to the casino management department, etc. A monitor or lamp that displays a warning or indication based on the determination result may be provided in another location.

[0098] As described above, fraudulent activity is detected by the control device 14, and at the time of detection or at an appropriate time, it emits a display signal to the indicator lamp 13 and error indicator lamp 16 of the card distribution device 3. In addition to issuing a warning, it may also activate a function of the card distribution device 3 that prevents the distribution of cards after the time fraud or error is detected, thereby preventing the distribution of card 1.

[0099] Below, an embodiment of the card distribution device 3 used in the table game system of the present invention will be described with reference to Figures 15 to 19. The card distribution device 3 includes a card storage section 102 for storing multiple shuffled playing cards 1s, a card guide section 105 for guiding the shuffled playing cards 1 when a dealer 5 or the like manually pulls them out one by one from the card storage section 102 toward the game table 4, an opening 106 for removing the card 1 guided by the card guide section 105, a card detection section (card detection sensors 22 and 23) for detecting when a shuffled playing card 1 is pulled out, and a card reading section 10 for reading information representing at least the number (rank) of the shuffled playing cards 1. The system includes 8, a control unit 109 that determines the outcome of the card game based on the number (rank) of shuffled playing cards 1 sequentially read by the card reader 108, a result display lamp 13 that displays the win / loss result determined by the control unit 109, a distribution restriction device 30 provided in the opening 106 that restricts the entry and exit of cards 1 from the card storage unit 102, and a management control unit 114 that has the same function as the control device 14, all of which are integrated, and the control device 14 has the function of preventing further cards from being drawn from the card distribution device 3 if a dealer error or fraudulent activity in the game is detected, either from the time of detection or at a predetermined timing.

[0100] Next, the distribution restriction device 30, which restricts the entry and exit of cards 1 from the card storage section 102, will be explained using Figures 17 and 18. The distribution restriction device 30 is provided on the card guide 107 of the card guide section 105, which guides cards 1, taken out one by one from the opening 106 in front of the card storage section 102, onto the game table 4. The distribution restriction device 30 has a structure in which a locking member 34 presses the card 1 when it passes through the slot 33 between the card guide section 105 and the guide cover of the card guide 107, thereby preventing the card 1 from entering or exiting the slot 33. The locking member 34 moves as shown by arrow m, driven by a drive unit 35 such as an electromagnetic solenoid or piezoelectric element, so that it can take two states: a position in which the card 1 is pressed (restriction position) and a passable position in which the card 1 can pass through. The drive unit 35 is controlled by a control unit 109, which is directly or indirectly connected to the control device 14 by wire or wireless, and moves the locking member 34 to two states: a position that presses against the card 1 and a passable position that allows the card 1 to pass through. The control unit 109 has the rules of the baccarat game pre-programmed and stored in it.

[0101] Next, a modified example of the distribution restriction device 30 will be explained with reference to Figure 18B. In the modified example, the distribution restriction device 40 has a structure in which a locking member 36 protrudes into the slot 33 when the card 1 passes through the slot 33 between the card guide portion 105 and the card guide 107 (guide cover) to prevent the movement of the card 1. The locking member 36 moves as shown by arrow m, so that it can take on two states: a position that prevents the movement of the card 1 (restriction position) and a passable position that allows the card 1 to pass, by a drive unit 37 such as an electromagnetic solenoid or a piezoelectric element. The drive unit 37 is controlled by a control unit 109 connected to a control device 14, which moves the locking member 36 to the two states: a position that prevents the movement of the card 1 and a passable position that allows the card 1 to pass.

[0102] Next, we will explain in detail the code reading unit 108, which reads a code 52 representing the number (rank) of card 1 from card 1 when card 1 is manually pulled out from card storage unit 102. Figure 17 is a plan view of the main part of the card distribution device 3. In the figure, the code reading unit 108 is provided in the card guide unit 105, which guides cards 1, which are manually pulled out one by one from the opening 106 in front of the card storage unit 102, onto the game table 4. The card guide unit 105 is an inclined surface, and card guides 107, which also serve as sensor covers, are attached to the edges on both sides. In addition, each of the two card guides 107 can be attached and detached with screws or the like (not shown). When the card guides 107 are removed, the sensor group 115 of the code reading unit 108 is exposed. The sensor group 115 consists of four sensors: two ultraviolet reaction sensors (UV sensors) 20, 21 and object detection sensors 22, 23.

[0103] The object detection sensors 22 and 23 are optical fiber type sensors that detect the presence or absence of card 1 and can detect the movement of card 1. Object detection sensor 22 is located on the upstream side of the card guide section 105, along the flow direction of card 1, while the other object detection sensor 23 is located on the downstream side. As shown in the figure, both object detection sensors 22 and 23 are provided on the upstream and downstream sides of the UV sensors 20 and 21. The UV sensors 20 and 21 are equipped with an LED (ultraviolet LED) that emits ultraviolet light and a detector. Card 1 has the mark M of code 52 printed on it using ultraviolet light-emitting ink that changes color when exposed to ultraviolet light. When ultraviolet light (black light) is shone on card 1, the reflected light of the mark M of code 52 on card 1 is detected by the detector. The UV sensors 20 and 21 are connected to the code reading section 108 and the control section 109 via cables. In the code reading unit 108, the combination of marks M is determined from the output signals of the UV sensors 20 and 21, and the number (rank) corresponding to each code 52 is determined.

[0104] The code reading unit 108 controls the start and end of reading from the UV sensors 20 and 21 based on the detection signals from the object detection sensors 22 and 23, with the control unit 109 controlling this. The control unit 109 also determines whether card 1 has passed through the card guide unit 105 normally, based on the detection signals from the object detection sensors 22 and 23. As shown in Figure 19, rectangular marks M representing the rank (number) and suit (hearts, spades, etc.) of the card are arranged in two columns and four rows around the edge of card 1. When the UV sensors 20 and 21 detect a mark M, they output an ON signal. The code reading unit 108 determines the relative relationship between the two signals input from the two UV sensors 20 and 21. As a result, the code reading unit 108 identifies the code based on the relative differences between the two marks M detected by the two UV sensors 20 and 21, and identifies the corresponding number (rank) and type (suit) of card 1.

[0105] Figure 19 shows the relationship between code 52 and the ON signal output of the two UV sensors 20 and 21. Based on the comparison of the relative changes in the ON signal output of the UV sensors 20 and 21, a predetermined combination of mark M can be identified. As a result, there are four combinations of mark M in the upper and lower two columns, and when these are printed in four columns, 4 to the power of 4 = 256 types of codes can be constructed. Each of the 52 types of playing cards is assigned to one of the 256 types of codes, and this is stored in memory or program as a reference table. The code reading unit 108 identifies each code 52, and the number (rank) and type (suit) of card 1 are identified from a predetermined reference table (not shown). Furthermore, since the 256 types of codes can be associated with the 52 types of cards in any combination and stored in the reference table, the combinations can be made complex, and the combination of the 256 types of codes and the 52 types of cards can be changed depending on the time and place. The codes are printed with paint that becomes visible when exposed to ultraviolet light, and it is desirable that they are printed in a position that does not overlap with the card type indication or index 103.

[0106] In the above embodiment, the image analysis device 12 and the control device 14 were devices that utilized artificial intelligence or had a deep learning structure. Specifically, the image analysis device 12 and the control device 14 may use a scale-invariant feature transform (SIFT) algorithm, a convolutional neural network (CNN), deep learning, machine learning, or similar methods to perform image analysis and the various controls described above. These technologies are techniques for recognizing objects contained in captured images by performing image recognition on the captured image. In particular, in recent years, deep learning techniques using multi-layered neural networks have been used to recognize objects with high accuracy. This deep learning technique generally recognizes objects with high accuracy by stacking multiple layers in the intermediate layers between the input and output layers of the neural network. In this deep learning technique, convolutional neural networks, in particular, have attracted attention because they have higher performance than conventional methods of recognizing objects based on image features.

[0107] In a convolutional neural network, labeled target images are trained to recognize the main object contained within those images. If multiple main objects exist within a training image, a region rectangle is used to specify it, and the corresponding image is labeled for training. Furthermore, a convolutional neural network can also determine the main object within an image and its position.

[0108] To further explain convolutional neural networks, the recognition process involves extracting candidate regions based on local features by performing edge extraction and other operations on the target image. These candidate regions are then input into a convolutional neural network to extract feature vectors, which are then classified. The candidate region with the highest confidence level is then recognized as the result. Confidence level is a measure of how much higher the similarity between a given image region and the subject of the image (as learned along with its label) is relative to the similarity between other classes.

[0109] Furthermore, devices utilizing artificial intelligence or having a deep learning structure are described in U.S. Patent No. 9361577, U.S. Patent Publication No. 2016-171336, U.S. Patent Publication No. 2015-036920, and Japanese Patent Publication No. 2016-110232, etc., and these descriptions are incorporated herein by reference.

[0110] Although various embodiments of the present invention have been described above, it goes without saying that the above embodiments can be modified by those skilled in the art within the scope of the present invention, and the apparatus of these embodiments may be appropriately modified as needed for the game to which they are applied. [Explanation of Symbols]

[0111] 1 Playing Card 1s Multiple shuffled playing cards 2 Surveillance cameras 3. Card distribution device 4 Gaming Tables 5 Dealer 6. Players (Game Participants / Players) 7 chairs 8 Betting Areas 9 chips 10 areas 10P Player Area 10B Bunker Area 11 Game recording device 12 Image analysis device 13. Result display lamp 14 Control device 14C Card Distribution Detection Device 15 Output (abnormality judgment results, etc.) 16. Error indicator lamp 30 Distribution restriction device 33 slots 34 Locking member 35 Drive unit 36 Locking member 37 Drive unit 40 Distribution restriction device 102 Card storage compartment 103 Index 105 Card Guide Section 106 Opening 107 Card Guide 109 Control Unit 112 Side monitor

Claims

1. A chip recognition system used in a game table having a chip tray for holding dealer chips and a table surface on which chips are placed in the form of a chip stack, A camera generates multiple images by photographing each of the multiple chip stacks placed on the game table from an oblique angle above, wherein the game table has multiple betting areas at different distances from the camera, and chips can be placed in any of the multiple betting areas, so that the distance from the camera may differ between the chip stacks shown in one image and the chip stacks shown in another image, and so that the size of the chip stacks in those images may differ. A processor that uses artificial intelligence technology or deep learning technology to analyze the multiple images, recognizes each chip stack appearing in the multiple images, and determines the number of chips in each recognized chip stack, A chip ID reader that reads the chip ID of the chip held in the chip tray, Equipped with, The chip recognition system includes at least one processor which distinguishes a chip stack placed by a dealer from a chip stack placed by a player, and determines the number of chips in each of the chip stacks.

2. The chip recognition system according to claim 1, wherein the at least one processor determines the position of the chip stack.

3. The chip recognition system according to claim 1, wherein the at least one processor recognizes the type of chip in the chip stack.

4. The chip recognition system according to claim 3, wherein the at least one processor determines the amount of each chip stack in the plurality of images based on the recognized type.

5. The chip recognition system according to claim 1, wherein the chip ID reader determines the type of chip held in the chip tray based on the read chip ID.

6. The aforementioned chip has a chip ID via RFID. The chip recognition system according to claim 1, wherein the chip ID reader reads the RFID of the chip held in the chip tray.

7. The chip recognition system according to claim 1, wherein the at least one processor performs image analysis of an image captured by the camera according to a predetermined trigger.

8. The chip recognition system according to claim 1, wherein the camera is configured to photograph the chip placed on the table surface by the dealer.

9. The chip recognition system according to claim 1, wherein the camera photographs the chip placed on the table surface by the player.

10. The chip recognition system according to claim 1, wherein the at least one processor recognizes the chip stack using the artificial intelligence technology or the deep learning technology even when a portion of the chip is obscured by a blind spot of the camera.

11. The chip recognition system according to claim 1, wherein the at least one processor recognizes the chip stack using the artificial intelligence technology or the deep learning technology, even when the chip stack is composed of a plurality of chips stacked in a staggered manner.

Images