DETECTING AND LINKING ANOMALOUS GAMING EVENTS VIA ANONYMOUS BIOMETRIC IDENTIFIERS
A method and system detect and link anomalous gaming events. A processor, using shuffler sensors, detects a first anomaly on a first card from a first game instance, the anomaly varying from prior card images. In response, a machine learning model analyzes image data from a first gaming table's image sensors to determine unique biometric facial features of a first participant. A first anonymous identifier is generated from these features and stored, linked to first anomaly data. Subsequently, a second anomaly on a second card in a second game instance is detected by shuffler sensors. Image data from a second gaming table is analyzed by the machine learning model to extract unique biometric facial features of a second participant. It is then determined that the second participant's facial features mathematically correspond to the stored first anonymous identifier. A counter associated with the first anonymous identifier is incremented, indicating multiple linked anomaly events.
1 . A method including:
detecting, by a processor using sensors of one or more shufflers in a shuffler network, a first anomaly on a first card associated with a first game instance, wherein the first anomaly varies from one or more previously taken images of the first card;
in response to detecting the first anomaly, determining, via analysis by a machine learning model of first image data captured by one or more image sensors at a first gaming table associated with the first game instance, unique biometric facial features of a first participant associated with the first game instance;
generating a first anonymous identifier mathematically derived from the unique biometric facial features of the first participant;
storing the first anonymous identifier linked to data describing the first anomaly;
subsequently detecting, by the processor using sensors of the one or more shufflers in the shuffler network, a second anomaly on a second card associated with a second game instance at a potentially different location or time, wherein the second anomaly varies from one or more previously taken images of the second card;
in response to detecting the second anomaly, analyzing second image data via the machine learning model captured by one or more image sensors at a second gaming table associated with the second game instance to extract unique biometric facial features of a second participant associated with the second game instance;
determining that the unique biometric facial features of the second participant correspond mathematically to the stored first anonymous identifier; and
incrementing a counter associated with the first anonymous identifier indicating multiple linked anomaly events.
2 . The method of claim 1 , wherein the first anomaly and the second anomaly are detected on cards of high value.
3 . The method of claim 1 , wherein the one or more previously taken images indicate an image of an original manufactured appearance of the respective card.
4 . The method of claim 1 , wherein the one or more previously taken images indicate an appearance of the respective card when previously shuffled by one of the one or more shufflers.
5 . The method of claim 1 , wherein the first anomaly or the second anomaly includes one or more indentations having a matching arc shape.
6 . The method of claim 1 , wherein the data describing the first anomaly includes at least one of a type of anomaly, a location of the anomaly on the first card, an image of the first anomaly, or a time of detection.
7 . The method of claim 1 , further including adjusting a collusion-confidence score associated with the first anonymous identifier in response to incrementing the counter.
8 . The method of claim 1 , wherein the first gaming table and the second gaming table are the same gaming table.
9 . The method of claim 1 , wherein the first gaming table and the second gaming table are different gaming tables.
10 . A system including:
one or more shufflers in a shuffler network, each shuffler having one or more shuffler sensors;
one or more image sensors associated with one or more gaming tables;
a processor communicatively coupled to the one or more shufflers and the one or more image sensors; and
a memory storing instructions that, when executed by the processor, cause the system to perform operations to:
detect, using the shuffler sensors, a first anomaly on a first card associated with a first game instance, wherein the first anomaly varies from one or more previously taken images of the first card;
in response to detection of the first anomaly, determine, via analysis by a machine learning model of first image data captured by the one or more image sensors at a first gaming table associated with the first game instance, unique biometric facial features of a first participant associated with the first game instance;
generate a first anonymous identifier mathematically derived from the unique biometric facial features of the first participant;
store in the memory the first anonymous identifier linked to data describing the first anomaly;
subsequently detect, using the shuffler sensors, a second anomaly on a second card associated with a second game instance at a potentially different location or time, wherein the second anomaly varies from one or more previously taken images of the second card;
in response to detecting the second anomaly, analyze second image data via the machine learning model captured by the one or more image sensors at a second gaming table associated with the second game instance to extract unique biometric facial features of a second participant associated with the second game instance;
determine that the unique biometric facial features of the second participant correspond mathematically to the stored first anonymous identifier; and
increment a counter associated with the first anonymous identifier, the counter stored in the memory, indicating multiple linked anomaly events.
11 . The system of claim 10 , wherein the first anomaly and the second anomaly are detected on cards of high value.
12 . The system of claim 10 , wherein the one or more previously taken images indicate an image of an original manufactured appearance of the respective card.
13 . The system of claim 10 , wherein the one or more previously taken images indicate an appearance of the respective card when previously shuffled by one of the one or more shufflers.
14 . The system of claim 10 , wherein the first anomaly or the second anomaly includes one or more indentations having a matching arc shape detected by the shuffler sensors.
15 . The system of claim 10 , wherein the data describing the first anomaly includes at least one of a type of anomaly, a location of the anomaly on the first card, an image of the first anomaly, or a time of detection.
16 . The system of claim 10 , wherein the operations further include adjusting a collusion-confidence score associated with the first anonymous identifier in response to incrementing the counter.
17 . The system of claim 10 , wherein the first gaming table and the second gaming table are the same physical gaming table.
18 . The system of claim 10 , wherein the first gaming table and the second gaming table are different physical gaming tables.
19 . The system of claim 10 , wherein the machine learning model is stored in the memory and executed by the processor.