Device, computer program and method for determining player's handball from images

Abstract

Disclosed are an apparatus, computer program, and method for determining a player's handball from an image. A method for determining a player's handball from an image is described. The method includes receiving an image from a camera, the image including a player's arm or hand and a ball; determining the player's posture when the ball makes contact with the player's arm or hand; comparing the player's posture with a predetermined posture; and determining that a handball event has occurred based on the comparison.

Description

Technical Field

[0001] The present technology relates to devices, computer programs, and methods. Background Art

[0002] The "background" description provided herein is for the purpose of generally presenting the context of the present disclosure. To the extent described in the background section, the work of the presently named inventors and aspects of the description that would not otherwise be considered prior art at the time of filing are neither explicitly nor implicitly admitted to be prior art to the present technology.

[0003] Video Assisted Referee (VAR) has been implemented in many sports events. VAR helps referees make correct decisions in sports events. For example, in football matches, VAR is sometimes used to help identify handballs, offsides, shots, fouls, etc.

[0004] As mentioned above, one of the decisions that VAR assists with is the handball determination. The International Football Association Board (IFAB) introduced a change to the handball rule in football so that a handball is called if a player uses his / her arms to bring his / her body into an "unnatural contour" when the ball strikes his / her arm or hand.

[0005] The purpose of this disclosure is to improve handball detection using VAR. Summary of the Invention

[0006] According to an embodiment, a method for determining a handball committed by a player from an image is provided, comprising: receiving an image from a camera, the image including the player's arm or hand and a ball; determining the player's posture when the ball contacts the player's arm or hand; comparing the player's posture with a predetermined posture; and based on the comparison, determining that a handball event has occurred.

[0007] The foregoing paragraphs have been provided by way of general introduction and are not intended to limit the scope of the following claims.The described embodiments, together with other advantages, may be best understood by reference to the following detailed description taken in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] A more complete appreciation of the present disclosure and many of its attendant advantages will become more readily understood from the following detailed description when considered in conjunction with the accompanying drawings, in which:

[0009] Figure 1 shows a plan view of a sporting event being held at a sports venue with cameras positioned at different locations around the sports venue;

[0010] Figure 2 A data structure is shown that stores a camera identifier in association with a camera position and a camera type;

[0011] Figure 3 An apparatus according to an embodiment of the present disclosure is shown;

[0012] Figures 4A to 4D shows a soccer player's stance according to an embodiment; and

[0013] Figure 5 A flow chart according to an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0014] Reference is now made to the drawings, wherein like reference numerals designate identical or corresponding parts throughout the several views.

[0015] Figure 1 A plan view of a sporting event taking place at a sports venue is shown with cameras positioned at different locations around the venue. Figure 1 In the embodiment of the present invention, the sports venue is a soccer stadium and the sports event is a soccer match. However, it should be understood that the present disclosure is not so limited and any sports event at any sports venue is contemplated. For example, the sports event may be a rugby match at a rugby stadium, an American football match at an American football stadium, a baseball game at a baseball stadium, etc.

[0016] In sports venue 100, multiple cameras 105A-105H are positioned around the court. Each of the multiple cameras 105A-105H is positioned at a different location around the court and captures a different view of the court. In other words, each of the multiple cameras is located at a different geographic or physical location around the court. In one embodiment, the court is surrounded by the multiple cameras. However, the present disclosure is not limited thereto, and the cameras may be positioned around a smaller area of the court than the court.

[0017] Each camera 105A-105H generates a camera feed, which is an image or image stream of the field of view captured by the camera. These camera feeds can be broadcast video feeds sent to viewers watching the sporting event at home on television or the internet, or can be used to film the event for use on displays within the stadium. The cameras 105A-105H can be specifically positioned to assist referees in video assisted refereeing (VAR) scenarios. Of course, the cameras can have a mix of uses, such as television cameras and VAR cameras.

[0018] although Figure 1 It is not explicitly shown in the figure, however, that the camera feed generated by each of the plurality of cameras 105A-105H located at different locations on the sports field is fed back to the apparatus according to an embodiment of the present disclosure. Figure 3This device will be described.

[0019] It is obvious to those skilled in the art that Figure 1 In the description of FIG, eight cameras located at different positions around the sports field are provided. Of course, the present disclosure is not limited thereto and any number of cameras can be envisioned.

[0020] exist Figure 1 In an embodiment, each camera 105A-105H is located at different locations around the stadium and has different viewing angles of the sporting event being broadcast on the stadium's pitch. Because each camera 105A-105H is located at different locations on the stadium, each camera has a different field of view of the sporting event. In embodiments, this means that each camera 105A-105H has a different field of view of the sporting event. Furthermore, each camera 105A-105H may have a different focal length or lens arrangement compared to other cameras. This means that different cameras can capture different angles of the event on the pitch. For example, in embodiments, camera 105D is located on a broadcast stand. This means that camera 105D captures the entire pitch within its field of view. However, due to the wide field of view, this means that it is difficult for camera 105D to capture detailed images of events occurring in the penalty box on the pitch.

[0021] In an embodiment, camera 105B is located behind the east goal. Camera 105B may be a camera with a narrow field of view and a short focal length, such as in the east penalty area. This means that camera 105B may capture detailed images of events occurring in the east penalty area, but may not capture events occurring in the west penalty area.

[0022] Accordingly, each camera 105A-105H will have a different view of the court due to the different positions of each camera. This means that each camera 105A-105H will have a different view of any event occurring on the court. Figure 2 As described, each camera may have a different focal length or a different lens arrangement, ie each camera will have a different field of view of any event. The different focal lengths or different lens arrangements are embodiments of different features or types of cameras.

[0023] refer to Figure 1 , the sports event is a football match. The first team, shown by the solid line, is playing against the second team, shown by the dotted line. Figure 1 In FIG, several athletes of each team are shown. Specifically, the first team includes athletes 110A-110H and the second team includes athletes 115A-115C. Figure 1A ball 120 is shown in FIG.

[0024] In the embodiments of the present disclosure, Figure 1 In an embodiment of the present disclosure, an event has occurred in the sporting event shown in FIG. Figure 1 The event represented in FIG is a handball by player 110G of the first team. Figure 1 The event shown in is a handball, however, any type of event is contemplated such as a player being offside, a goal being scored, a foul, etc. Of course, other events are contemplated where the sporting event is not football. Therefore, an event is defined as an event that requires an operator to review the captured video footage. In an embodiment, using Figure 3 The device shown in , ie, for example, a VAR device.

[0025] exist Figure 1 In the embodiment, the incident occurs near the east penalty area on the court.

[0026] In an embodiment, the court may be divided into quadrants. Figure 1 These quadrants are shown as Quadrants 1-4 in FIG. Specifically, an event occurs in Quadrant 2 of the pitch. As will become apparent later, this allows for the approximate location of the event on the pitch to be defined. Accordingly, the pitch can be divided in any manner and may include dividing the pitch into 2, 8, 16, or more zones. Generally, the more zones there are, the more precisely the location of the event can be defined. Furthermore, the shape of the zones can vary. For example, in a zone where a greater number of events occur relative to other zones of the pitch (e.g., in the penalty area), that zone may have more zones than the rest of the pitch.

[0027] Now turn Figure 2 , Figure 2 The data structure for storing the camera identifier in association with the camera position and the camera type is shown. In the embodiment of the present disclosure, the data structure may be a database, etc. Of course, the present disclosure is not limited thereto.

[0028] In data structure 200, three camera identifiers are shown. These are shown in camera identifier column 205. Associated with each camera identifier is the location of the camera.

[0029] Although this is described as the location of a sporting event, the location of the camera may be GPS coordinates or some other unique location that identifies the location of the camera at the stadium. This location is shown in column 210. In other words, the geographic location of the camera 20 having the identifier shown in column 205 is shown in column 210.

[0030] Furthermore, the data structure 200 has a camera type column 215. The camera type column 215 stores the camera type having the camera identifier shown in the camera identifier column 205. In other words, the camera type column 215 shows the camera type associated with the camera identifier shown in the camera identifier column 205.

[0031] As described above, in embodiments of the present disclosure, the camera type refers to the field of view of the camera with the camera identifier in the camera identifier column 205. For example, the camera located on the "South Broadcast Pedestal" is a wide-angle camera type that displays the entire field. Because this camera has a field of view that covers the entire field, it is sometimes referred to as camera number 1 or the main camera.

[0032] In other embodiments, camera identifier 105B shows a camera positioned behind a goal on the east side of a sporting event. This type of camera is typically a close-up camera, i.e., it follows the movement of the ball or a particular player and its field of view includes the ball or a particular player.

[0033] In the third embodiment, the camera identified by camera identifier 105A is located on the northeast corner of the sports field. This camera has a wider field of view and is designed to capture the northeast quadrant ( Figure 1 As shown in column 215, this camera type is called a landscape camera.

[0034] Finally, a quadrant camera column 220 is shown. In the quadrant camera column 220, the quadrants of the field of view of the camera associated with the camera identifier 205 are marked. In other words, camera 105D has a field of view covering all quadrants, camera 105B has a field of view covering quadrants 2 and 4, and camera 105A has a field of view covering quadrant 2.

[0035] refer to Figure 3 , shows an apparatus 400 according to an embodiment of the present disclosure. The apparatus 400 includes a processing circuit 405. The processing circuit 405 is any type of circuit that controls the apparatus 400 to perform a method according to an embodiment of the present disclosure. Thus, the processing circuit 405 may be an application-specific integrated circuit (ASIC) or a microprocessor operating under the control of software code.

[0036] The software code may be a computer program stored on a memory 410 located within the apparatus 400 and connected to the processing circuit 405. The memory 410 may be any type of computer readable memory such as solid-state memory, magnetically readable memory, or optically readable memory.

[0037] The processing circuit 405 is also connected to the communication circuit 415. The communication circuit 415 can be any type of circuit that allows the device 400 according to an embodiment of the present disclosure to communicate with the cameras located at the sports venue. In other words, the communication circuit 415 allows the device 400 to communicate with the cameras located at the sports venue. Figure 1 The apparatus 400 may communicate with a plurality of cameras 105A-105H shown at the sports venue. Such communication may occur via a direct connection between the apparatus 400 and the cameras or via a network such as the Internet. The purpose of the communication circuit 415 is to receive an image or image feed (which may include multiple images) from the cameras located at the sports venue.

[0038] refer to Figure 4A , shows a pose graph 500 of a player in a soccer match. A pose graph is a diagram showing the player's poses. It is typically composed of a plurality of straight lines and is derived from an image of the player. The mechanism for obtaining pose graph 500 is known and is therefore not described for simplicity. However, it should be appreciated that one method of generating a pose graph is to obtain an image of the soccer player from a captured image and apply a model such as the PoseNet model.

[0039] As described above, a handball is determined if a player uses their arms to create an "unnatural contour" of their body when the ball strikes their arm or hand. Various techniques can be used to determine whether the ball strikes a player's hand or arm. In embodiments, the ball can be tracked and tracked at any one time, thereby tracking the movement of the ball on the field and thereby the position of the ball.

[0040] These ball tracking techniques can be used in conjunction with athlete tracking to determine the position of an athlete to determine when the ball hits an athlete's arm or hand. This is advantageous because ball tracking and athlete tracking techniques can infer the position of the ball and / or athlete between captured frames. This allows the handball to be accurately detected. Furthermore, using ball and athlete tracking enables an automatic mechanism to be implemented to determine when the ball hits an athlete's hand or arm. Of course, it should be appreciated that other methods of detecting whether a ball makes contact with an arm or hand are known.

[0041]

[0004] Despite the foregoing, techniques for detecting whether an athlete has caused his / her body to present "unnatural contours" are unknown.

[0042] Figure 4B FIG shows a posture diagram 510 according to an embodiment of the present disclosure. It should be appreciated that Figure 4B The postures performed by the athletes in Figure 4A. However, a natural silhouette pose 515 is also shown. The natural silhouette pose 515 is a border defining the natural silhouette of the athlete. In embodiments, the border defining the natural silhouette of the athlete may be opaque, but the natural silhouette pose 515 may be semi-transparent.

[0043] This allows the boundary and the player to be seen at the same time. This allows the operator to see the natural silhouette posture 515 superimposed on the player to determine whether the body is in an unnatural silhouette. The natural silhouette posture 515 defines the boundary of the natural silhouette of the player's arms and hands. Therefore, when the ball hits the player's arm or hand, if the player's posture is on or within the boundary, it is judged that the player has not caused his / her body to present an "unnatural silhouette" and a handball will not be determined. However, when the ball hits the player's arm or hand, if the player's posture is outside the boundary, it will be judged that the player has caused his / her body to present an "unnatural silhouette" and a handball will be determined. For ease of illustration, a posture diagram is shown for a front view of the player and a posture diagram is shown for a back view of the player. It should be recognized that Figure 4B The posture diagrams in are examples only and any shape or size of the boundary may be envisioned.

[0044] In an embodiment, a natural silhouette pose 515 can be generated for an athlete of a specific height and body type. In other words, a natural silhouette pose 515 can be defined for an athlete who is 1.80 meters tall and has a slender build. Thus, if the athletes in the image have different heights or different body types, the natural silhouette pose 515 can be modified accordingly. Thus, for example, if the athlete is 2.00 meters tall and has a muscular build, their natural silhouette pose 515 will be modified accordingly. The athlete's height and / or body type can be stored as metadata associated with the athlete. In fact, any metadata associated with the athlete can be used to adjust the natural silhouette pose. Thus, the natural silhouette pose can be modified based on the athlete's metadata.

[0045] refer to Figure 4C , the player's natural silhouette posture 515 is superimposed on the player's image at the moment the ball 120 hits the player's arm or hand. Figure 4CIn an exemplary embodiment of the present invention, the player's arm and hand of gesture graph 510 are within the boundaries defined by the player's natural outline posture 515. Therefore, even though the ball 120 hits the player's arm, because the player does not cause his or her body to present an unnatural contour (as defined by the gesture graph), a handball will not be determined. In an embodiment, the gesture graph is referred to as a predetermined gesture. This determination is made by comparing the player's gesture graph with the player's natural outline posture 515. If the gesture graph is within the boundaries of the player's natural outline posture 515 at the moment the ball hits the player's arm or hand, a handball will not be determined.

[0046] However, reference Figure 4D As can be seen from the athlete's posture graph 510, at the moment the ball impacted the athlete's arm or hand, at least one arm and / or hand was outside the athlete's natural contour posture 515. Accordingly, it is determined that the athlete presented an unnatural contour through his or her body at the moment the ball 120 impacted the athlete's arm or hand. Therefore, assuming that the ball made contact with the athlete's arm or hand at the moment of impact, and the arm and / or hand was outside the natural contour, a handball will be determined.

[0047] In an embodiment, if the free kick is outside of the natural outline posture 515 when the ball contacts the player's arm or hand, a handball determination will be given. Of course, the present disclosure is not so limited, and if a majority of the ball is outside of the natural outline posture 515 when the ball contacts the player's arm or hand, a handball determination will be given. By counting the number of pixels of the ball that are outside of the natural outline posture 515 and comparing these to the number of pixels of the ball that are on or within the natural outline posture 515, a determination of the majority of the ball can be made. If the number of pixels outside of the natural outline posture 515 is greater than the number of pixels that are on or within the natural outline posture 515 when the ball contacts the player's arm or hand, a handball determination will be given.

[0048] Although it has been described above that postures are multiple linear segments derived from software such as PoseNet, the present disclosure is not limited thereto. In other embodiments, a model of the athlete can be derived from an image. The model can be a 3D image that allows the athlete's natural outline posture 515 to be superimposed on the athlete's 3D image. The model can be generated using known techniques such as character binding, which is a known technique and causes the 3D model to present a specific posture based on the derived athlete's posture. Generating a 3D model allows the operator to view the event more clearly. Specifically, by viewing the event in 3D, the operator is allowed to view the event from all angles. This allows for making more accurate decisions about handball. Moreover, by viewing the event in 3D, the decision can be more clearly shown to the audience of the game. This helps to prove that the decision is correct. Another advantage of using a 3D model is that there is no camera with a field of view where the ball hits the athlete's arm and / or hand. This situation may occur when the camera's field of view is blocked by other athletes or the contact point on the hand and / or arm is on the other side of the athlete's body. In this case, the tracking system will provide the exact position of the athlete's hand and / or arm and the ball, and can show this position in the 3D model view. This enables the referee to make a decision about handball. In this case, the athlete's natural contour posture 515 is modeled into a 3D image and superimposed on the model. In other words, the athlete's posture and the predetermined posture are a 3D model. Again, when the ball hits a hand or arm, if the athlete's hand or arm is outside the athlete's natural contour posture 515, a handball will be determined. This will allow a more accurate representation of the natural contour.

[0049] The athlete's natural silhouette posture 515 can be generated based on the athlete's normal movements during a soccer game. For example, an athlete may position their arms in an unusual position when tackling, or put their arms in an "L" shape when running. This can change the shape of the athlete's natural silhouette posture 515 based on the athlete's movements. In other words, in an embodiment, the method includes selecting a predetermined posture based on the movement made by the athlete when the ball contacts the athlete's arm or hand. Accordingly, when checking whether a handball has occurred, multiple natural silhouette postures of the athlete can be generated and superimposed on the athlete's posture. This will provide a more accurate natural silhouette of the athlete.

[0050] In some implementations, when checking whether a handball has occurred, the various different natural postures of the player may be fused into a single natural profile player posture that may be applied to the player's posture regardless of the player's action.

[0051] In some embodiments, a player's natural silhouette gesture 515 can be generated through machine learning. Specifically, when making a handball decision, the player's natural silhouette gesture 515 can be derived from examining a plurality of different gestures performed by the soccer player during previous instances. In other words, in an embodiment, the following steps are provided: a predetermined gesture is generated using the player's hand or arm position during previous handball incidents. This allows decisions using VAR to be consistent with previous decisions made without VAR.

[0052] In one embodiment, known techniques can be used to determine the velocity of the ball when it strikes the player's arm or hand. This can be displayed to the operator and / or referee. The velocity can include both the speed and direction of the ball. In one embodiment, the speed or direction of the ball is displayed. The purpose of providing additional information is to provide context for the event, allowing the operator and / or referee to determine whether a handball has occurred. Other factors such as whether the ball deflected before striking the player's arm can be displayed. Furthermore, the distance between the deflection and the arm or hand can also be displayed. This additional information provides further context for the operator and referee and can help determine whether a handball has occurred.

[0053] In an embodiment, the additional information can be compared with previous similar events recorded in a database. The results of the previous events can also be displayed to the operator and / or referee to help the operator and / or referee determine whether a handball has occurred.

[0054] In an embodiment, the additional information can be used to automatically detect handballs. In this case, the additional information can be compared with previous similar events, and the result selected by the device is the same as the previous event with the most similar features. In this case, the additional information does not need to be displayed.

[0055] Various modifications and variations of the present disclosure are possible in light of the above teachings.It is therefore to be understood that within the scope of the appended claims, the present disclosure may be practiced otherwise than as specifically described herein.

[0056] Thus far, while the present disclosure has been described as an embodiment implemented at least in part by a data processing apparatus controlled by software, it should be recognized that non-transitory machine-readable media such as optical disks, magnetic disks, semiconductor memories, etc. carrying the software are also considered to represent embodiments of the present disclosure.

[0057] It will be appreciated that, for clarity, the above description has been described with reference to different functional units, circuits, and / or processors. However, it will be appreciated that any suitable distribution of functionality between different functional units, circuits, and / or processors may be used without departing from the embodiments.

[0058] refer to Figure 5 , illustrates the process performed by processing circuit 405. Process 700 begins at step 705. The process proceeds to step 710, where an image is received from a camera, the image including the player's arm or hand and the ball. The process then proceeds to step 715, where the player's posture when the ball contacted the player's arm or hand is determined. The process then proceeds to step 720, where the player's posture is compared to a predetermined posture. The process then proceeds to step 725, where a determination is made based on the comparison whether a handball event occurred. The process then ends at step 730.

[0059] The described embodiments can be realized in any suitable manner including hardware, software, firmware or any combination thereof. Alternatively, the described embodiments can be implemented at least in part as computer software running on one or more data processors and / or digital signal processors. The elements and components of any embodiment can be physically, functionally, and logically realized in any suitable manner. Indeed, functions can be realized in a single unit, multiple units, or as a part for other functional units. Therefore, the disclosed embodiments can be realized in a single unit or the disclosed embodiments can be physically and functionally distributed between different units, circuits, and / or processors.

[0060] Although the present disclosure has been described in conjunction with certain embodiments, it is not intended to be limited to the specific forms set forth herein. Furthermore, although features may be explicitly described in conjunction with specific embodiments, those skilled in the art will recognize that the various features of the described embodiments may be combined in any manner suitable for implementing the present technology.

[0061] The embodiments of the present technology can be generally described by the following numbered items:

[0062] 1. A method for determining a handball by a player from an image, comprising:

[0063] receiving an image from a camera, the image including an arm or hand of a player and a ball;

[0064] Determine the player's posture when the ball makes contact with their arm or hand;

[0065] comparing the athlete's posture to a predetermined posture; and

[0066] Based on the comparison, it is determined that a handball incident has occurred.

[0067] 2. The method according to item 1, wherein the athlete's posture and the predetermined posture are 3D models.

[0068] 3. A method according to claim 1 or 2, comprising selecting a predetermined gesture based on an action performed by the player when the ball contacts the player's arm or hand.

[0069] 4. A method according to any preceding claim, comprising generating the predetermined gesture using the position of the player's hands or arms in previous handball incidents.

[0070] 5. A method as described in any preceding claim including using player tracking and ball tracking to determine when the ball makes contact with a player's arm or hand.

[0071] 6. A method according to any preceding claim, wherein the predetermined gesture has an opaque border, and wherein the comparing step comprises superimposing the predetermined gesture on the player's gesture.

[0072] 7. A method according to any preceding claim, comprising: determining the speed of the ball at the moment the ball makes contact with the player's arm or hand; and determining that a handball event has occurred based on the speed of the ball and a previous handball decision.

[0073] 8. A computer program product comprising computer-readable instructions which, when loaded into a computer, configure the computer to perform the method according to any one of items 1 to 7.

[0074] 9. A device for determining a handball committed by a player from an image, comprising:

[0075] Communication circuitry configured to receive an image from a camera, the image including an arm or hand of a player; and processing circuitry configured to:

[0076] Determine the player's position when the ball makes contact with the player's arm or hand;

[0077] comparing the athlete's posture to a predetermined posture; and

[0078] Based on the comparison, it is determined that a handball incident has occurred.

[0079] 10. The apparatus according to item 9, wherein the athlete's posture and the predetermined posture are 3D models.

[0080] 11. The apparatus of item 9 or 10, wherein the processing circuit is configured to select the predetermined gesture based on an action performed by the player when the ball contacts the player's arm or hand.

[0081] 12. The apparatus of any one of items 9 to 11, wherein the processing circuit is configured to generate the predetermined gesture using the player's hand or arm position in a previous handball incident.

[0082] 13. The apparatus of any one of items 9 to 11, wherein the processing circuit is configured to use player tracking and ball tracking to determine when the ball makes contact with an arm or hand of a player.

[0083] 14. The apparatus of any one of items 9 to 13, wherein the predetermined gesture has an opaque border, and wherein the comparing step comprises superimposing the predetermined gesture on the athlete's gesture.

[0084] 15. An apparatus according to any preceding claim, wherein the processing circuit is configured to: determine the speed of the ball at the moment the ball makes contact with the player's arm or hand; and determine that a handball event has occurred based on the speed of the ball and previous handball decisions.

Claims

1. A method for determining a handball by a player from an image, comprising: receiving an image from a camera, the image including the player's arm or hand and a ball; determining the posture of the player when the ball makes contact with the player's arm or hand; comparing the posture of the athlete to a predetermined posture; and Based on the comparison, it is determined that a handball incident has occurred, wherein the comparing comprises determining a boundary defining a natural profile posture of the athlete in the predetermined posture and determining whether the posture of the athlete is outside the boundary by superimposing the natural profile posture of the athlete on the posture of the athlete, If the posture of the player is on or within the boundary, determining that no handball incident occurs; If the posture of the player is outside the boundary, it is determined that a handball incident has occurred.

2. The method according to claim 1, wherein The posture of the athlete and the predetermined posture are 3D models.

3. The method according to claim 1, comprising: The predetermined gesture is selected based on an action performed by the player when the ball contacts the arm or hand of the player.

4. The method according to claim 1, comprising: The predetermined gesture is generated using the player's hand or arm position in previous handball incidents.

5. The method according to claim 1, comprising: Player tracking and ball tracking are used to determine when the ball makes contact with the arm or hand of the player.

6. The method according to claim 1, wherein The predetermined gesture has an opaque border.

7. The method according to claim 1, comprising: determining the velocity of the ball at the moment the ball makes contact with the arm or hand of the player; And based on the speed of the ball and the previous handball penalty, it is determined that a handball incident has occurred.

8. A computer program product comprising computer readable instructions which, when loaded onto a computer, configure the computer to perform the method according to claim 1.

9. A device for determining a handball by a player from an image, comprising: communications circuitry configured to receive an image from a camera, the image including an arm or hand of the athlete; as well as The processing circuit is configured to: determining the posture of the player when the ball makes contact with the player's arm or hand; comparing the posture of the athlete to a predetermined posture; and Based on the comparison, it is determined that a handball incident has occurred, Wherein, the processing circuit is further configured to: determining a boundary defining a natural profile posture of the athlete in the predetermined posture and determining whether the posture of the athlete is outside the boundary by superimposing the natural profile posture of the athlete on the posture of the athlete, If the posture of the player is on or within the boundary, determining that no handball incident occurs; If the posture of the player is outside the boundary, it is determined that a handball incident has occurred.

10. The apparatus according to claim 9, wherein The posture of the athlete and the predetermined posture are 3D models.

11. The apparatus according to claim 9, wherein The processing circuit is configured to select the predetermined gesture based on an action performed by the player when the ball contacts the arm or hand of the player.

12. The apparatus according to claim 9, wherein The processing circuit is configured to generate the predetermined gesture using a player's hand or arm position in a previous handball incident.

13. The apparatus according to claim 9, wherein The processing circuitry is configured to determine when the ball makes contact with the arm or hand of the player using player tracking and ball tracking.

14. The apparatus according to claim 9, wherein The predetermined gesture has an opaque border.

15. The apparatus according to claim 9, wherein The processing circuit is configured to: determine the speed of the ball at the moment the ball contacts the arm or hand of the player; and determine that a handball event has occurred based on the speed of the ball and a previous handball penalty.

Images