FR2926656A1 - Gesture recognition device for e.g. board game, has monitoring unit monitoring trajectory of one of spherical or hemispherical shaped color markers and comparison unit comparing trajectory of marker with predetermined trajectory - Google Patents

Abstract

The device (5) has a camera (8) positioned on an upper end of a case (6) and connected to an image processing processor i.e. digital signal processor, that processes images transmitted by the camera. Spherical or hemispherical shaped color markers (12) are positioned on a member of user. A monitoring unit formed of coordinates restoring unit and trajectory reconstruction unit, monitors a trajectory of one of the markers and a comparison unit compares the trajectory of marker with a preset trajectory. The camera has a complementary MOS sensor with a signal processing module. An independent claim is also included for a gesture recognition method for capturing a series of images.

Description

The present invention relates to a gesture recognition method and device particularly suitable for the production of board games, for functional rehabilitation or for sports training in particular.

In the field of board games, it is well known so-called board games or cards for fun and educational purposes. This is the case for example board games such as monopoly (trademark) or the trivial pursuit (trademark) for example. However, there are no board games such as board games or cards allowing for a playful way of developing co-ordination skills in particular. Moreover, it is well known that after a trauma, following an accident or a surgical operation for example, it is necessary to re-educate the patient so that he can again perform normal actions such as walking or grasping including . For this purpose, the patient follows a functional rehabilitation treatment generally dispensed by a physiotherapist or the like. This re-education treatment essentially consists of gestural work to strengthen the limbs and regain a better laxity so that the patient can find normal gestures. However, the duration of treatment is particularly dependent on the availability of physiotherapists, only the latter being able to validate and correct the movements of the patient, so that this functional rehabilitation is often long and expensive. These treatments also promote the development of the coordination of movements using the main senses of the patient (hearing, sight, etc. ...) and therefore apply perfectly to the development of stimulation products and cognitive rehabilitation.

Finally, it is well known that the precision of the gesture is paramount in the effectiveness of a sports movement, such as for rifle shooting, golf for swing, archery, etc ... However, it is often difficult for the athlete himself, but also for the coach of the latter, to perceive the flaws in the gesture to correct it. One of the aims of the invention is therefore to remedy these drawbacks by proposing a method and a gesture recognition device of simple design, inexpensive, providing high precision and intended for a board game, a rehabilitation device or sports training.

For this purpose, and in accordance with the invention, there is provided a remarkable gestural recognition method in that it comprises at least the following steps of capturing a succession of images including at least one color marker capable of being positioned on a member of a user and / or on a moving object driven by said user, determining the trajectory of said marker, and comparing the trajectory of said marker with at least one predetermined trajectory. Particularly advantageously, the method comprises a step of adjusting the image capture parameters by the CMOS sensor according to the characteristics of the ambient light.

Said adjustment step comprises at least the following steps of capturing an image in RGB format, transforming the image in RGB format into a grayscale image, calculating the histogram of the image, for calculating the energy in the RGB image, transforming the RGB image to TLS information (Hue, Luminescence, Saturation), calculating the overall energy of the image from luminance values, comparing said global energy with a threshold value, extracting data relating to the contrast and to the colors of the image from the histogram if the overall energy is greater than or equal to the threshold value, adjusting the parameters image capture of the CMOS sensor according to the values of the histogram. The step of tracking the trajectory of said marker comprises at least the following steps of recovering the coordinates of said marker and reconstruction of the trajectory of said marker from said coordinates. The step of recovering the coordinates of said marker comprises at least the following steps of capturing the RGB image, of transforming the RGB image into a TLS image, of determining the pixels of the image for which the hue of the image is between two threshold values, the luminance is greater than a threshold luminance value and a saturation greater than a threshold saturation value, the selection of the pixel groups corresponding to said marker and a determination of a center of gravity of a pixel group whose pixel density is within a reference range, the coordinates of said center of gravity corresponding to the coordinates of said marker. The step of reconstructing the trajectory consists in constructing a set of regularly spaced points from the coordinates of said previously recorded marker. Advantageously, the method according to the invention comprises a tremor correction step comprising at least the following steps of determining the vector's directing coefficient connecting two successive positions of the marker, determining the vector's directing coefficient connecting the two successive successive positions. of the marker, for determining the directing coefficient of a plurality of successive positions of said marker, said major director coefficient, of comparing the sign of the vector directing coefficients and of the major directing coefficient, and of transforming the coordinates of the position of said marker when the sign key coefficients change.

Another object of the invention relates to a gestural recognition device which is remarkable in that it consists of at least one camera connected to a processing processor of the images transmitted by said camera, of at least one color marker suitable for being positioned on a member of a user, tracking means of the trajectory of said marker and means for comparing the trajectory of said marker with at least one predetermined trajectory. Said marker consists of a hemispherical or spherical shape marker comprising gripping means.

Preferably, the device comprises two markers intended to be entered respectively by a hand of a user. Said camera consists of a CMOS sensor comprising a signal processing module and the processor consists of a digital signal processor called DSP processor.

Advantageously, the device comprises means for adjusting the image capture parameters by the CMOS sensor according to the characteristics of the ambient light. Said adjustment means consist of means for capturing an image in RGB format, means for transforming the image in RGB format into a grayscale image, means for calculating the histogram of the image. image, means for calculating the energy in the RGB image, means for transforming the RGB image into TLS information (Hue, Luminescence, Saturation), means for calculating the overall energy of the image from the luminance values, means for comparing said global energy with a threshold value, means for extracting data relating to the contrast and to the colors of the image from the histogram if the overall energy is greater than or equal to the threshold value, and means for adjusting the image capture parameters of the CMOS sensor according to the values of the histogram.

Said means for tracking the trajectory of said marker consist of means for recovering the coordinates of said marker and means for reconstructing the trajectory of said marker from said coordinates.

Said means for recovering the coordinates of said marker comprise means for capturing the RGB image, means for transforming the RGB image into an image in TLS format, means for determining the pixels of the image for which the hue of the image the image is between two threshold values, the luminance is greater than a threshold luminance value and a saturation greater than a threshold saturation value, means for selecting the pixel groups corresponding to said marker and means for determining a center of gravity of a pixel group whose pixel density is within a reference range, the coordinates of said center of gravity corresponding to the coordinates of said marker. Said path reconstruction means consist of means capable of constructing a set of regularly spaced points from the coordinates of said previously recorded marker. Advantageously, the device according to the invention comprises means for correcting tremors consisting of means for determining the vector's directing coefficient connecting two successive positions of the marker, means for determining the vector's directing coefficient connecting the two following successive positions. of the marker, means for determining the steering coefficient of a plurality of positions of said marker, said major steering coefficient, means for comparing the sign of the vector directing coefficients and the major steering coefficient, and means for transforming the coordinates of the position of said marker when the sign of the control coefficients changes.

Furthermore, the device comprises an autonomous housing adapted to receive the camera, the image processing processor transmitted by said camera, the tracking means of the trajectory of said marker, the means for comparing the trajectory of said marker with at least one trajectory. predetermined and an autonomous source of energy. Other advantages and features will emerge more clearly from the description which follows, of several variant embodiments, given by way of non-limiting examples, of the method and the device for sign recognition according to the invention, from the accompanying drawings on which: - Figure 1 is a perspective representation of a board game board game type comprising a gesture recognition device according to the invention; - Figure 2 is a schematic perspective view of the gesture recognition device according to the invention; FIG. 3 is a schematic representation of the gestural recognition device according to the invention; FIGS. 4A and 4B are schematic representations of the position of a player with respect to the gestural recognition device according to the invention; FIG. 5 is a flow chart of the different steps of ambient light analysis of the gesture recognition device. According to the invention, FIG. 6 is a flow diagram of the different steps for recovering the color markers of the gesture recognition device according to the invention; FIG. 7 is a flow chart of the step of reconstructing the movement of the device; 8A and 8B are diagrammatic representations of the movements respectively before and after reconstruction, FIGS. 9A to 9C are schematic representations of the various steps of correction of the tremors of the gestural recognition device according to FIG. FIG. 10 is a flow diagram of the various stages of the board game comprising a gesture recognition device according to the invention; FIG. 11 is a flowchart of the different steps of the analysis of a following movement. a geometrical shape by means of the gesture recognition device according to the invention, FIG. 12 is a flow diagram of the various contour detection steps according to the invention; FIG. 13 is a flow chart of the various steps of analysis of the coordination of the movements according to the invention; FIG. 14 is a flow chart of the various steps of analysis of the momentum according to the invention, - Figure 15 is a flowchart of the different stages of analysis of the position of the markers according to the invention, - Figure 16 is a perspective representation of a variant. embodiment of the card game type board game comprising a gesture recognition device according to the invention, FIG 17 is a schematic representation of a functional rehabilitation aid device comprising a gesture recognition device according to the invention. the invention. A gestural recognition device particularly suitable for the production of a board game and / or card game and a functional rehabilitation aid comprising a sign recognition device will be described below. according to the invention; However, it is obvious that the device according to the invention can find many applications in the field of sports including to improve the accuracy of the actions of athletes for example. With reference to FIG. 1, the board game of the board game type consists of a game board 1 on the upper face of which is drawn a path 2 constituted by a succession of boxes 3 able to receive pins 4, a counter is assigned to each player. The game also comprises a gestural recognition device 5 consisting of a generally rectangular housing 6 comprising a base 7, a camera 8 positioned at the upper end of said housing, a first display area 9, a second area of the housing. display 10 divided into two subassemblies 10a and 10b, control buttons 11 and two markers 12. It is obvious that the housing 6 of the gesture recognition device may have any shape without departing from the scope of the invention. Said markers 12 have a generally hemispherical shape and comprise gripping means 13. The hemispherical portions of said markers 12 have an orange and green color, respectively. The color of said markers 12 is uniform and their texture is perfectly smooth. Said markers 12 are held in the right and left hands of the player by placing the gripping means 13 between two successive fingers of a hand so that the hemispherical portion of said marker 12 extends at the level of the palm of the hand. . It is obvious that said markers 12 may have different colors, shapes and textures without departing from the scope of the invention.

The first display area 9 located just above the base 7 and the control buttons 11 comprises geometric shapes such as a triangle, a square, a circle and a rhombus for example, able to be illuminated during of the game as it will be detailed later.

The second display area 10 located between the first display area 9 and the camera 8 comprises a plurality of pictograms corresponding to a stylized representation of the markers 12 and able to be illuminated during the game as will be seen later.

With reference to FIGS. 1 to 3, the gesture recognition device 5 preferably consists of a CMOS sensor 14 according to the English acronym Complementary Metal Oxide Semiconductor, of a digital signal processor known as DSP 15 according to the acronym digital signal processor connected to said CMOS sensor 14, a SDRAM memory 16 according to the English acronym Synchronous Dynamic Random Access Memory connected to the input EMI of the DSP 15, a flash memory 17 also connected to the EMI input of the DSP processor 15 and power supply means 18 such as batteries or a transformer connected to the domestic electrical network for example.

Incidentally, the gesture recognition device 5 may comprise sound transmission means 19 such as an amplifier and loudspeakers for example. Furthermore, the gesture recognition device 5 comprises an optical lens 20 placed in front of the CMOS sensor 14. In a particularly advantageous manner, said lens is integral with the CMOS sensor 14 so as to reduce the manufacturing costs of said device by avoiding the use of a device. The gesture recognition device 5 thus makes it possible to define, with reference to FIG. 2, a field of vision 21 of approximately 2m by 1.5m situated at a distance of approximately 1m from said device. Thus, with reference to FIGS. 4A and 4B, the player or players will have to sit in front of the gestural recognition device 5 placed on a fixed support such as a table 22 for example, at a distance of about 1 m from the latter. The player can stand as well sitting (Figure 4A) as standing (Figure 4B), as long as his hands do not come out of the field of vision 22 as previously defined during the execution of the game. In addition, the device of gesture recognition 5 comprises means for adjusting the image capture parameters by the CMOS sensor as a function of the characteristics of the ambient light and tracking means of the trajectory of said marker recorded in the DSP processor 15 of the device according to the invention .

Means for Adjusting the Capture Parameters Referring to FIG. 5, said capture parameter adjustment means makes it possible to automatically adjust the image pickup parameters of the CMOS sensor 14 according to the characteristics of the ambient light at the camera. moment t. Ambient light characteristics are understood to mean the contrast, luminance and saturation of light in particular. Said adjustment means consist of means for capturing an image in RGB format 30, means 31 for analyzing the histogram of the captured RGB image, means 32 for transforming the image into RGB format. a grayscale image, calculation means 33 of the histogram of the grayscale image, energy calculation means 34 in the RGB image, transformation means 35 of the RGB image to TLS information (Hue, Luminescence, Saturation), means 36 for calculating the overall energy of the image from the luminance values, means 37 for comparing said overall energy with a threshold value, Data extraction means 38 (White_pic, Black peak, Color max, Color min, Color_moy) relating to the contrast and to the colors of the image from the histogram if the overall energy is greater than or equal to the threshold value , and means 39 for adjusting the image capture parameters of the CMOS sensor 14 (Con Trace, Saturation, Luminance) according to the values of the histogram such as Color max, Color_min and Color_moy for example. RGB image according to the English acronym Red Green Blue is an image whose description of colors is defined by the mixture of three basic components, red, green and blue. Said means previously described consist of algorithms recorded in the DSP 15 of the gesture recognition device 5. Thus, when the gesture recognition device 5 is turned on, an RGB image is captured by the CMOS sensor 14 in a first step, then the energy E in said RGB image is calculated, then the RGB image is transformed into a TLS image (Hue, Luminance, Saturation), then the overall energy of the image is calculated from the luminance values, then said overall energy is compared with a reference threshold value. At the same time, a step of analyzing the histogram of the RGB image is performed, then the TGB image is converted into a greyscale image, and the histogram of said greyscale image. is calculated.

If the overall energy of the image is less than the reference threshold value, an error message 40 is emitted by the sound emission means 19 of the gesture recognition device 5 indicating that the lighting of the room in which the device is installed is insufficient. If the overall energy of the image is greater than or equal to the reference threshold value, data (White_pic, Black_pic, Color max, Color min, Color_moy) relating to the contrast and colors of the image are extracted from the histogram of the grayscale image, then the values of the Color max data are compared to the values of the Black_pic and White_pic data.

If the values of the data Color max are lower than the values of the data Black_pic or White_pic, an error message 41 is emitted by the sound transmission means 19 of the gesture recognition device 5 indicating that the backlight is too important.

If the Color max data values are greater than the Black_pic or White_pic data values, the CMOS sensor 14 image capture settings (Contrast, Saturation, Luminance) are adjusted according to the histogram values such as Color max , Color_min and Color_moy for example.

This algorithm is executed in a loop in order to modify the parameters of image capture during the game, the lighting and backlighting conditions may vary over time.

Means for Monitoring the Traiectory of the Marker Said means for tracking the trajectory of said marker consist of means for recovering the coordinates of said marker and means for reconstructing the trajectory of said marker from said coordinates. Said means consist of algorithms registered in the DSP processor 15 of the gesture recognition device 5.

Means for recovering the coordinates of said marker

Said means for recovering the coordinates of said marker comprise, with reference to FIG. 6, means 30 for capturing the image in RGB format, means 40 for transforming the RGB image into a TLS format image and means for determining the pixels of the image for which the hue of the image is between two threshold values, the luminance is greater than a threshold luminance value and the saturation is greater than a threshold saturation value, said threshold values being directly dependent on a threshold luminance value; CMOS sensor characteristics. It will be observed that the transformation means 40 of the RGB image into a TLS format image may consist in the means for transforming the RGB image into a TLS format image of means for adjusting the capture parameters. The means for recovering the coordinates of the marker also comprise means for selecting the pixel groups corresponding to said marker and means for determining a center of gravity of a pixel group whose pixel density is in an interval. reference, the coordinates (x, y) of said center of gravity corresponding to the coordinates of said marker.

Means for reconstructing the trajectory With reference to FIGS. 7, 8A and 8B, the path reconstruction means consist of means capable of constructing a set of regularly spaced points from the coordinates of said previously recorded marker.

Thus, with reference to FIG. 7, these means consist of an algorithm comprising a reconstruction function 60 which determines an array 61 of points (of coordinates (x ', y') regularly spaced two by two, corresponding to the trajectory represented on FIG. Figure 8B, from a table 62 of points (coordinates (x, y)), corresponding to the trajectory shown in Figure 8A, determined by the recovery means 42 of the coordinates of the marker.

Furthermore, the gesture recognition device 5 advantageously comprises means for correcting tremors.

Means for correcting tremors

In order to eliminate the irregularities of the movement due to more or less pronounced tremors of people and thus avoid recognition errors, the device according to the invention comprises tremor correction means which comprise means for determining the directing coefficient of the connecting vector. two successive positions of the marker from the array of regularly spaced points previously determined, means for determining the vector's directing coefficient connecting the two following successive positions of the marker, means for determining the directing coefficient of the following positions of said marker said major governing coefficient means for comparing the sign of the vector director coefficients and the major director coefficient, and means for transforming the coordinates of the position of said marker when the sign of the direction coefficients changes. Said means consist of algorithms recorded in the DSP processor 15 of the gesture recognition device 5. Thus, with reference to FIGS. 9A to 9C, the method used consists of comparing the directing coefficients of two successive segments. With reference to FIG. 9A, the director coefficient α between two successive points of the regularly spaced point array and the director coefficient α 2 between the next two points of said regularly spaced point array are determined.

The coefficients a, and a2 are then compared. If said directional coefficients a, and a2 do not show a sign change, it means that there is no change of direction in the trajectory.

On the other hand, if said directional coefficients a, and a2 have a sign change, as shown in FIG. 9A, this means that there is a change of direction in the trajectory. The algorithm then verifies that the change of sign is temporary, that is to say that the change of sign is due to tremors, or durable, that is to say that the sign change is due to a change voluntary direction of the trajectory. For this purpose, with reference to FIG. 9B, the algorithm determines a so-called major director coefficient corresponding to the direction coefficient of all the following positions of said marker from the regularly spaced point array, then the coefficient a, is compared to the main coefficient amajeur • If the change of sign is confirmed, then no modification is made to the table of points.

On the other hand, if the change of sign is not confirmed, as represented in my FIG. 9C, then the scoreboard is modified in order to align the points due to the tremor with those of the major direction.

An example of a board game using the gesture recognition device will now be explained with reference to FIGS. 1 and 10 to 15. Referring to FIG. 1, each player or team chooses a piece 4 which is then placed on the Start of the Way box. 2. In this exemplary embodiment, said path 2 comprises 49 boxes including a Start box on which each player decides the challenge he will have to perform to start the game, 36 boxes Challenges, the game with 6 types of challenges, 5 Bonus boxes allowing the player to advance the number of squares corresponding to the challenge multiplied by 2 if the challenge is successful and to move back one box if the player fails the challenge, 6 squares Duels allowing the player who falls on one of these boxes, to choose another player against whom he will take up the indicated challenge, the winner then advancing the number of boxes corresponding to the challenge, while the loser returns directly to the box art, and an Arrival checkbox, the game being won by the first player to reach or over the finish square.

For example, players start from the smallest to the largest and each plays one after the other, whether they win or lose their challenges. For example, after successfully completing the challenge, the player advances the corresponding number of boxes and falls on a new challenge. He does not run it right away, but passes the hand to the next player. When his turn returns, he then executes the indicated challenge, moves forward or backward according to his result, and so on until a player reaches or crosses the finish square. With reference to FIG. 10, when the gesture recognition device 5 is turned on, the latter initializes, in a step 105, all of its parameters. The player then presses one of the buttons 11 (step 110) of the device 5 corresponding to the challenge of the box on which his piece 4 is located and said device 5 transmits in a step 115 an audible message indicating to the player that he must set up in front of the gesture recognition device 5 at the field of vision 21, the sound message consisting for example in In place! .

Prior to the execution of the challenge, the gesture recognition device 5 checks in a step 120 the ambient brightness level and adjusts its image capture parameters accordingly. If an optimum setting of the parameters is not possible due to a too low light level or to the presence of a backlight too important, the module warns the user by the following sound message: Insufficient lighting ... or Backlight too important

. as it was described previously. Players must then reposition the game in a more appropriate environment before starting the game. This step 120 of checking the ambient light level is performed before the launch of each challenge. If the parameters relating to the lighting and backlighting could be performed, the gesture recognition device 5 executes in a step 125 the algorithm of the challenge selected by the player. In this particular embodiment, the gesture recognition device has 6 different challenges: Memorization, Discovery, Coordination, Music, Skill and Mirror, and 2 different Duels: Music and skill described below ... DTD: Challenge Memorization

Referring to Figure 11, the orange marker is positioned in the right hand or left hand of the player who is positioned in front of the module at the field of vision as described above. When the player is ready, one of the other players just press the button 11 corresponding to the challenge and the gesture recognition device beeps: In place! . The geometric shapes of the first display area 9 illuminate and then go out, one after the other, to form a sequence that the candidate will have to reproduce with his hand carrying the orange marker. After this step, the module announces: Ready to go. The player executes the first geometric shape, then closes the hand to hide the color marker. A beep will confirm the successful completion of the gesture and allow the player to reopen his hand to follow on the next form. In case of bad performance, indicated by a different beep, the player will have to repeat the same gesture until the success of it. And so on until the complete reproduction of the series in the allotted time.

In order to recognize the form performed by the player, the gesture recognition device 5 comprises pattern recognition algorithms from the regularly spaced point array. For a circle, the algorithm checks whether all the points are included in a crown whose width corresponds to the tolerance threshold, the tolerance corresponding preferably to a percentage of the average radius of the circle executed by the player. For a triangle, the algorithm constructs a graph representing the distance of each of the points at the center of gravity, said graph having a sinusoidal function form comprising vertices and valleys. The algorithm then verifies that the sum of the vertices and troughs of the graph is equal to 6 and that they are all included in a band whose width corresponds to the tolerance threshold, said tolerance threshold being preferably equal to a percentage the average distance of the vertices and the average distance of the hollows of said graph. For a square, the algorithm calculates the 2 dimensions of the square and checks that the difference between the 2 values is well below the tolerance threshold. Then she checks the horizontality and verticality of each side. Indeed, the square is considered correct only if each side is included in a band whose width corresponds to the tolerance threshold. The tolerance threshold for one side of the square is equal to a certain percentage of the length of the considered side. The algorithm calculates in particular the difference between two adjacent sides, said difference being to be less than the tolerance of one of the sides of the square, and determines if all the points of the square drawn by the player are situated in a band whose width corresponds at the tolerance threshold of the considered side. For a rhombus, the algorithm inclines the 45 ° geometric shape to transform it into a square. Then, the algorithm as previously described for a square is executed to check if the geometric shape is correctly reproduced.

Discovery Challenge Referring to Figure 12, the orange marker is positioned in the right hand or left hand of the player who is positioned in front of the module at the field of vision as described above. When the player is ready, one of the other players just press the button corresponding to the challenge. The module announces: In place !!! followed 2s later by the message Ready to Go. In the virtual space corresponding to the field of view, there exists a geometrical shape randomly selected by the gesture recognition device 5. By walking his hand in the virtual space, the player will hear a beep every time the marker passes over the outline of the sign. the geometric form. The player will have to mentally reconstruct the shape to determine which of the four possible geometric shapes is the right one. Before the end of the allotted time, he will have to make his choice: - Choice 1: The announced geometric shape is in conformity with that indicated by the module. The player advances one space. - Choice 2: The player prefers not to give an answer. He moves back one space. - Choice 3: The geometric shape announced is different from that indicated by the module. The player moves back 3 spaces. 25 Coordination Challenge

Referring to FIG. 13, the orange marker is positioned in the player's right or left hand and the green marker is positioned in the right or left hand of the player who is positioned in front of the module at the field of view. as previously described.

When the player is ready, one of the other players just press the button corresponding to the challenge. The module announces: In place !!! then illuminate 2 geometrical shapes selected randomly. After this step, the module announces: Ready to go. The player executes both forms simultaneously, then simultaneously closes both hands to hide the 2 color markers. In case of bad realization, a beep will indicate to the player that he must repeat his attempt. And so on until the correct reproduction of these 2 forms in the time allotted. The module may display only one geometric shape. The player will then all the more chance to succeed since these 2 hands will have to reproduce simultaneously the same geometric form instead of 2 different forms in the general case.

Challenge Music Referring to Figure 14, the orange marker is positioned in the right or left hand of the player who is positioned in front of the module at the field of vision as described above. When the player is ready, one of the other players just press the button corresponding to the challenge. The module announces: In place !!! 2s later follow the message Ready to go and then broadcast the music on which the player will keep moving his hand in the virtual zone as defined above. The goal is not to go as fast as possible, but to make the widest possible movements while remaining in the virtual space. Indeed, it is the amount of movement that is measured, both during the Music phase and during the Silence phase where the player must, conversely, remain perfectly immobile. And if the amount of movement during the Silence phase remains below a reference threshold, and the momentum during the Music phase is greater than another reference threshold throughout the duration of the challenge, then the latter will be successful. . Conversely, a beep followed by a failure message will be broadcast by the module.

Challenge Skill

With reference to FIG. 15, the orange marker is positioned in the right or left hand of the player and the green marker is positioned in the right or left hand of the player who is positioned in front of the module at the level of the field of vision as described. previously. When the player is ready, one of the other players just press the button corresponding to the challenge. The module announces: In place !!! then illuminates 2 markers positioned on the front face, one on the right side, the other side on the left, indicating the initial position of the hands in the virtual space before the actual start of the challenge. The player gets ready and the module announces: Ready to go. The player must bring his hands in the virtual space towards the positions indicated by the leds which lights up then go off randomly on the diagram on the front face. It is important to understand that the position of the markers on the diagram on the front face corresponds to the position of the hands that the player must follow in the virtual space. If he succeeds in following the rhythm imposed by the module without being mistaken for the duration of the challenge, then the latter will be successful. Otherwise, an audible beep followed by a failure message will be broadcast by the module.

Mirror Challenge

The orange marker is positioned in the right or left hand of the player and the green marker is positioned in the right or left hand of the player who is positioned in front of the module at the field of vision as described above. When the player is ready, one of the other players just press the button corresponding to the challenge. The module announces: In place !!! then illuminate 1 marker positioned on the front face. The player must then position his hand opposite to what is indicated to him. This is the actual starting point of the challenge. The module then announces: Ready to go. The LEDs illuminate and then go off randomly on the diagram on the front panel. Unlike the previous challenge, this time the player must bring his hands to the opposite position indicated by the LEDs.

Duel Music

The orange marker is positioned in the right or left hand of a first player, the player who has fallen on the square corresponding to this duel choosing the hand he prefers, and the green marker is positioned in the left or right hand respectively the second player. Both players are positioned next to each other in front of the module at the field of vision as previously described.

When the players are ready, the music challenge described above is executed; However, there is no more limited time. In the case of the duel, the first player who exceeds the reference values has lost. The winner is announced by the module via the message Winner: Player 'Orange'! or Winner: Player `Green '! .

Dual Skill

The orange marker is positioned in the right or left hand of a first player, the player who has fallen on the square corresponding to this duel choosing the hand he prefers, and the green marker is positioned in the left or right hand respectively the second player. Both players are positioned next to each other in front of the module at the field of vision as previously described.

When the players are ready, the skill challenge described above is executed. Note that the right player and the left player must respectively position their hand in the virtual space of right and left, according to the indications given by the leds in front of the module. In the case of the duel, there is no more limited time. The first player who can not keep up with the pace imposed by the module has lost. The winner is announced by the module via the message Winner: Player 'Orange'! or Winner: Player `Green '! .

According to an alternative embodiment of the gesture recognition device 5, not shown in the figures, the latter comprises in the same manner as before a generally rectangular housing 6 having a base 7, a camera 8 positioned at the upper end of said housing, a first display zone 9, a second display zone 10 divided into two subsets 10a and 10b, control buttons 11 and two markers 12. In this particular embodiment, it consists of a sensor CMOS 14 integrating the functions of digital processing of adjustment of the capture parameters, tracking the trajectory of said marker and recovering the coordinates of said marker, a microprocessor connected to the CMOS sensor for reconstruction of the trajectory and incidentally the correction of the tremors and in which are recorded and are executed the algorithms corresponding to different challenges and dual such than previously described.

Such an architecture of the gesture recognition device 5 makes it possible in particular to significantly reduce the cost of the device by using a low-cost microprocessor for the execution of challenges and dual instead of the DSP processor which is expensive.

According to an alternative embodiment of the board game according to the invention, with reference to FIG. 16, the latter comprises, in the same manner as previously, a gesture recognition device 5 consisting of a generally rectangular box 6 comprising a base 7, a camera 8 positioned at the upper end of said housing, a first display area 9, a second display area 10 divided into two subsets 10a and 10b, control buttons 11 and two markers 12 of orange color and green for example. The game is distinguished by the fact that the game board 1 and the pieces 4 are substituted by a card game 70, totems 71 and so-called check pawns 72.

This game can for example include 24 Totems, following the drawing of a Totem card, the player recovering a checker Failure if he does not succeed the challenge and recovering the totem if he succeeds the challenge, 12 Checkers failed, following the draw of a Challenge card, the player recovering a Failure token if he does not succeed the challenge in question and separating himself from a pawn if he succeeds the challenge, and a deck of 52 cards including 36 Challenge cards, 6 Totem cards , 6 Duel cards and 4 Bonus cards. Challenge cards represent one of the 6 types of challenges to perform as described in the execution variant of the game: In case of failure, the player retrieves a failed counter. In the opposite case, he separates himself if, of course, he has some. On the other hand, as soon as he has 3 checkers, he is forced to part with one of his totems. The 3 pawns then take their place in the pile with the others. If a Totem card is drawn, the player picks the corresponding totem from the draw pile if he succeeds in the challenge.

If a Duel card is drawn, the player chooses one of the other players with one of the totems of interest. The two players will play against each other and whoever wins the challenge will choose one of his opponent's totems. If the loser does not have totems, or if the winner already has all the totems of the loser, then the winner will take the totem that interests him in the draw. In the event of a draw of a bonus card, the player has no challenge to raise, but may separate from a failed counter if he has at least one. The goal of the game remains the same as the board game version described above and consists of raising and winning enough challenges to win the game. It is for participants to reproduce gestures or perform particular movements for a limited time. This duration is fixed randomly by an electronic module which later judges the quality of realization of the challenges. By shooting, one after the other, game cards giving the challenge to achieve, we win or lose totems based on the results announced by the module and the winner is the one who collects the first 6 totems corresponding to 6 types possible challenges. Each card is associated with a theme, and therefore a particular type of challenge.

Initially, the deck of cards is shuffled and placed on the table, back of the deck of cards up. Players start from the smallest to the largest and each plays one clock after another. The player draws a card, performs the action as indicated above, puts the card under the deck and passes the hand to the next player, and so on. The game is won by the first player who collects the first 6 totems corresponding to 6 types of possible challenges. However, nothing prevents to set a playing time and to make the accounts at the end of this period. Whoever wins the most challenges then wins the game.

Referring to Figure 17 and following a particularly interesting application of the device according to the invention, there is described a functional rehabilitation device consisting of a gesture recognition device 5 as described above comprising at least one generally rectangular housing 6 having a base 7, a camera 8 positioned at the upper end of said housing 6, control buttons 11 and at least one marker 12. In this embodiment, the device comprises three markers 12 positioned respectively at the hip, knee and the ankle of a patient for the functional rehabilitation of his leg. Following an accident, surgery, normal aging or any other cause directly affecting the motor organs, some people use functional rehabilitation techniques in order to find a normal motor, or at least prevent it do not wither. In some cases, simple daily actions adapted to the motor problem of the person can effectively rehabilitate failing motor organs. Take, for example, the case of a person operated on the knee following tearing ligaments. After an immobilization more or less long, it must follow rehabilitation sessions in order to find as quickly and as effectively as possible the use of his knee. The realization of the movements should be as precise as possible in order to achieve optimal rehabilitation. The amplitude of these movements will have to be progressive over time according to an evolutionary training program allowing to bring the patient from his current state to his normal state before operation. The device according to the invention makes it possible to provide an aid to the functional rehabilitation of persons by monitoring the precision of movements, the speed of movements and the evolution of the amplitude of said movements. For this purpose, said device comprises a training program capable of adjusting these parameters to the progress to be made by the patient or the progress actually made by the patient. Said device according to the invention is previously parameterized by the practitioner, such as a physiotherapist for example, with the data corresponding to the pathology of the patient. This information is provided to the gesture recognition device 5 by a direct connection of a PC or PDA type computer to an input of the device or by a remote connection, preferably by means of a secure network 81 such as a connection secure internet, via an application program running on a computer type PC located in the cabinet 82 of the physiotherapist for example.

These parameters can also be modified a posteriori in order to be able to adjust the rehabilitation program according to the actual progress of the patient. The analysis of the movements made by the patient is made possible by the real-time monitoring of the markers 12 positioned at the hip, knee and ankle of the patient, the position of these markers 12 as well as the position of the patient in front of the patient. the gestural recognition device being previously defined by the physiotherapist. The device advantageously comprises a video monitor 83, such as a television or a screen of a PC-type computer for example, connected to a video output of the gesture recognition device 5 to ensure a visual feedback allowing the patient to perform the movements with maximum precision. The gesture recognition device 5 comprises algorithms for adjusting the capture parameters, tracking the trajectory of said marker, recovering the coordinates of said marker, reconstructing the trajectory and, incidentally, correcting the tremors as described previously in order to be able to manage the precision of the movements, the speed of the movements and the evolution of the amplitude of said movements.

In order to verify the progress made by the patient, the device summarizes the data relating to the sessions performed, thus enabling the physiotherapist to control the smooth running of the rehabilitation phase and to intervene effectively in case of any drifts. Said physiotherapist can intervene remotely via a PC application allowing the recovery of results and the parameterization of data related to the pathology of the patient thus avoiding the daily movement of the latter or the patient's movement in the office of said physiotherapist. Similarly, said physiotherapist can intervene in the patient's home by directly connecting the gesture recognition device 5 to a laptop or the like. It may possibly be possible to display the summary information by consulting them via the video monitor 83 connected to the gesture recognition device 5. Finally, another particularly interesting application of the gesture recognition device according to the invention concerns the field of sports activities. * Indeed, one of the keys to success of many sports lies in the precision of the gestures and movements performed by the athlete. This is particularly the case of the golf swing, the basketball free throw or the biathlon competition, etc ... The precision and the ability to repeat their gesture is acquired during training. The gesture recognition device 5 according to the invention makes it possible to verify in real time the good execution of these movements in space and the level of repetitiveness of these same movements over time. Said gesture recognition device 5 comprises in the same way as previously at least one generally rectangular housing 6 comprising a base 7, a camera 8 positioned at the upper end of said housing 6, control buttons 11 and at least one marker 12. In addition, the gesture recognition device 5 comprises algorithms for adjusting the capture parameters, tracking the trajectory of said marker, recovering the coordinates of said marker, reconstructing the trajectory and, incidentally, correcting the tremors as described above. in order to be able to manage the precision of the movements, the speed of the movements and the evolution of the amplitude of said movements.

For example, for a biathlete, the device has three markers positioned respectively on the shoulder thereof, the end of the barrel of the rifle and the end of the butt of said rifle. In this particular example, three so-called reference curves corresponding to the movements of the three markers then define the ideal movement that the athlete must reproduce. In real time, during the execution of the movement, the gesture recognition device determines the trajectories of the three markers and compares them with the reference curves to determine whether the markers have deviated from their ideal position or not.

During the execution of the movement, the athlete then immediately knows where his error lies simply by listening to the sound information transmitted by the gesture recognition device 5, the latter emitting an audible signal as soon as the difference between the reference curve and the curve of said marker is greater than a tolerance threshold. A different sound for each marker may for example be defined. At the end of the execution of his gesture, the athlete can visualize on a screen the plot of his movement and compare it with the reference movement to better understand where the potential problem lies.

Advantageously, the tolerance threshold may be set according to the accuracy requirements of the athlete, the tolerance threshold being small plus the accuracy of the gesture to be large. Finally, it is obvious that the examples which have just been given are only particular illustrations which are in no way limiting as to the scope of the invention.

Claims (6)

1 - Gesture recognition device characterized in that it consists of at least one camera connected to a processing processor of the images transmitted by said camera, at least one color marker capable of being positioned on a member of a user, means for tracking the trajectory of said marker and means for comparing the trajectory of said marker with at least one predetermined trajectory. 2 ù Device according to the preceding claim characterized in that said marker consists of a hemispherical or spherical shape marker having gripping means. 3 ù Device according to any one of claims 1 or 2 characterized in that it comprises two markers intended to be entered respectively by a hand of a user. 4 ù Device according to any one of claims 1 to 3 characterized in that the camera consists of a CMOS sensor comprising a signal processing module. 5. Device according to any one of Claims 1 to 4, characterized in that the processor consists of a digital signal processor called the DSP processor. 6 ù Device according to any one of claims 4 or 5 characterized in that it comprises means for adjusting the image capture parameters by the CMOS sensor according to the characteristics of the ambient light. Device according to Claim 6, characterized in that the said adjustment means consist of means for capturing an image in RGB format, means for transforming the image in RGB format into a grayscale image. , means for calculating the histogram of the image, means for calculating the energy in the RGB image, means for transforming the RGB image to TLS information (Hue, Luminescence, Saturation), means for calculating the overall energy of the image from the luminance values, means for comparing said global energy with a threshold value, means for extracting data relating to the contrast and to the colors of the image from of the histogram if the overall energy is greater than or equal to the threshold value, and means for adjusting the image capture parameters of the CMOS sensor according to the values of the histogram. 8 ù Device according to any one of claims 1 to 7 characterized in that the tracking means of the trajectory of said marker consist of means for recovering the coordinates of said marker and means for reconstructing the trajectory of said marker from said coordinates . 9 ù Device according to Claim 8, characterized in that the means for recovering the coordinates of said marker comprise means for capturing the RGB image, means for transforming the RGB image into a TLS format image, determination means pixels of the image for which the hue of the image is between two threshold values, the luminance is greater than a threshold luminance value and a saturation greater than a threshold saturation value, the pixel group selection means corresponding to said marker and means for determining a center of gravity of a pixel group whose pixel density is within a reference interval, the coordinates of said center of gravity corresponding to the coordinates of said marker. 10 ù Device according to claims 8 or 9 characterized in that the path reconstruction means consist of means capable of constructing a set of regularly spaced points from the coordinates of said marker previously recorded.11 - Device according to claim 10 characterized in that it comprises means for correcting tremors consisting of means for determining the vector's directing coefficient connecting two successive positions of the marker, means for determining the vector's directing coefficient connecting the next two successive positions of the marker, means for determining determining the directing coefficient of a plurality of positions of said marker, said major directing coefficient, means for comparing the sign of the vector directing coefficients and the major directing coefficient, and means for transforming the coordinates of the position of said marker when the sign key coefficients change. 12 - Device according to any one of claims 1 to 11 characterized in that it comprises an autonomous housing adapted to receive the camera, the image processing processor transmitted by said camera, the tracking means of the trajectory of said marker, the means for comparing the trajectory of said marker with at least one predetermined trajectory and an autonomous energy source. 13 - sign recognition process characterized in that it comprises at least the following steps of: - capturing a succession of images including at least one color marker capable of being positioned on a member of a user and / or on a moving object driven by said user, - determining the trajectory of said marker, and - comparing the trajectory of said marker with at least one predetermined trajectory. 14 - Process according to claim 13, characterized in that it comprises a step of adjusting the image capture parameters by the CMOS sensor as a function of the ambient light characteristics. 15. A process according to claim 14, characterized in that adjustment step comprises at least the following steps of: - capturing an image in RGB format, - transforming the RGB image into a grayscale image, - calculating the histogram of the image. image, - calculation of the energy in the RGB image, - transformations of the RGB image to TLS information (Hue, Luminescence, Saturation), - calculation of the overall energy of the image from the luminance values comparing said global energy with a threshold value, extracting contrast and color data from the histogram if the overall energy is greater than or equal to the threshold value, adjusting the parameters i capture the CMOS sensor according to the histogram values. 16 - Process according to any one of claims 13 to 15 characterized in that the step of tracking the trajectory of said marker comprises at least the following steps of: - recovery of the coordinates of said marker and - reconstruction of the trajectory of said marker from said coordinates. 17. Process according to claim 16, characterized in that the step of recovering the coordinates of said marker comprises at least the following steps of: capture of the RGB image, transformation of the RGB image into a TLS format image, determination of the pixels of the image for which the hue of the image is between two threshold values, the luminance is greater than a threshold luminance value and a saturation higher than a threshold saturation value; selection of the corresponding pixel groups said marker and - determining a center of gravity of a pixel group whose pixel density is within a reference range, the coordinates of said center of gravity corresponding to the coordinates of said marker. 18. Process according to claim 16, characterized in that the step of reconstructing the trajectory consists of constructing a set of regularly spaced points from the coordinates of said previously recorded marker. 19. Process according to claim 18 characterized in that it comprises a step of correction of tremors comprising at least the following stages of: determination of the vector's directing coefficient connecting two successive positions of the marker; determination of the vector's directing coefficient connecting the two following successive positions of the marker, - determination of the directing coefficient of a plurality of successive positions of said marker said major director, 20 - of comparison of the sign of the vector director coefficients and the major director coefficient, and - coordinate transformation the position of said marker when the sign of the directional coefficients changes.