Sign language recognition device based on deep learning

Abstract

The utility model relates to the technical field of sign language recognition devices, and discloses a sign language recognition device based on deep learning, which comprises a sign language recognition device main body, a power line is fixedly mounted on the right side of the sign language recognition device main body, and a camera mechanism is fixedly mounted at the upper end of the sign language recognition device main body. The upper end of the sign language recognition device body is fixedly connected with a selfie display screen mechanism, and the upper end of the sign language recognition device body is fixedly connected with a translation display screen. According to the sign language recognition device based on deep learning, through cooperative use of a damping rotating disc, a first damping rotating shaft, a second damping rotating shaft and a third damping rotating shaft of the camera shooting mechanism, the adjusting base can be rotated, the angles of a first supporting arm, a second supporting arm and a camera can be adjusted, and the shooting angle can be flexibly adjusted to adapt to the heights and body types of people with different sign languages; therefore, the best shooting angle is achieved, various postures and action details of the hand can be completely and clearly captured, and the user experience is enhanced.

Description

Technical Field

[0001] The utility model relates to the technical field related to sign language recognition devices, and specifically to a sign language recognition device based on deep learning. Background Art

[0002] Deep learning models, especially convolutional neural networks and recurrent neural networks, are widely used in sign language recognition. Convolutional neural networks are good at processing image data and can extract the spatial features of sign language gestures, such as gesture shape and hand position, while recurrent neural networks are good at processing time series data and can capture the dynamic features of sign language gestures, such as gesture migration and speed. Sign language recognition devices are based on this and use cameras to extract gesture images and translate them to ensure normal communication between hearing-impaired people and the outside world.

[0003] Existing sign language recognition devices are mostly used in public places such as government offices, hospitals, and shopping malls to ensure normal communication between hearing-impaired people and the outside world. However, the recognition camera is usually fixed during use, which is not convenient for hearing-impaired people of different heights to use, affecting accuracy. Therefore, it is necessary to propose a sign language recognition device based on deep learning. Utility Model Content

[0004] The purpose of this utility model is to provide a sign language recognition device based on deep learning to solve the problems raised in the above background technology.

[0005] To achieve the above objectives, the present invention provides the following technical solutions: a deep learning-based sign language recognition device, comprising a sign language recognition device body, a power cord fixedly mounted on the right side of the sign language recognition device body, a camera mechanism fixedly mounted on the upper end of the sign language recognition device body, a selfie display mechanism fixedly connected to the upper end of the sign language recognition device body, and a translation display fixedly connected to the upper end of the sign language recognition device body;

[0006] The camera mechanism includes a base, a damping turntable, a pedestal, a damping shaft 1, a first arm, a damping shaft 2, a second arm, a damping shaft 3 and a camera. The base is fixedly connected to the upper end of the sign language recognition device body, the upper end of the base is fixedly installed with a damping turntable, the upper end of the damping turntable is fixedly installed with a pedestal, the upper end of the base is fixedly installed with a damping shaft 1, the inner side of the damping shaft 1 is provided with a first arm, the upper end of the first arm is fixedly installed with a damping shaft 2, the outer side of the damping shaft 2 is provided with a second arm, the inner side of the second arm is fixedly installed with a damping shaft 3, and the left side of the damping shaft 3 is provided with a camera.

[0007] Preferably, the selfie display screen mechanism includes a bracket, a damping shaft four, a rotating arm and a selfie display screen body. The bracket is fixedly connected to the upper end of the sign language recognition device body, the upper end of the bracket is fixedly installed with a damping shaft four, a rotating arm is provided on the right side of the damping shaft four, and the selfie display screen body is fixedly installed on the front side of the rotating arm.

[0008] Preferably, the bracket is fixed obliquely to the upper end of the main body of the sign language recognition device, and rotating arms are provided on both sides of the damping shaft four.

[0009] Preferably, the base, the damping turntable and the pedestal are all circular, and the damping turntable is fixedly installed between the base and the pedestal by screws.

[0010] Preferably, a second support arm is provided on both sides of the damping shaft 2, and a damping shaft 3 is provided on both sides of the camera.

[0011] Preferably, the camera is square as a whole, and the damping shaft three is located between the camera and the second arm.

[0012] Preferably, the translation display screen, camera and selfie display screen body are all electrically connected to the sign language recognition device body, and the translation display screen is tiltedly arranged at the upper end of the sign language recognition device body.

[0013] Compared with the prior art, the beneficial effects of the present invention are:

[0014] 1. This deep learning-based sign language recognition device, through the coordinated use of the damping dial, damping shaft 1, damping shaft 2, and damping shaft 3 of the camera mechanism, can rotate and adjust the base as well as the angles of the first arm, second arm, and camera. This allows for flexible adjustment of the shooting angle to accommodate the heights and body shapes of signers, thereby achieving the optimal shooting angle and ensuring complete and clear capture of various hand gestures and movement details, enhancing the user experience.

[0015] 2. This deep learning-based sign language recognition device can provide intuitive feedback on the signer's movements and the camera's captured image through the selfie display screen of the selfie display mechanism. This allows the signer to immediately see their movements reflected in the camera and can self-correct the camera's shooting angle to avoid incomplete images that affect sign language recognition. The damping shaft and the arm can be used to easily adjust the angle of the selfie display screen, making it suitable for people of different heights. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a schematic diagram of the first overall structure of the utility model;

[0017] Figure 2 This is a schematic diagram of the second overall structure of the utility model;

[0018] Figure 3 This is a schematic diagram of the installation structure of the first camera mechanism of the utility model;

[0019] Figure 4 This is a schematic diagram of the installation structure of the second camera mechanism of the present invention;

[0020] Figure 5 This is a schematic diagram of the installation structure of the selfie display screen mechanism of the present invention.

[0021] Among them: 1. Sign language recognition device body; 2. Power cord; 3. Camera mechanism; 301. Base; 302. Damping turntable; 303. Base; 304. Damping rotation axis 1; 305. First arm; 306. Damping rotation axis 2; 307. Second arm; 308. Damping rotation axis 3; 309. Camera; 4. Selfie display screen mechanism; 401. Bracket; 402. Damping rotation axis 4; 403. Rotating arm; 404. Selfie display screen body; 5. Translation display screen. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0023] See also Figure 1-5The utility model provides a technical solution: a sign language recognition device based on deep learning, comprising a sign language recognition device body 1, a power cord 2 is fixedly installed on the right side of the sign language recognition device body 1, a camera mechanism 3 is fixedly installed on the upper end of the sign language recognition device body 1, a selfie display mechanism 4 is fixedly connected to the upper end of the sign language recognition device body 1, a translation display 5 is fixedly connected to the upper end of the sign language recognition device body 1, the camera mechanism 3 comprises a base 301, a damping turntable 302, a base 303, a damping shaft 1 304, a first arm 305, a damping shaft 2 306, a second arm 307, a damping shaft 3 308 and a camera 309, the base 301 is fixedly connected to the upper end of the sign language recognition device body 1, a damping turntable 302 is fixedly installed on the upper end of the base 301, a base 303 is fixedly installed on the upper end of the damping turntable 302, A damping shaft 1 304 is fixedly installed on the upper end, a first arm 305 is provided on the inner side of the damping shaft 1 304, a damping shaft 2 306 is fixedly installed on the upper end of the first arm 305, a second arm 307 is provided on the outer side of the damping shaft 2 306, a damping shaft 308 is fixedly installed on the inner side of the second arm 307, and a camera 309 is provided on the left side of the damping shaft 308. By cooperating with the damping turntable 302, the damping shaft 1 304, the damping shaft 2 306 and the damping shaft 3 308 of the camera mechanism 3, the base 303 can be rotated and the angles of the first arm 305, the second arm 307 and the camera 309 can be adjusted. The shooting angle can be flexibly adjusted to adapt to the height and body shape of different signers, thereby achieving the best shooting angle, ensuring that various hand postures and movement details are captured completely and clearly, and enhancing the user experience.

[0024] See also Figure 2-5In this embodiment, the selfie display screen mechanism 4 includes a bracket 401, a damping shaft 402, a rotating arm 403 and a selfie display screen body 404. The bracket 401 is fixedly connected to the upper end of the sign language recognition device body 1. The upper end of the bracket 401 is fixedly installed with a damping shaft 402. The right side of the damping shaft 402 is provided with a rotating arm 403. The front side of the rotating arm 403 is fixedly installed with a selfie display screen body 404. The selfie display screen body 404 can provide an intuitive feedback to the signer's movements and the picture taken by the camera 309, allowing the signer to immediately see the picture of his or her movements in the camera 309, and can self-correct the shooting angle of the camera 309 to avoid incomplete pictures taken by the camera 309, which affects sign language recognition. The bracket 401 is tilted and fixed to the upper end of the sign language recognition device body 1. A rotating arm 403 is provided on both sides of the fourth axis 402. The setting of the damping rotating axis 402 and the rotating arm 403 can facilitate the adjustment of the angle of the selfie display screen body 404, which is suitable for people of different heights. The base 301, the damping turntable 302 and the base 303 are all circular. The damping turntable 302 is fixedly installed between the base 301 and the base 303 by screws. A second support arm 307 is provided on both sides of the damping rotating axis 2 306. A damping rotating axis 308 is provided on both sides of the camera 309. The camera 309 is square as a whole. The damping rotating axis 308 is located between the camera 309 and the second support arm 307. The translation display screen 5, the camera 309 and the selfie display screen body 404 are all electrically connected to the sign language recognition device body 1, and the translation display screen 5 is tilted and arranged at the upper end of the sign language recognition device body 1.

[0025] Working Principle: First, connect the power cord 2 to the socket to power on the sign language recognition device body 1, and connect the network cable. When in use, the damping dial 302, damping shaft 1 304, damping shaft 2 306 and damping shaft 3 308 of the camera mechanism 3 are used in conjunction to rotate and adjust the base 303 and adjust the angles of the first arm 305, second arm 307 and camera 309. The shooting angle can be flexibly adjusted to suit the height and body shape of different signers, thereby achieving the best shooting angle and ensuring that various hand gestures are captured completely and clearly. The gestures and action details can be enhanced to the user experience. The selfie display screen body 404 of the selfie display screen mechanism 4 can provide an intuitive feedback to the signer's actions and the shooting picture of the camera 309, so that the signer can immediately see the action picture of his or her own action in the camera 309, and can self-correct the shooting angle of the camera 309 to avoid incomplete shooting picture of the camera 309 and affect sign language recognition. The setting of the damping shaft 402 and the rotating arm 403 can facilitate the adjustment of the angle of the selfie display screen body 404, which is suitable for people of different heights.

[0026] Although the embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations may be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A sign language recognition device based on deep learning, comprising a sign language recognition device body (1), characterized in that: A power cord (2) is fixedly installed on the right side of the sign language recognition device body (1), a camera mechanism (3) is fixedly installed on the upper end of the sign language recognition device body (1), a selfie display mechanism (4) is fixedly connected to the upper end of the sign language recognition device body (1), and a translation display (5) is fixedly connected to the upper end of the sign language recognition device body (1); The camera mechanism (3) comprises a base (301), a damping turntable (302), a pedestal (303), a damping rotation shaft 1 (304), a first support arm (305), a damping rotation shaft 2 (306), a second support arm (307), a damping rotation shaft 3 (308) and a camera (309), wherein the base (301) is fixedly connected to the upper end of the sign language recognition device body (1), the damping turntable (302) is fixedly mounted on the upper end of the base (301), and the pedestal (303) is fixedly mounted on the upper end of the damping turntable (302). (303), a damping shaft 1 (304) is fixedly mounted on the upper end of the base (303), a first arm (305) is provided on the inner side of the damping shaft 1 (304), a damping shaft 2 (306) is fixedly mounted on the upper end of the first arm (305), a second arm (307) is provided on the outer side of the damping shaft 2 (306), a damping shaft 3 (308) is fixedly mounted on the inner side of the second arm (307), and a camera (309) is provided on the left side of the damping shaft 3 (308).

2. The deep learning-based sign language recognition device according to claim 1, characterized in that: The Selfie display screen mechanism (4) comprises a bracket (401), a damping rotating shaft four (402), a rotating arm (403) and a Selfie display screen body (404); the bracket (401) is fixedly connected to the upper end of the sign language recognition device body (1); the damping rotating shaft four (402) is fixedly mounted on the upper end of the bracket (401); a rotating arm (403) is provided on the right side of the damping rotating shaft four (402); and the Selfie display screen body (404) is fixedly mounted on the front side of the rotating arm (403).

3. The deep learning-based sign language recognition device according to claim 2, characterized in that: The bracket (401) is fixed obliquely to the upper end of the sign language recognition device body (1), and rotating arms (403) are provided on both sides of the damping rotating shaft (402).

4. The deep learning-based sign language recognition device according to claim 1, wherein: The base (301), the damping turntable (302) and the pedestal (303) are all circular, and the damping turntable (302) is fixedly installed between the base (301) and the pedestal (303) by screws.

5. The deep learning-based sign language recognition device according to claim 1, characterized in that: A second support arm (307) is provided on both sides of the second damping shaft (306), and a third damping shaft (308) is provided on both sides of the camera (309).

6. The deep learning-based sign language recognition device according to claim 1, characterized in that: The camera (309) is square in shape as a whole, and the damping shaft (308) is located between the camera (309) and the second support arm (307).

7. The deep learning-based sign language recognition device according to claim 1, characterized in that: The translation display screen (5), the camera (309) and the Selfie display screen body (404) are all electrically connected to the sign language recognition device body (1), and the translation display screen (5) is tiltedly arranged at the upper end of the sign language recognition device body (1).

Images