Sign language recognition and translation system and method

Abstract

The invention belongs to the technical field of sign language translation software, and discloses a sign language recognition and translation system and method. The system comprises a front end module, a network connection module, an interaction synchronization module, an algorithm module, a data acquisition and processing module, a sign language data management module and a resource management module; each module is responsible for a function required in each step in the recognition process, and management of the data content supporting system continuous running. Data processing and feature extraction are performed according to the acquired sensor data. Pattern recognition and multi-element classification are performed on the processed data, and sign language actions are accurately and completely converted into natural language text information. We have designed the system and algorithm to ensure good recognition accuracy in the recognition process and to ensure a good user experienceduring use.

Description

technical field [0001] The invention belongs to the technical field of sign language translation software, and in particular relates to a sign language recognition and translation system and method with a complete client, server and data management that can be applied to multiple scenarios. Background technique [0002] Usually, the deaf-mute communicates through sign language, which requires both parties to communicate to understand sign language, and it takes a lot of time and energy for ordinary people to learn sign language. In order to solve the communication problem between the deaf-mute and non-sign language users, a sign language translation system with complete functions, high recognition accuracy and strong practicability is needed. [0003] Existing sign language translation systems are mainly based on computer vision or data gloves and other equipment for sign language recognition. Computer vision-based methods are easily affected by factors such as occlusion and...

AI Technical Summary

Problems solved by technology

[0003] Existing sign language translation systems are mainly based on computer vision or data gloves and other equipment for sign language recognition. Computer vision-based methods are easily affected by factors such as occlusion and lighting conditions, and have poor portability; while data-glove-based methods have easy Disadvantages of damage, inconvenient putting on and taking off and expensive

Although some sign language recognition methods and devices based on wearable devices have emerged, these methods and devices are still in the research and development stage, and a practical sign language translation product that combines mature algorithms and systematic management has not yet appeared

[0004] At the same time, at the current stage, there are many inconveniences in communication with service personnel for deaf-mute people in public service places such as hospitals and banks, and there is a lack of an effective solution. The current related solutions are either in the conceptual state or There are disadvantages such as incomplete function, low usability and high price

Images