INNOVATIVE METHOD AND DEVICE FOR INSTANT TRANSLATION OF SIGN LANGUAGE, AND VICE VERSA, BASED ON ARTIFICIAL INTELLIGENCE

Abstract

The invention relates to an application solution usable on mobile devices (smartphones, tablets) characterized by the instant translation of speech recognition (analysis of incoming acoustic signals and conversion to audio) or written text into French Sign Language, rendered as a video. (1) Conversely, the solution translates signs filmed by mobile devices into written text or audio in French. (1) This solution is based on artificial intelligence (2) with real-time access to cloud databases (remote computer servers hosted on the Internet). (3) It instantly detects, identifies, and interprets the user's gestures, speech, or written text. The solution includes a correction system based on spelling and grammar rules and common usage to produce a perfect text or sign, thus avoiding any adverse impact on accuracy.The solution also incorporates machine learning, which allows it to follow changes in the language and continuously improve its results (4).

Description

Title of the invention: INNOVATIVE METHOD AND DEVICE FOR THE INSTANT TRANSLATION OF SIGN LANGUAGE, AND VICE VERSA BASED ON ARTIFICIAL INTELLIGENCE

[0001] A device that allows sign language and spoken language to be translated using artificial intelligence. It detects sign movements and then converts them into spoken or readable form.

[0002] And regarding oral languages, this is converted into a sign language version in the form of a video.

[0003] In addition, it has automatic learning of its languages, so the translation will be more reliable and more constructive as its learning progresses.

[0004] Thanks to this invention, communication between individuals using sign language and people using oral languages ​​will be more accessible.

[0005] [Fig.l]

[0006] This figure represents the device for translating sign and spoken languages ​​using artificial intelligence.

[0007] The deployed AI fulfills several objectives:

[0008] -> Detect sign language signs instantly using data collected via the camera of mobile devices.

[0009] -> Instantly correct errors in spoken language. Machine Learning intervenes during this phase: When a message presents an error after translation, the device notices it and compensates for it.

[0010] -> Translate languages

[0011] The cloud is the element used to store the language data collected to carry out the translation.

[0012] Example of the procedure for translating a sign language into an oral language:

[0013] The AI ​​begins by detecting the signs (gestures) made by the individual to register. Based on the data stored in the cloud (which acts as a dictionary), the AI ​​offers a translation of the signed message.

[0014] The functionality resulting from machine learning then takes over to detect and compensate for any errors. For example, if the individual signs "How old are you?", the machine learning will have to return the following corrected message: "How old are you?"

[0015] Additionally, the next time the "How old are you?" message is signed, the tool will have learned this error and will automatically correct it.

[0016] Once the translation is done, the AI ​​will generate an audio version of the message.

[0017] As the device is used, the AI ​​will learn new message translations (by storing them in the cloud) to reinvest them later.

Claims

Claims

1. 1) The integration of instant machine learning when translating sign language text or audio into video format

2. 2) The integration of artificial intelligence when translating a text into sign language and vice versa • 2.1) Use of artificial intelligence for instant gesture recognition • 2.2) Use of artificial intelligence for instant data correction

3. 3) Generate the sign language translation in video form

Images