# US 2025/0375710 A1 · Real Time Translation Method for Games using Machine Learning Model

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2025/0375710 A1
- Document ID: US20250375710A1
- Status: Published application
- Kind: A1
- Application number: 18/736,546
- Filed: 2024-06-07
- Published: 2025-12-11
- Earliest claimed: 2024-06-07
- Inventors: Shih-Pang Wang; Guan-Lin Chen
- Web page: https://patentgrep.com/patent/US20250375710A1

## CPC

- A63F 13/67: adaptively or by learning from player actions, e.g. skill level adjustment or by storing successful combat sequences for re-use
- A63F 13/52: involving aspects of the displayed game scene
- G06F 40/58: Use of machine translation, e.g. for multi-lingual retrieval, for server-side translation for client devices or for real-time translation
- G06V 20/46: {Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames}
- G06V 30/19147: {Obtaining sets of training patterns; Bootstrap methods, e.g. bagging or boosting}
- A63F: CARD, BOARD, OR ROULETTE GAMES; INDOOR GAMES USING SMALL MOVING PLAYING BODIES; VIDEO GAMES; GAMES NOT OTHERWISE PROVIDED FOR
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING

## Abstract

A real time translation method for a game includes extracting features from video frames using computer vision and a database, performing an association process to find a machine learning model best matching the features for translation, obtaining texts in the game through optical character recognition (OCR), preprocessing the texts, translating the texts using the machine learning model to generate translated texts, and rendering the translated texts to images of the video frames for displaying the images with the translated texts on a display device.

## Claims

### Claim 1

A real time translation method for a game, comprising:
extracting features from video frames using computer vision and a database;
performing an association process to find a machine learning model best matching the features for translation;
obtaining texts in the game through optical character recognition (OCR);
preprocessing the texts;
translating the texts using the machine learning model to generate translated texts; and
rendering the translated texts to images of the video frames for displaying the images with the translated texts on a display device.

### Claim 2 (depends on 1)

The method in claim 1, further comprising:
after preprocessing the texts, loading contexts from retrieval-augmented generation (RAG) and completion cache.

### Claim 3 (depends on 2)

The method in claim 2, further comprising:
if the translated texts are not reliable, loading contexts from the retrieval-augmented generation (RAG) and the completion cache again.

### Claim 4 (depends on 2)

The method in claim 2, wherein rendering the translated texts to the images of the video frames for displaying the images with the translated texts on the display device is performed if the translated texts are reliable.

### Claim 5 (depends on 1)

The method in claim 1, wherein preprocessing the texts comprises embedding, text splitting, clustering, map reducing, and/or refining the texts.

### Claim 6 (depends on 1)

The method in claim 1, wherein performing the association process to find the machine learning model best matching the features for translation comprises:
selecting the machine learning model best matching the features from N machine learning models;
wherein N is a positive integer.

### Claim 7 (depends on 6)

The method in claim 6, further comprising:
training the N machine learning models.

### Claim 8 (depends on 7)

The method in claim 7, wherein training the N machine learning models comprises:
preprocessing training texts;
performing another association process to find a machine learning model best matching features of training images containing the training texts from N machine learning models; and
training the machine learning model best matching the features of the training images with the training texts and answers.

### Claim 9 (depends on 8)

The method in claim 8, wherein preprocessing the training texts comprises embedding, text splitting, clustering, map reducing, and/or refining the training texts.

### Claim 10

A real time translation method for a game, comprising:
extracting features from video frames using computer vision and a database;
performing an association process to find weightings of N machine learning models best matching the features for translation;
obtaining texts in the game through optical character recognition (OCR);
preprocessing the texts;
translating the texts using the N machine learning models with the weightings to generate translated texts; and
rendering the translated texts to images of the video frames for displaying the images with the translated texts on a display device.

### Claim 11 (depends on 10)

The method in claim 10, further comprising:
after preprocessing the texts, loading contexts from retrieval-augmented generation (RAG) and completion cache.

### Claim 12 (depends on 11)

The method in claim 11, further comprising:
if the translated texts are not reliable, loading contexts from the retrieval-augmented generation (RAG) and the completion cache again.

### Claim 13 (depends on 11)

The method in claim 11, wherein rendering the translated texts to the images of the video frames for displaying the images with the translated texts on the display device is performed if the translated texts are reliable.

### Claim 14 (depends on 10)

The method in claim 10, wherein preprocessing the texts comprises embedding, text splitting, clustering, map reducing, and/or refining the texts.

### Claim 15 (depends on 10)

The method in claim 10, further comprising:
training the N machine learning models.