Skip to main content

OCR Image Translation Settings

In this panel, you can not only select the recognition language that best suits your game scene and configure the translation language, but also customize image enhancement and the way text blocks are reorganized and merged into paragraphs through fine adjustment parameters, so as to handle various UI interfaces with complex layouts and noise. In addition, the overlay style, screenshot frame, and fixed mode of OCR image translation are also configured here.

Translation Language Settings​

  • OCR Recognition Language: Select the language family model used to recognize screenshot text. After upgrading to PP-OCRv6, the original three options "Chinese-English-Japanese / Traditional Chinese / Latin" have been merged into a unified "Multilingual" model:
    • Multilingual (Chinese/English/Japanese · Latin scripts) (default): A single model covers Chinese (including Traditional), English, Japanese, and Latin scripts, and is the first choice for the vast majority of game interfaces.
    • Korean, Cyrillic (Russian, etc.), Arabic, Greek: For cases where only one specific language family appears entirely, you can switch manually to eliminate possible false matches from the general model.
  • Translation Source Language: The source language of the OCR-recognized text; can be set to "Auto" for automatic language detection.
  • Translation Target Language: The language you want to translate into (usually your native language, such as Chinese).

Translation Overlay Settings​

Controls the text alignment, text and background colors, and depth presentation of the OCR image translation result overlay.

  • Text Alignment: The alignment of the translated text within the text block, choosing Auto / Left / Center / Right (default: Auto).
  • Text Color Mode: Choose Auto or Custom.
    • Auto: Automatically uses black or white text according to the brightness of each text block's background.
    • Custom: Uses the text color selected below.
  • Text Color: Only shown when the text color mode is "Custom"; select the font color (RGB) for the translated content.
  • Background Color Mode: Choose Auto or Custom.
    • Auto: Extracts the average color of the four corner pixels of each text block as its background color.
    • Custom: Uses the background color selected below.
  • Background Color: Only shown when the background color mode is "Custom"; default is white.
  • Background Transparency: If you are worried about blocking key game elements directly in front, you can pull this slider from 1.0 (completely opaque) to 0.0 (completely transparent). It takes effect in both color modes.
  • Text Block Depth Mode: Controls how the depth of translation text blocks is calculated in 3D space.
    • Overall Flat: All text blocks uniformly use the average depth after removing outliers, positioned on the same plane (recommended).
    • Spatially Independent: Each text block uses its own calculated depth, and outlier results that are too large or too small are replaced with the average depth.

Translation Screenshot Area Settings​

Provides recognition and selection area settings when the controller selection frame appears. You can also set the default projection area without relying on the controller's physical press to adjust the border.

  • Enable Controller Adjust Screenshot Frame: Whether to allow adjusting the screenshot frame size by holding the grip button and moving the controller in the game.
  • Screenshot Frame Color: The render color (RGB) of the frame. Default is light blue-cyan.
  • Screenshot Frame Transparency: Adjust the transparency of the screenshot frame (0.1 to 1.0, default 0.8); too low may make the frame line hard to see.
  • Screenshot Frame Width / Screenshot Frame Height: The proportion of the frame in the field of view (width default 33%, height default 21%).
  • Screenshot Frame X Position / Y Position: Fine-tune the value of the frame located slightly to the upper right and lower areas of the center of the field of view (35% to 65%, default 50%).
  • Screenshot Frame Depth Distance (m): Set the absolute three-dimensional depth pushed forward in screen space (0.2 to 5.0, default 2.0).

Image Enhancement​

Game screens are often not the highest quality "black text on white background" layouts. These switches are used to "organize" the image quality before passing the screenshot to the OCR engine:

  • Enhance Contrast: If the game scene is too dark (such as deep in a cave or backlit resulting in dim captured fonts), this can improve the distinction between text brightness and background.
  • Sharpen: When some VR games have too high anti-aliasing settings or insufficient resolution, font edges tend to be blurry. Enable this to reshape and sharpen native text edges.
  • Denoise: Suitable for image quality with heavy grain or lots of interference effects (such as retro filters, pixel art environments, fog particle interference, etc.), can avoid misreading as strange symbols.

OCR Text Post-processing​

Since game interface dialog boxes often have their own wild character spacing and line breaking rules, some line breaks or empty lines will be simply treated as independent sentences, resulting in a "car crash scene" for a single translation. All the sliders below (with reset ↺ function) can greatly change the system's merging judgment behavior.

Preset Merging Parameters One-click Strategy​

Pre-set combinations of experience values. When you arbitrarily pull any of the detailed sliders below, it will also automatically jump to the "Custom" configuration slot.

  • Vertical Preset: None / Narrow / Standard (default) / Wide (tolerates more distance breaks caused by messy lines) / Custom

Text Block Horizontal Merging Tolerance​

  • Line Grouping Y Coordinate Tolerance: Limits the overall height difference between two connected text blocks. Through this filter, you can avoid the danger of "UI element impurities" with the same width but different sizes being merged into one sentence. (Default 0.2)
  • Horizontal Merge Height Tolerance: When processing long single-line text, limits the proportional extreme value that can cross the highest and lowest fonts within the horizontal distance. (Default 0.1)
  • Horizontal Overlap Threshold (px): Set from -20 to 20px. Negative numbers allow character spacing, positive numbers require mutual overlap or character overlap. (Default -2)
  • Horizontal Merge Y Difference Tolerance: The longitudinal cross-section difference range established to determine whether two horizontally arranged word groups "belong to the same line". (Default 0.5)

Text Paragraph Vertical Reconstruction and Line Break Judgment​

  • Vertical Merge Height Tolerance: Limits the vertical limit tolerance when associating context. (Default 0.3 / 0.5)
  • Vertical Merge Alignment Tolerance: Limits how many character widths the offset of the first character alignment of two adjacent lines above and below cannot exceed. (Default 1.5)
  • Vertical Spacing Maximum Ratio: The empty line truncation limit allowed in the same translation request; exceeding this spacing will be discarded or treated as different sentences. (Default 0.15)
  • Vertical Merge Angle Tolerance (degrees): Deals with some dialogue panels with inclined designs or radians. Unit is degrees. (Default 2.0°)
  • Vertical Merge Minimum Overlap Ratio: The minimum benchmark allowed when vertically connected elements have top-bottom edge overlap. (Default 0.0)
  • Vertical Paragraph Spacing Multiplier: When detecting that the vertical distance exceeds how many times the system standard line spacing, the content below is recognized as a new paragraph, triggering a forced line break. (Default 1.5)
  • Vertical Minimum Paragraph Spacing (px): Same as the above judgment requirement, sets the minimum fixed physical pixel threshold for forced new paragraphs. (Default 80 to 150px)

Translation Overlay Fixed Mode​

Controls the display method of OCR translation result patches:

  • Fixed in World Space: The translation patch is anchored to a specific position in the VR scene, and rotating the head will not follow the movement, with a real three-dimensional sense of space. (Default option)
  • Fixed in Front Field of View: The translation patch always follows your line of sight direction, displayed at a fixed position directly in front of the field of view, suitable for scenes where you need to view translations at any time.