ScoreSight - Free Open Source OCR tool for Gaming and Scoreboards

ScoreSight - Free Open Source OCR tool for Gaming and Scoreboards v0.3.2

Kind of a shame that this wasn't developed as an OBS plugin, I'd like to display the same capture device both in OBS and ScoreSight but the device can only be used on on either one of them and not both. Probably could circumvent this using OBS Virtual Camera but it'd add CPU overhead to the operation. I'm trying to use the other OCR plugin you've developed though (no luck detecting any text unlike this app though).
 
Does anyone know if ScoreSight runs on the latest macOS versions? I noticed the last update of ScoreSight was in 2024 and it's still version 0.1. I’d love to start using it, but I can't test it myself right now since I still need to install an SDI capture card enclosure. Is anyone currently using ScoreSight in a live production environment?

The Scoresight tutorial videos look very cool!
 
royshilkrot updated ScoreSight - Free Open Source OCR tool for Gaming and Scoreboards with a new update entry:

ScoreSight is back baby! Overhauled UI, new backend, faster and better than ever!

ScoreSight is actively maintained again. This release modernizes the whole
stack, redesigns the window, fixes the most reported problems, and is tested
end to end on every platform before it ships. Highlights:

  • A redesigned window. The field list with large live values, the picture
    and a settings panel (Box and Outputs) side by side, and a Live
    layout
    (Ctrl+2) for game time. Each output (text files, HTTP server, OBS,
    vMix, UNO, HTTP API) has an on/off...

Read the rest of this update entry...
 
royshilkrot updated ScoreSight - Free Open Source OCR tool for Gaming and Scoreboards with a new update entry:

v0.3.2 - A new OCR engine, live overlays, and many more broadcast products.

Added
  • PP-OCR, a much more accurate OCR engine. PaddleOCR's PP-OCRv6 models
    read 95% of real scoreboard crops exactly in our evaluation, against 31%
    for Tesseract with its default settings. PP-OCR (automatic) is the
    default for new setups: it uses the accurate model, and switches to the
    fast one (90%) when the computer can't keep up, and says so. Existing setups
    keep their Tesseract model until you choose another. OCR also runs on its
    own thread now, so a slow reading...

Read the rest of this update entry...
 
Back
Top