VideOCR v1.5.1

Hotfix release:

• Fixes an issue where the program did not correctly track the start time of subtitles
• Removes stitching of images in the recognition step due to degraded accuracy in comparison to v1.4.2
• Fixes a crash during ETA calculation while using Google Lens as the OCR engine

Main Highlight of the v1.5.x release:

• Adds support for Google Lens OCR as a second OCR engine for highly accurate text recognition
  --> 200+ languages are now supported

Notes:

• Utilizes standalone version of PaddleOCR from https://github.com/timminator/PaddleOCR-Standalone
• Utilizes standalone version of Chrome-Lens-OCR from https://github.com/timminator/Chrome-Lens-OCR
• Linux version requires glibc 2.35 or newer (can be checked by running ldd --version in a terminal)

Download Tips:

• Windows:
  • CPU version:
    VideOCR-CPU-v1.5.1-setup-x64.exe
  • GPU version:
    VideOCR-GPU-v1.5.1-CUDA-11.8-setup-x64.exe (Nvidia 10 Series graphics cards)
    VideOCR-GPU-v1.5.1-CUDA-12.9-setup-x64.exe (Nvidia 16 - 50 Series graphics cards)
• Linux:
  • CPU version:
    VideOCR-CPU-v1.5.1-Linux.7z
  • GPU version:
    VideOCR-GPU-v1.5.1-CUDA-11.8-Linux.7z (Nvidia 10 Series graphics cards)
    VideOCR-GPU-v1.5.1-CUDA-12.9-Linux.7z.001 (Nvidia 16 - 50 Series graphics cards)
    VideOCR-GPU-v1.5.1-CUDA-12.9-Linux.7z.002 (Nvidia 16 - 50 Series graphics cards)