QualityScaler 8.0 - image/video deeplearning upscaler with any GPU (BSRGAN/Pytorch)

CHANGELOG 8.0.

NEW

• Support for transparent png images  (IMPORTANT)
• Rewritten video-frame extraction using multi-thread approach
  •   now much faster  
• Rewritten video frame resize using multi-thread approach  
  •   now much faster
• Completely re-designed left bar with more options to select:
  • section "upscaled image/video frame extension" (.png / .jpg)
    •   .png is the best choise for quality, but is slower and upscaled files could be much bigger
    •   .jpg is 10% faster and upscaled files can be lighter but images and videos could have some artifacts
  •  number of cpus to use:
    • this value will influence video upscaling when extract frames from video and when resize frames

GUI

• Removed paypal button
• General UI changes
• Updated clean button icon

BUGFIX / IMPROVEMENTS

• Rewrote image management for better reliability and speed
• Fixed a bug that, when upscaling multiple photos the app could crash
• FIxed a problem that original .png images were being overwritten by temporary images produced by the app
• Updated external packages
• General code cleaning and many bugfix and improvements