Bulk Comic Panel Text Removal

I have hundreds of coloured comic-book panels that need their speech balloons and captions removed while keeping every hue and gradient of the original artwork untouched. I already have a rough Python proof-of-concept, so you won’t be starting from scratch—I simply need it matured into a reliable, user-friendly executable that can churn through whole folders of images in one go.

Key requirements
• Source images are coloured PNG/JPEG files. Even better if whole CBR files
• Text must be detected automatically and erased by an inpainting method rather than clone-stamping or manual masking, so OpenCV, PyTorch or similar libraries are welcome.
• After processing, the script should run an automated check to confirm no visible text remains; an OCR pass is fine for this.
• Final deliverable is a one-click desktop executable (Windows) plus the updated, well-commented Python code.

Nice to have
– Simple GUI or command-line flags for input/output folders, image format filtering and log verbosity.
– Parallel processing to speed up large batches.
– A brief README covering setup, dependencies and how the verification step works.

I’ll provide the current script, a small sample set and expected outputs. Once the executable handles the sample flawlessly, we’ll test it on a larger batch before sign-off. Looking forward to seeing your approach!