D-Ogi/WatermarkRemover-AI

AI-Powered Watermark Remover using Florence-2 and LaMA: Remove watermarks from images and videos, including AI-generated content from Sora, Runway, and others. Features a modern PyWebview GUI.

What it solves

It provides an automated way to remove watermarks from images and videos, specifically targeting AI-generated content from platforms like Sora and Runway, while maintaining a natural look in the resulting media.

How it works

The tool uses a two-step AI process: first, Microsoft's Florence-2 vision model identifies the location of the watermark. Second, the LaMA (Large Mask Inpainting) model fills in the removed area by predicting the surrounding pixels to create a seamless blend.

Who it’s for

Content creators and users who need to clean up watermarks from visual media, including those working with AI-generated videos.

Highlights

  • AI-Powered Detection: Uses Florence-2 for precise watermark identification.
  • Video Support: Handles multiple video formats and supports audio preservation via FFmpeg.
  • Batch Processing: Ability to process entire folders of files at once.
  • Flexible Control: Includes a preview mode to check detections before processing and options to handle watermarks that fade in or out.
  • Accessible UI: Features a modern, multi-language GUI built with PyWebview.

Related

  • Project
    photonixapp/photonix

    A self-hosted photo management app that uses object recognition and color analysis to automatically organize and filter photo collections.

  • Project
    zs1083339604/FaceWinUnlock-Tauri

    A Windows facial recognition unlock tool that uses OpenCV and a custom Credential Provider to enable face-unlock capabilities on devices without native Windows Hello hardware.

  • Project
    microsoft/GlobalMLBuildingFootprints

    Microsoft’s Global Building Footprints repo releases a >1 billion‑building, open‑licensed dataset of polygon footprints (and many height estimates) generated by deep‑learning segmentation models. The data are provided as gzipped line‑delimited GeoJSON files, indexed by Bing quad‑keys, with confidence scores and documentation for large‑scale GIS or AI‑driven urban analysis.

  • Project
    apple-aiml-research/ml-cvnets

    A computer vision toolkit for training and evaluating standard and novel mobile and non-mobile models for tasks like classification, detection, and segmentation.