User works with a collection of photos representing full-page layouts—such as scanned publications or photographs of photo album pages. Each page contains multiple embedded images.
Workflow involves:
- Applying global transformations to the entire page (e.g., rotation, color correction, perspective adjustment).
- Automatically or manually extracting each embedded image from the layout.
- Applying a custom set of transformations to each extracted image individually.
Desired Capabilities
- Automatic Sub-Image Detection as a Transformation Node
Introduce a new transformation type: "Detect Embedded Images", which identifies and isolates individual images within a page layout.
This transformation would iterate over all detected sub-images, treating each as a separate unit for downstream processing. - User-Guided Iteration and Control in the Transformation Tree
Enable manual intervention at a designated breakpoint in the transformation tree:- User can manually crop or warp to define a sub-image.
- After writing the processed image to file, provide a "Loop" or "Go Back" node that lets the user:
- Continue extracting and processing additional sub-images from the same page.
- Or proceed to the next page in the batch.