Add preview pane with visual interface to improve crop UX

[drockee]

Issue: Add preview pane with visual interface to improve crop UX

Summary

• Add preview pane that can show one or more visual interfaces for viewing input scan images and determining 3D reconstruction region limits.

• Proposed cropping interfaces

  • View 1: Side view axial rotation via adjustable slider from perspective of the detector with crop window overlay. No other view is required to set crop parameters for X, Y, and Z axis.
  • View 2: Top view resizable translucent circular FOV mask overlay. May require a preliminary low resolution 3D reconstruction. Requires a secondary view to define z axis cropping.

• Proposing that the visual interface can set any of the following 3D reconstruction volume limiting parameters:

  • X, Y and Z sample count (set_numX, set_numY, set_numZ)
  • Circular FOV mask (set_diameterFOV)
  • XYZ offset (set_offsetX, set_offsetY, set_offsetZ)

What to implement:

• Preview pane to hold side and top views

• Side view:

  • Image display + crop overlay (drag/resize rectangle) + input image slider.
  • Mapped x,y,w,h (px) readout
  • Example for side view crop UI

• Top view:

  • Produce a small, fast reconstruction (downsampled) to obtain a top view image.
  • Able to select one of the following crop overlays
    • Resizable translucent circular FOV mask overlay
    • Circular mask diameter (mm) readout.
    • Drag/resize rectangle

• Checkbox settings:

  • Select which volume limiting parameters to update when crop is applied. Chose between the following
    • X sample count and offset
    • Y sample count and offset
    • Z sample count and offset
    • Circular FOV mask (set_diameterFOV) and offset (set_offsetX, set_offsetY)

• Buttons: Apply Crop.

• On Apply Crop:

  • Map crop overlay to selected volume limiting parameters.

Why:

• Improves user confidence and reduces errors by letting operators visually select the 3D reconstruction region rather than guessing numeric grid and offset values.
• Speeds workflow: a visual crop maps directly to LEAP's setters (set_numX/Y/Z, set_offsetX/Y/Z, set_diameterFOV), removing manual conversion.
• Reduces compute and/or memory waste by encouraging tight grid bounds or FOV masks before a full reconstruction is completed.
• Low-risk: reuses existing parameter setters. No changes to core LEAP reconstruction implementation required.

Acceptance criteria

1. UI elements

   • A preview pane exists in leapctrails/leapctrails.py with two sub-views: Side view (image slider + rectangular crop overlay) and Top view (small downsampled reconstruction + resizable circular FOV overlay).
   • Checkboxes are present to select which parameters to update on apply: X count, Y count, Z count, circular FOV.
   • An Apply Crop button is present and enabled when at least one parameter checkbox is selected.

2. Side view behavior

   • The side view shows input projections and supports dragging/resizing a rectangular overlay.
   • A numeric readout displays:
     • mapped x,y,w,h in detector pixels
     • circular FOV diameter and offset computed from the vertical crop lines
     • XYZ sample counts and offsets

3. Top view behavior

   • The top view shows a quick low-resolution reconstruction (downsampled) of the scan volume sufficient to place a circular mask.
   • If selected, a translucent circular overlay is resizable and displays current diameter in mm. Alternatively, show standard drag/resizable crop rectangle overlay.

4. Mapping & wiring

   • On Apply Crop, the UI maps the selected overlays to the corresponding LEAP API calls:
     • Rectangle crop
       1. Computed numX/numY/numZ and offsetX/offsetY/offsetZ where checkboxes selected.
       2. Computed FOV mask set_diameterFOV(d_mm) and set_offsetX/Y() in place of numX/numY if checkbox selected.
     • Circle mask crop
       1. call set_diameterFOV(d_mm) and set_offsetX/Y() to center the FOV.
   • Implementation must use existing methods to set params

Files / entry points to change (where to implement)

• leapctrails/leapctrails.py — add preview readout, toggle, crop UI elements, and wiring for Apply.
• leapctrails/ct_volume_page.py — use existing push_numX/push_numY/push_numZ/ set_diameterFOV and offset push methods to set limits.

[drockee]

I'm going to implement this functionality for the LEAP CT GUI. As I work through this, looking for feedback on the proposed cropping interface views, specifically:

• Are the proposed cropping interfaces and overlays sensible from the perspective of both the user and LEAP's capabilities.
• Are there additional parameters or constraints I should be aware of.
• Are there performance or other concerns for producing the quick down-sampled top view reconstruction for the fov diameter crop mask?

Any suggestions to consider before I start are welcome.