AI-Powered Canvas Maze Solver & CV Pipeline

1. Source Image Feed

Drag & drop your maze drawing here, or browse

Accepts PNG, JPG, JPEG (Adaptive high-fidelity scaling supported)

2. CV Preprocessing Pipeline

Max Processing Resolution Use higher resolutions (e.g., 1200px-1600px) to prevent thin walls and pathways from merging during downscaling. 800px

Otsu Threshold Offset Offset applied to the calculated global Otsu threshold. Fine-tunes edge separation. 0

Invert Palette

White Walls

8-Way Search

Corner-Cut Safe

Morphological Filter (Erosion/Dilation) Negative values erode corridors (thicken walls to seal gaps). Positive values dilate corridors (thins walls to bridge corridor spaces). 0 (None)

3. Path-Centering & Snapping

A* Wall-Centering Strength Higher values heavily penalize movements close to walls. Snaps the solved path to the perfect center of corridors. 4.0

Snapping Placement Tool

4. Overlay Visualization

Show CV Binarized Image Overlay Corridor Distance Map

Original Image Opacity 100%

Path Color

#00ffcc

Line Width

Metrics & Diagnostics

Input Grid Size -

Otsu Threshold -

CV Process Time -

A* Solve Time -

Nodes Evaluated -

Path Length -

System Status

Waiting for Maze Image Input...

Interactive Design Area

Select a sample preset above, or drop custom maze images inside the left panel to launch computer vision workspace

Start Anchor (Green) End Anchor (Red) Centered Corridor Vector Black Boundaries (Walls)