Match Shapes

This function block compares two shapes (contours) and returns a resemblance score between 0 and 1. It is useful when you want a single numeric measure of how similar two contours are — where values closer to 1 mean the shapes are more similar.

📥 Inputs

Contour — First contour to compare (reference).

Contour — Second contour to compare (query).

📤 Outputs

Resemblence — A number between 0 and 1 expressing how similar the two contours are (higher = more similar).

🕹️ Controls

Resemblence — A read-only label that displays the current resemblance value for quick feedback in the UI.

⚙️ Running mechanism

When the block runs it accepts two contour inputs, converts them into comparable representations, and computes a single resemblance score. The score is presented both as the output value and on the block's Resemblence label so you can see results immediately.

🎯 Features

• Quick numeric similarity score for two contours.

• UI feedback via a live resemblance label.

• Works with contours produced by other shape analysis and detection blocks.

📝 How to use

1. Provide two contours to the inputs. Contours can come from detector or contour-processing blocks.

2. Run the flow; the block returns Resemblence and updates the label for instant inspection.

3. Use the numeric output for decisions, logging, or further processing.

💡 Tips and Tricks

• If you only have an image, extract contours first using the Find Contour block, then feed the resulting contours into this block.

• To simplify complex contours (reduce noise or number of points) try Approximate Contour before matching; simpler contours often yield more stable resemblance values.

• If you need the contour cropped into an image for visual comparison, use Contour to Image and preview the result with Show Image.

• To compare multiple candidate contours against a reference, combine this block with Most Similar Shape to automatically pick the best match.

• When matching shapes in specific regions, crop the image first using Image ROI Select so contours come from the expected area.

• If contours originate from images with different scales, normalize sizes with Image Resize or convert contours to cropped images of the same dimensions using Contour to Image.

🛠️ Troubleshooting

• If the resemblance seems unexpectedly low: ensure both contours are similarly scaled and oriented. Try simplifying the contours with Approximate Contour or normalizing their image sizes.

• If the output is unstable across frames, check the contour extraction step (e.g., Find Contour) for noisy detections and consider preprocessing with Blur or Image Threshold.

• No output or invalid contour input: verify the contour-producing block actually returns contour data (use Debug Input or preview intermediate images with Show Image).