# 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`).