> For the complete documentation index, see [llms.txt](https://docs.augelab.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.augelab.com/function-blocks/blocks-reference/detections-shapes/shape-analysis/most-similar-shape.md).
# Most Similar Shape
This function block compares a reference contour against a list of candidate contours and selects the most similar one. It returns the best-matching contour, a similarity score (0-1), and an annotated image showing the selected contour.
## đĨ Inputs
* `Contour` Reference contour used for comparison
* `Contour(s)` List of candidate contours to compare against
* `Image` Optional image used for visual output/annotation
## đ¤ Outputs
* `Most Similar` The contour from the candidate list that best matches the reference
* `Score (0-1)` Similarity score (1.0 means identical, lower values indicate less similarity)
* `Image Any` Optional annotated image with the selected contour highlighted
## đšī¸ Controls
This function block has no user-configurable controls. It runs automatically when inputs are provided.
## đ¨ Features
* Fast comparison of shapes to find the closest match in a list.
* Produces both a numeric similarity score and a visual result for verification.
* Works with any contour format accepted by the system.
## đ Usage Instructions
1. Provide a reference contour to `Contour` input.
2. Provide one or more candidate contours to `Contour(s)` input.
3. Optionally supply an image to `Image` to receive an annotated visual showing which contour was chosen.
4. Run the scenario; outputs will include the selected contour, its similarity score, and the annotated image if supplied.
## đ Evaluation
When evaluated, this block measures resemblance between the reference contour and each candidate. It returns the candidate with the best resemblance (highest similarity) along with a numeric score and an annotated image (if provided).
## đĄ Tips and Tricks
* To extract contours from a binary image first, use `Find Contour` to get contours suitable for comparison.
* Use `Approximate Contour` to simplify contours (reduce point count) before comparison when shapes are noisy or overly detailed.
* Combine with `Hull Convex` to compare convex hulls instead of raw contours when outer shape similarity is desired.
* Use `Minimum Rectangle`, `Minimum Ellipse`, or `Minimum Circle` to convert contours into simple geometric descriptors for alternate comparison strategies.
* Preprocess input images with `Blur`, `Denoising`, or `Image Threshold` to reduce noise and improve contour extraction.
* If you want to visually inspect the selected result, connect the annotated image output to `Show Image`.
* Use `Contour to Image` to crop or isolate the region around a contour for further inspection or downstream processing.
* Draw the selected contour on live output using `Draw Result On Image` or `Draw Detections` for clearer visualization in dashboards.
* If you only need to compare contours within a sub-area, crop first with `Image ROI` or `Image ROI Select` to limit candidates and speed up processing.
* For batch comparisons over multiple regions, use `Image ROI Polygon` or `Image ROI Polygon Multi` to generate multiple areas and feed their contours into this block.
## đ ī¸ Troubleshooting
* If similarity scores are unexpectedly low, ensure contours are extracted from a clean binary image (use `Image Threshold` and `Blur` as needed).
* If multiple candidates produce similar scores, simplify contours with `Approximate Contour` or compare bounding shapes (`Minimum Rectangle`, `Minimum Ellipse`) to reduce sensitivity to small differences.
* If the visual output does not show the selected contour, verify an image was provided to the `Image` input or connect the block output to `Show Image` for preview.