How do you find similarity percentage from euclidean distance of the embeddings

[Code-beep12]

Two images which showed a 98.838% similarity has a euclidean distance of 0.7619. How would i go about finding the similarity percentage. My aim is to store embeddings and later find the similarity of faces rather than calling the verify api of compreface repeatedly

[pospielov]

You can find the formula that we use to calculate the similarity from Euclidean distance here: https://github.com/exadel-inc/CompreFace/blob/d10cf2f1d8d6f4a55ba918f391509a402e5b59eb/java/api/src/main/java/com/exadel/frs/core/trainservice/component/classifiers/EuclideanDistanceClassifier.java#L142 And coefficients here: https://github.com/exadel-inc/CompreFace/blob/d10cf2f1d8d6f4a55ba918f391509a402e5b59eb/embedding-calculator/src/services/facescan/plugins/facenet/facenet.py#L141 https://github.com/exadel-inc/CompreFace/blob/d10cf2f1d8d6f4a55ba918f391509a402e5b59eb/embedding-calculator/src/services/facescan/plugins/insightface/insightface.py#L136 However, if you don't need to compare similarities with CompreFace, you can just use distance and choose the threshold based on it. We created similarity just to simplify understanding of choosing threshold.

[Code-beep12]

Thank you so much