Faces: Update README.md #5167

Signed-off-by: Michael Mayer <michael@photoprism.app>
This commit is contained in:
Michael Mayer 2025-10-08 15:08:37 +02:00
parent 8571600a1b
commit 3efb69483e

View file

@ -59,7 +59,7 @@ All embeddings, regardless of origin, are normalized to unit length (‖x‖₂
- Face clusters update their sample statistics (`Samples`, `SampleRadius`) from the latest matches via `Face.UpdateMatchStats`, avoiding stale radii during optimize loops.
- Cluster materialisation now pre-sizes buffers; `BenchmarkClusterMaterialize` reports ~14.8µs/op with 64 allocations (≈56KB) versus the legacy ~29.8µs/op with 384 allocations (≈105KB).
This guarantees that Euclidean distance comparisons are equivalent to cosine comparisons, aligning our thresholds with FaceNet literature.
This guarantees that Euclidean distance comparisons are equivalent to cosine comparisons, aligning our thresholds with [FaceNet](https://maucher.pages.mi.hdm-stuttgart.de/orbook/face/faceRecognition.html) literature.
#### Face Kind Reference