dtoro/mule-image
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix: raise face detection threshold to 0.85, tighten clustering

Browse Source 
Eliminates false positives (dog faces, painting faces) by requiring
score >= 0.85. Tighten cluster eps to 0.25 for better separation.
Tested: 4 real faces from 2 photos, no false positives.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This commit is contained in:
authentik Default Admin
2026-04-10 11:09:43 +02:00
parent 5f3fa5240e
commit 40d570f2c2
2 changed files with 4 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
backend/app/config.py
View File

@@ -50,8 +50,8 @@ class FacesSettings(BaseModel):
"""YuNet + SFace face detection/recognition settings""" """YuNet + SFace face detection/recognition settings"""
enabled: bool = True enabled: bool = True
min_face_size: int = 40 min_face_size: int = 40
recognition_threshold: float = 0.7 recognition_threshold: float = 0.85
cluster_eps: float = 0.35 cluster_eps: float = 0.25
class VisionSettings(BaseModel): class VisionSettings(BaseModel):
"""AI vision pipeline settings. Disabled when running on SQLite """AI vision pipeline settings. Disabled when running on SQLite

4
mulita.yml
View File

@@ -44,6 +44,6 @@ vision:
faces: faces:
enabled: true enabled: true
min_face_size: 40 min_face_size: 40
recognition_threshold: 0.7 recognition_threshold: 0.85
cluster_eps: 0.35 cluster_eps: 0.25
worker_concurrency: 2 worker_concurrency: 2