UniFace: A Unified Face Analysis Library for Python | Detection, alignment, landmarks, face-mesh, recognition, parsing, gaze, attributes and anti-spoofing under one API.
Python computer-visiondeep-learningface-alignmentface-analysisface-anonymizationface-anti-spoofingface-attributesface-detectionface-meshface-parsingface-quality-assessmentface-recognitionface-segmentationface-trackingfacial-landmarksgaze-estimationhead-pose-estimationonnxruntimeportrait-mattingpython
feat: Update FaceAnalyzer defaults and remove Spanish README (#150)
Yakhyokhuja Valikhujaev8e30484351b0
diff --git a/CHANGELOG.md b/CHANGELOG.mdindex 53826c2..0b6d623 100644--- a/CHANGELOG.md+++ b/CHANGELOG.md@@ -5,6 +5,12 @@ the autogenerated notes on the [GitHub releases page](https://github.com/yakhyo/## Unreleased+### Changed++- `FaceAnalyzer()` now defaults to RetinaFace `MNET_V2` + EdgeFace `XXS`+ instead of SCRFD `500M_KPS` + ArcFace `MNET`. Pass the old models explicitly+ to keep using them.+### Added- AdaFace `IR_50` weights (WebFace4M, 166 MB): 95.44% IJB-B, 96.98% IJB-C.diff --git a/README.es.md b/README.es.mddeleted file mode 100644index 15d7ddf..0000000--- a/README.es.md+++ /dev/null@@ -1,190 +0,0 @@-\[[English](README.md)\] | \[[Chinese](README.zh-CN.md)\] | \[[Korean](README.ko.md)\] | \[[Spanish](README.es.md)\] | \[[Uzbek](README.uz.md)\]--<h1 align="center">UniFace: Una biblioteca unificada de análisis facial para Python</h1>--Esta página es la traducción al español del [README en inglés](README.md). La documentación enlazada está principalmente en inglés. Al actualizar la traducción, mantén el contenido sincronizado con la versión inglesa.--<div align="center">--[](https://pypi.org/project/uniface/)-[](https://www.python.org/)-[](https://opensource.org/licenses/MIT)-[](https://github.com/yakhyo/uniface/actions)-[](https://pepy.tech/projects/uniface)-[](https://www.kaggle.com/yakhyokhuja/code)-[](https://huggingface.co/spaces/yakhyo/uniface)--</div>--<div align="center">- <img src="https://raw.githubusercontent.com/yakhyo/uniface/main/.github/logos/uniface_rounded_q80.webp" width="90%" alt="UniFace — Una biblioteca unificada de análisis facial para Python">-</div>--<p align="center">- UniFace es una biblioteca ligera de Python, lista para producción, para detección y reconocimiento facial,<br>- seguimiento, análisis de puntos de referencia, segmentación de partes del rostro, estimación de la mirada y atributos faciales.-</p>--<p align="center">- <a href="https://yakhyo.github.io/uniface/quickstart/"><img src="https://img.shields.io/badge/Primeros%20pasos-1f6feb?style=for-the-badge&logoColor=white" alt="Primeros pasos"></a>- - <a href="https://yakhyo.github.io/uniface/models/"><img src="https://img.shields.io/badge/Cat%C3%A1logo%20de%20modelos-30363d?style=for-the-badge&logoColor=white" alt="Catálogo de modelos"></a>- - <a href="https://yakhyo.github.io/uniface/notebooks/"><img src="https://img.shields.io/badge/Cuadernos-30363d?style=for-the-badge&logo=jupyter&logoColor=white" alt="Cuadernos"></a>- - <a href="https://yakhyo.github.io/uniface/"><img src="https://img.shields.io/badge/Documentaci%C3%B3n%20completa-30363d?style=for-the-badge&logoColor=white" alt="Documentación completa"></a>-</p>--```bash-pip install "uniface[cpu]" # CPU y Apple Silicon-pip install "uniface[gpu]" # NVIDIA CUDA-pip install --pre "uniface[cpu]" # última versión preliminar-```--<details>-<summary><b>Un primer script</b></summary>--<br>--`FaceAnalyzer` ejecuta la detección, la alineación y el reconocimiento en una sola llamada. Los modelos de atributos se añaden de forma explícita.--```python-import cv2-from uniface import FaceAnalyzer, FairFace--analyzer = FaceAnalyzer(predictors=[FairFace()])--for face in analyzer.analyze(cv2.imread("photo.jpg")):- print(face.bbox, face.sex, face.age_group, face.embedding.shape)-```--`bbox`, `confidence`, `landmarks` y `embedding` siempre tienen un valor. La edad, el sexo, la raza, la emoción, la calidad-y los estados del rostro permanecen en `None` hasta que se añade el predictor correspondiente.--</details>--<details>-<summary><b>Las quince tareas y sus modelos</b></summary>--<br>--| Tarea | Modelos |-| --- | --- |-| Detección facial | RetinaFace, SCRFD, CenterFace, YOLOv5-Face, YOLOv8-Face, BlazeFace |-| Reconocimiento facial | AdaFace, ArcFace, EdgeFace, MobileFace, SphereFace |-| Seguimiento facial | BYTETracker, identificadores persistentes entre fotogramas |-| Puntos de referencia faciales | 2d106det (106), PIPNet (98 / 68), Face Mesh (468 / 478, 3D) |-| Segmentación de partes del rostro | BiSeNet (19 clases), máscaras con XSeg |-| Recorte de retratos | MODNet, sin trimap |-| Estimación de la mirada | MobileGaze (ResNet-18 / 34 / 50, MobileNetV2) |-| Pose de la cabeza | representación de rotación 6D, cabeceo / guiñada / alabeo |-| Atributos demográficos | AgeGender, FairFace (grupo de edad, sexo, raza) |-| Emoción | AffectNet-7 y AffectNet-8 |-| Estados del rostro | FaceAttribNet: estado de los ojos, gafas, gafas de sol, mascarilla |-| Calidad facial | eDifFIQA (T / S / M / L) |-| Detección de suplantación facial | detección de vida con MiniFASNet |-| Anonimización | 5 métodos de desenfoque |-| Almacén de vectores | búsqueda de embeddings con FAISS |--Funciona en CPU, Apple Silicon y CUDA. Los pesos se descargan al usarlos por primera vez y se verifican con SHA-256.--</details>--<br>--### Detectar rostros y medir sus características--**Detección facial** · [documentación](https://yakhyo.github.io/uniface/modules/detection/)--<img src="https://raw.githubusercontent.com/yakhyo/uniface/main/assets/demo/detection.jpg" width="100%">--**Puntos de referencia faciales** · [documentación](https://yakhyo.github.io/uniface/modules/landmarks/)--<img src="https://raw.githubusercontent.com/yakhyo/uniface/main/assets/demo/landmarks.jpg" width="100%">--**Malla facial** · [documentación](https://yakhyo.github.io/uniface/modules/landmarks/#face-mesh-468-or-478-points-3d)--<img src="https://raw.githubusercontent.com/yakhyo/uniface/main/assets/demo/face_mesh.jpg" width="100%">--**Calidad facial** · [documentación](https://yakhyo.github.io/uniface/modules/quality/)--<img src="https://raw.githubusercontent.com/yakhyo/uniface/main/assets/demo/quality.jpg" width="100%">--### Separar rostros y retratos--**Segmentación de partes del rostro** · [documentación](https://yakhyo.github.io/uniface/modules/parsing/)--<img src="https://raw.githubusercontent.com/yakhyo/uniface/main/assets/demo/parsing.jpg" width="100%">--**Segmentación facial** · [documentación](https://yakhyo.github.io/uniface/modules/parsing/#xseg)--<img src="https://raw.githubusercontent.com/yakhyo/uniface/main/assets/demo/segmentation.jpg" width="100%">--<img src="https://raw.githubusercontent.com/yakhyo/uniface/main/assets/demo/segmentation_occluded.jpg" width="100%">--**Recorte de retratos** · [documentación](https://yakhyo.github.io/uniface/modules/matting/)--<img src="https://raw.githubusercontent.com/yakhyo/uniface/main/assets/demo/matting.jpg" width="100%">--<img src="https://raw.githubusercontent.com/yakhyo/uniface/main/assets/demo/matting_alt.jpg" width="100%">--### Estimar la orientación de la cabeza y la mirada--**Pose de la cabeza** · [documentación](https://yakhyo.github.io/uniface/modules/headpose/)--<img src="https://raw.githubusercontent.com/yakhyo/uniface/main/assets/demo/headpose.jpg" width="100%">--**Estimación de la mirada** · [documentación](https://yakhyo.github.io/uniface/modules/gaze/)--<img src="https://raw.githubusercontent.com/yakhyo/uniface/main/assets/demo/gaze.jpg" width="100%">--### Analizar atributos faciales--**Edad y sexo** · [documentación](https://yakhyo.github.io/uniface/modules/attributes/)--<img src="https://raw.githubusercontent.com/yakhyo/uniface/main/assets/demo/demography.jpg" width="100%">--**Emoción** · [documentación](https://yakhyo.github.io/uniface/modules/attributes/#emotion)--<img src="https://raw.githubusercontent.com/yakhyo/uniface/main/assets/demo/emotion.jpg" width="100%">--**Estados del rostro** · [documentación](https://yakhyo.github.io/uniface/modules/attributes/#faceattribnet)--<img src="https://raw.githubusercontent.com/yakhyo/uniface/main/assets/demo/face_states.jpg" width="100%">--<img src="https://raw.githubusercontent.com/yakhyo/uniface/main/assets/demo/face_states_alt.jpg" width="100%">--### Distinguir un rostro real de una reproducción--**Detección de suplantación facial** · [documentación](https://yakhyo.github.io/uniface/modules/spoofing/)--<img src="https://raw.githubusercontent.com/yakhyo/uniface/main/assets/demo/spoofing.jpg" width="100%">--### Comparar rostros o anonimizarlos--**Reconocimiento facial** · [documentación](https://yakhyo.github.io/uniface/modules/recognition/)--<img src="https://raw.githubusercontent.com/yakhyo/uniface/main/assets/demo/verification.jpg" width="100%">--**Anonimización facial** · [documentación](https://yakhyo.github.io/uniface/modules/privacy/)--<img src="https://raw.githubusercontent.com/yakhyo/uniface/main/assets/demo/anonymization.jpg" width="100%">--<br>--<div align="center">--**[Primeros pasos](https://yakhyo.github.io/uniface/quickstart/)** · -[Catálogo de modelos](https://yakhyo.github.io/uniface/models/) · -[Cuadernos](https://yakhyo.github.io/uniface/notebooks/) · -[Licencias de los modelos](https://yakhyo.github.io/uniface/license-attribution/) · -[Contribuir](CONTRIBUTING.md) · -[Discord](https://discord.gg/wdzrjr7R5j) · -[Incidencias](https://github.com/yakhyo/uniface/issues)--Funciona en CPU, Apple Silicon y CUDA. Los pesos se descargan al usarlos por primera vez y se verifican con SHA-256.<br>-UniFace se distribuye bajo la licencia [MIT](LICENSE); algunos pesos preentrenados tienen otras licencias, así que consulta las-[licencias](https://yakhyo.github.io/uniface/license-attribution/) antes de distribuirlo comercialmente.<br>-Este proyecto no está afiliado a [Uniface](https://uniface.com/) de Rocket Software.--</div>diff --git a/README.ko.md b/README.ko.mdindex d0e0411..f30bfec 100644--- a/README.ko.md+++ b/README.ko.md@@ -1,4 +1,4 @@-\[[English](README.md)\] | \[[Chinese](README.zh-CN.md)\] | \[[Korean](README.ko.md)\] | \[[Spanish](README.es.md)\] | \[[Uzbek](README.uz.md)\]+\[[English](README.md)\] | \[[Chinese](README.zh-CN.md)\] | \[[Korean](README.ko.md)\] | \[[Uzbek](README.uz.md)\]<h1 align="center">UniFace: Python용 통합 얼굴 분석 라이브러리</h1>diff --git a/README.md b/README.mdindex 435902d..41a52d6 100644--- a/README.md+++ b/README.md@@ -1,4 +1,4 @@-\[[English](README.md)\] | \[[Chinese](README.zh-CN.md)\] | \[[Korean](README.ko.md)\] | \[[Spanish](README.es.md)\] | \[[Uzbek](README.uz.md)\]+\[[English](README.md)\] | \[[Chinese](README.zh-CN.md)\] | \[[Korean](README.ko.md)\] | \[[Uzbek](README.uz.md)\]<h1 align="center">UniFace: A Unified Face Analysis Library for Python</h1>diff --git a/README.uz.md b/README.uz.mdindex a4479b8..773850c 100644--- a/README.uz.md+++ b/README.uz.md@@ -1,4 +1,4 @@-\[[English](README.md)\] | \[[Chinese](README.zh-CN.md)\] | \[[Korean](README.ko.md)\] | \[[Spanish](README.es.md)\] | \[[Uzbek](README.uz.md)\]+\[[English](README.md)\] | \[[Chinese](README.zh-CN.md)\] | \[[Korean](README.ko.md)\] | \[[Uzbek](README.uz.md)\]<h1 align="center">UniFace: Python uchun yagona yuz tahlili kutubxonasi</h1>diff --git a/README.zh-CN.md b/README.zh-CN.mdindex 0392aff..161af29 100644--- a/README.zh-CN.md+++ b/README.zh-CN.md@@ -1,4 +1,4 @@-\[[English](README.md)\] | \[[Chinese](README.zh-CN.md)\] | \[[Korean](README.ko.md)\] | \[[Spanish](README.es.md)\] | \[[Uzbek](README.uz.md)\]+\[[English](README.md)\] | \[[Chinese](README.zh-CN.md)\] | \[[Korean](README.ko.md)\] | \[[Uzbek](README.uz.md)\]<h1 align="center">UniFace:统一的 Python 人脸分析库</h1>diff --git a/docs/quickstart.md b/docs/quickstart.mdindex 4684a78..3f2c173 100644--- a/docs/quickstart.md+++ b/docs/quickstart.md@@ -13,7 +13,7 @@ models you hand it. Start here if you want a whole pipeline rather than one modeimport cv2from uniface import FaceAnalyzer-analyzer = FaceAnalyzer() # SCRFD + ArcFace by default, no arguments needed+analyzer = FaceAnalyzer() # RetinaFace + EdgeFace by default, no arguments neededfaces = analyzer.analyze(cv2.imread("photo.jpg"))for face in faces:diff --git a/tests/test_analyzer.py b/tests/test_analyzer.pyindex fd54e53..f1c8af1 100644--- a/tests/test_analyzer.py+++ b/tests/test_analyzer.py@@ -121,3 +121,14 @@ def test_repr_lists_predictors():def test_constructor_is_keyword_only():with pytest.raises(TypeError):FaceAnalyzer(StubDetector())+++def test_defaults_use_permissively_licensed_weights():+ """InsightFace weights (SCRFD, ArcFace) are non-commercial, so they must stay opt-in."""+ from uniface.detection import RetinaFace+ from uniface.recognition import EdgeFace++ analyzer = FaceAnalyzer()++ assert isinstance(analyzer.detector, RetinaFace)+ assert isinstance(analyzer.recognizer, EdgeFace)diff --git a/tests/test_detector_base.py b/tests/test_detector_base.pyindex 8421d74..ae537fc 100644--- a/tests/test_detector_base.py+++ b/tests/test_detector_base.py@@ -42,7 +42,7 @@ def test_capability_flags_are_opt_in():def test_analyzer_disables_recognition_for_box_only_detector():- """A boxes-only detector must take the recognition-disabled branch, not load ArcFace."""+ """A boxes-only detector must take the recognition-disabled branch, not load EdgeFace."""analyzer = FaceAnalyzer(detector=BoxOnlyDetector())assert analyzer.recognizer is Nonediff --git a/uniface/analyzer.py b/uniface/analyzer.pyindex c12c7ad..60122f0 100644--- a/uniface/analyzer.py+++ b/uniface/analyzer.py@@ -9,8 +9,11 @@ from typing import Anyimport numpy as npfrom uniface.attribute.base import BaseAttribute+from uniface.constants import EdgeFaceWeights, RetinaFaceWeights+from uniface.detection import RetinaFacefrom uniface.detection.base import BaseDetectorfrom uniface.log import Logger+from uniface.recognition import EdgeFacefrom uniface.recognition.base import BaseRecognizerfrom uniface.types import Face@@ -31,13 +34,13 @@ class FaceAnalyzer:via the `predictors` list. Each predictor's `predict(image, face)`is called once per detected face, enriching the `Face` in-place.- When called with no arguments, uses SCRFD (500M) for detection and- ArcFace (MobileNet) for recognition — the smallest and fastest variants.+ When called with no arguments, uses RetinaFace (MobileNetV2) for detection+ and EdgeFace (XXS) for recognition.Args:- detector: Face detector instance. Defaults to `SCRFD(SCRFD_500M_KPS)`.+ detector: Face detector instance. Defaults to `RetinaFace(MNET_V2)`.recognizer: Face recognizer for extracting embeddings.- Defaults to `ArcFace(MNET)`. Pass `None` to disable recognition.+ Defaults to `EdgeFace(XXS)`. Pass `None` to disable recognition.predictors: Optional list of `BaseAttribute` predictors to run oneach detected face (e.g. `[AgeGender()]`).@@ -59,10 +62,7 @@ class FaceAnalyzer:predictors: list[BaseAttribute] | None = None,) -> None:if detector is None:- from uniface.constants import SCRFDWeights- from uniface.detection import SCRFD-- detector = SCRFD(model_name=SCRFDWeights.SCRFD_500M_KPS)+ detector = RetinaFace(model_name=RetinaFaceWeights.MNET_V2)# Checked before the _UNSET branch below so an unusable recognizer is never loaded.if not getattr(detector, 'supports_alignment', False):@@ -75,9 +75,7 @@ class FaceAnalyzer:recognizer = Noneif recognizer is _UNSET:- from uniface.recognition import ArcFace-- recognizer = ArcFace()+ recognizer = EdgeFace(model_name=EdgeFaceWeights.XXS)self.detector = detectorself.recognizer = recognizer