yolov11n and yolov11m added to model selection

benchmarks/Evaluate-Results.ipynb

@@ -29,8 +29,8 @@
    "outputs": [],
    "source": [
     "alignment = [False, True]\n",
-    "models = [\"Facenet512\", \"Facenet\", \"VGG-Face\", \"ArcFace\", \"Dlib\", \"GhostFaceNet\", \"SFace\", \"OpenFace\", \"DeepFace\", \"DeepID\"]\n",
-    "detectors = [\"retinaface\", \"mtcnn\", \"fastmtcnn\", \"dlib\", \"yolov8\", \"yolov11n\", \"yolov11m\", \"yunet\", \"centerface\", \"mediapipe\", \"ssd\", \"opencv\", \"skip\"]\n",
+    "models = [\"Facenet512\", \"Facenet\", \"VGG-Face\", \"ArcFace\", \"Dlib\", \"GhostFaceNet\", \"SFace\", \"OpenFace\", \"DeepFace\", \"DeepID\", \"yolov8\", \"yolov11n\", \"yolov11s\", \"yolov11m\"]\n",
+    "detectors = [\"retinaface\", \"mtcnn\", \"fastmtcnn\", \"dlib\", \"yolov8\", \"yolov11n\", \"yolov11s\", \"yolov11m\", \"yunet\", \"centerface\", \"mediapipe\", \"ssd\", \"opencv\", \"skip\"]\n",
     "distance_metrics = [\"euclidean\", \"euclidean_l2\", \"cosine\"]"
    ]
   },

deepface/DeepFace.py

@@ -56,7 +56,7 @@ def build_model(model_name: str, task: str = "facial_recognition") -> Any:
             - VGG-Face, Facenet, Facenet512, OpenFace, DeepFace, DeepID, Dlib,
                 ArcFace, SFace, GhostFaceNet for face recognition
             - Age, Gender, Emotion, Race for facial attributes
-            - opencv, mtcnn, ssd, dlib, retinaface, mediapipe, yolov8, 'yolov11n', 'yolov11m', yunet,
+            - opencv, mtcnn, ssd, dlib, retinaface, mediapipe, yolov8, 'yolov11n', 'yolov11s','yolov11m', yunet,
                 fastmtcnn or centerface for face detectors
             - Fasnet for spoofing
         task (str): facial_recognition, facial_attribute, face_detector, spoofing
@@ -96,7 +96,7 @@ def verify(
             OpenFace, DeepFace, DeepID, Dlib, ArcFace, SFace and GhostFaceNet (default is VGG-Face).
 
         detector_backend (string): face detector backend. Options: 'opencv', 'retinaface',
-            'mtcnn', 'ssd', 'dlib', 'mediapipe', 'yolov8', 'yolov11n', 'yolov11m', 'centerface' or 'skip'
+            'mtcnn', 'ssd', 'dlib', 'mediapipe', 'yolov8', 'yolov11n', 'yolov11s', 'yolov11m', 'centerface' or 'skip'
             (default is opencv).
 
         distance_metric (string): Metric for measuring similarity. Options: 'cosine',
@@ -187,7 +187,7 @@ def analyze(
             Set to False to avoid the exception for low-resolution images (default is True).
 
         detector_backend (string): face detector backend. Options: 'opencv', 'retinaface',
-            'mtcnn', 'ssd', 'dlib', 'mediapipe', 'yolov8', 'yolov11n', 'yolov11m', 'centerface' or 'skip'
+            'mtcnn', 'ssd', 'dlib', 'mediapipe', 'yolov8', 'yolov11n',  'yolov11s', 'yolov11m', 'centerface' or 'skip'
             (default is opencv).
 
         distance_metric (string): Metric for measuring similarity. Options: 'cosine',
@@ -298,7 +298,7 @@ def find(
             Set to False to avoid the exception for low-resolution images (default is True).
 
         detector_backend (string): face detector backend. Options: 'opencv', 'retinaface',
-            'mtcnn', 'ssd', 'dlib', 'mediapipe', 'yolov8', 'yolov11n', 'yolov11m', 'centerface' or 'skip'
+            'mtcnn', 'ssd', 'dlib', 'mediapipe', 'yolov8', 'yolov11n', 'yolov11s', 'yolov11m', 'centerface' or 'skip'
             (default is opencv).
 
         align (boolean): Perform alignment based on the eye positions (default is True).
@@ -396,7 +396,7 @@ def represent(
             (default is True).
 
         detector_backend (string): face detector backend. Options: 'opencv', 'retinaface',
-            'mtcnn', 'ssd', 'dlib', 'mediapipe', 'yolov8', 'yolov11n', 'yolov11m', 'centerface' or 'skip'
+            'mtcnn', 'ssd', 'dlib', 'mediapipe', 'yolov8', 'yolov11n', 'yolov11s', 'yolov11m', 'centerface' or 'skip'
             (default is opencv).
 
         align (boolean): Perform alignment based on the eye positions (default is True).
@@ -462,7 +462,7 @@ def stream(
             OpenFace, DeepFace, DeepID, Dlib, ArcFace, SFace and GhostFaceNet (default is VGG-Face).
 
         detector_backend (string): face detector backend. Options: 'opencv', 'retinaface',
-            'mtcnn', 'ssd', 'dlib', 'mediapipe', 'yolov8', 'yolov11n', 'yolov11m', 'centerface' or 'skip'
+            'mtcnn', 'ssd', 'dlib', 'mediapipe', 'yolov8', 'yolov11n', 'yolov11s', 'yolov11m', 'centerface' or 'skip'
             (default is opencv).
 
         distance_metric (string): Metric for measuring similarity. Options: 'cosine',
@@ -517,7 +517,7 @@ def extract_faces(
             as a string, numpy array (BGR), or base64 encoded images.
 
         detector_backend (string): face detector backend. Options: 'opencv', 'retinaface',
-            'mtcnn', 'ssd', 'dlib', 'mediapipe', 'yolov8', 'yolov11n', 'yolov11m', 'centerface' or 'skip'
+            'mtcnn', 'ssd', 'dlib', 'mediapipe', 'yolov8', 'yolov11n', 'yolov11s', 'yolov11m', 'centerface' or 'skip'
             (default is opencv).
 
         enforce_detection (boolean): If no face is detected in an image, raise an exception.
@@ -601,7 +601,7 @@ def detectFace(
             added to resize the image (default is (224, 224)).
 
         detector_backend (string): face detector backend. Options: 'opencv', 'retinaface',
-            'mtcnn', 'ssd', 'dlib', 'mediapipe', 'yolov8', 'yolov11n', 'yolov11m', 'centerface' or 'skip'
+            'mtcnn', 'ssd', 'dlib', 'mediapipe', 'yolov8', 'yolov11n', 'yolov11s', 'yolov11m', 'centerface' or 'skip'
             (default is opencv).
 
         enforce_detection (boolean): If no face is detected in an image, raise an exception.

deepface/commons/weight_utils.py

@@ -127,7 +127,7 @@ def download_all_models_in_one_shot() -> None:
         MODEL_URL as SSD_MODEL,
         WEIGHTS_URL as SSD_WEIGHTS,
     )
-    from deepface.models.face_detection.Yolo import (
+    from deepface.models.YoloModel import (
         WEIGHT_URLS as YOLO_WEIGHTS,
         WEIGHT_NAMES as YOLO_WEIGHT_NAMES,
         YoloModel
@@ -170,6 +170,10 @@ def download_all_models_in_one_shot() -> None:
             "filename": YOLO_WEIGHT_NAMES[YoloModel.V11N.value],
             "url": YOLO_WEIGHTS[YoloModel.V11N.value],
         },
+        {
+            "filename": YOLO_WEIGHT_NAMES[YoloModel.V11S.value],
+            "url": YOLO_WEIGHTS[YoloModel.V11S.value],
+        },
         {
             "filename": YOLO_WEIGHT_NAMES[YoloModel.V11M.value],
             "url": YOLO_WEIGHTS[YoloModel.V11M.value],

deepface/models/YoloClientBase.py

@@ -0,0 +1,37 @@
+# built-in dependencies
+from typing import Any
+
+# project dependencies
+from deepface.models.YoloModel import YoloModel, WEIGHT_URLS, WEIGHT_NAMES
+from deepface.commons import weight_utils
+from deepface.commons.logger import Logger
+
+logger = Logger()
+
+
+class YoloClientBase:
+    def __init__(self, model: YoloModel):
+        self.model = self.build_model(model)
+
+    def build_model(self, model: YoloModel) -> Any:
+        """
+        Build a yolo detector model
+        Returns:
+            model (Any)
+        """
+
+        # Import the optional Ultralytics YOLO model
+        try:
+            from ultralytics import YOLO
+        except ModuleNotFoundError as e:
+            raise ImportError(
+                "Yolo is an optional detector, ensure the library is installed. "
+                "Please install using 'pip install ultralytics'"
+            ) from e
+
+        weight_file = weight_utils.download_weights_if_necessary(
+            file_name=WEIGHT_NAMES[model.value], source_url=WEIGHT_URLS[model.value]
+        )
+
+        # Return face_detector
+        return YOLO(weight_file)

deepface/models/YoloModel.py

@@ -0,0 +1,21 @@
+from enum import Enum
+
+
+class YoloModel(Enum):
+    V8N = 0
+    V11N = 1
+    V11S = 2
+    V11M = 3
+
+
+# Model's weights paths
+WEIGHT_NAMES = ["yolov8n-face.pt",
+                "yolov11n-face.pt",
+                "yolov11s-face.pt",
+                "yolov11m-face.pt"]
+
+# Google Drive URL from repo (https://github.com/derronqi/yolov8-face) ~6MB
+WEIGHT_URLS = ["https://drive.google.com/uc?id=1qcr9DbgsX3ryrz2uU8w4Xm3cOrRywXqb",
+               "https://github.com/akanametov/yolo-face/releases/download/v0.0.0/yolov11n-face.pt",
+               "https://github.com/akanametov/yolo-face/releases/download/v0.0.0/yolov11s-face.pt",
+               "https://github.com/akanametov/yolo-face/releases/download/v0.0.0/yolov11m-face.pt"]

deepface/models/face_detection/Yolo.py

@@ -1,61 +1,22 @@
 # built-in dependencies
 import os
-from typing import Any, List
-from enum import Enum
+from typing import List
 
 # 3rd party dependencies
 import numpy as np
 
 # project dependencies
+from deepface.models.YoloClientBase import YoloClientBase
+from deepface.models.YoloModel import YoloModel
 from deepface.models.Detector import Detector, FacialAreaRegion
-from deepface.commons import weight_utils
 from deepface.commons.logger import Logger
 
 logger = Logger()
 
-# Model's weights paths
-WEIGHT_NAMES = ["yolov8n-face.pt",
-                "yolov11n-face.pt",
-                "yolov11m-face.pt"]
 
-# Google Drive URL from repo (https://github.com/derronqi/yolov8-face) ~6MB
-WEIGHT_URLS = ["https://drive.google.com/uc?id=1qcr9DbgsX3ryrz2uU8w4Xm3cOrRywXqb",
-               "https://github.com/akanametov/yolo-face/releases/download/v0.0.0/yolov11n-face.pt",
-               "https://github.com/akanametov/yolo-face/releases/download/v0.0.0/yolov11m-face.pt"]
-
-
-class YoloModel(Enum):
-    V8N = 0
-    V11N = 1
-    V11M = 2
-
-
-class YoloClient(Detector):
+class YoloDetectorClient(YoloClientBase, Detector):
     def __init__(self, model: YoloModel):
-        self.model = self.build_model(model)
-
-    def build_model(self, model: YoloModel) -> Any:
-        """
-        Build a yolo detector model
-        Returns:
-            model (Any)
-        """
-
-        # Import the optional Ultralytics YOLO model
-        try:
-            from ultralytics import YOLO
-        except ModuleNotFoundError as e:
-            raise ImportError(
-                "Yolo is an optional detector, ensure the library is installed. "
-                "Please install using 'pip install ultralytics'"
-            ) from e
-
-        weight_file = weight_utils.download_weights_if_necessary(
-            file_name=WEIGHT_NAMES[model.value], source_url=WEIGHT_URLS[model.value]
-        )
-
-        # Return face_detector
-        return YOLO(weight_file)
+        super().__init__(model)
 
     def detect_faces(self, img: np.ndarray) -> List[FacialAreaRegion]:
         """
@@ -80,21 +41,24 @@ def detect_faces(self, img: np.ndarray) -> List[FacialAreaRegion]:
         # For each face, extract the bounding box, the landmarks and confidence
         for result in results:
 
-            if result.boxes is None or result.keypoints is None:
+            if result.boxes is None:
                 continue
 
             # Extract the bounding box and the confidence
             x, y, w, h = result.boxes.xywh.tolist()[0]
             confidence = result.boxes.conf.tolist()[0]
 
-            # right_eye_conf = result.keypoints.conf[0][0]
-            # left_eye_conf = result.keypoints.conf[0][1]
-            right_eye = result.keypoints.xy[0][0].tolist()
-            left_eye = result.keypoints.xy[0][1].tolist()
+            right_eye = None
+            left_eye = None
+            if result.keypoints is not None:
+                # right_eye_conf = result.keypoints.conf[0][0]
+                # left_eye_conf = result.keypoints.conf[0][1]
+                right_eye = result.keypoints.xy[0][0].tolist()
+                left_eye = result.keypoints.xy[0][1].tolist()
 
-            # eyes are list of float, need to cast them tuple of int
-            left_eye = tuple(int(i) for i in left_eye)
-            right_eye = tuple(int(i) for i in right_eye)
+                # eyes are list of float, need to cast them tuple of int
+                left_eye = tuple(int(i) for i in left_eye)
+                right_eye = tuple(int(i) for i in right_eye)
 
             x, y, w, h = int(x - w / 2), int(y - h / 2), int(w), int(h)
             facial_area = FacialAreaRegion(
@@ -111,16 +75,21 @@ def detect_faces(self, img: np.ndarray) -> List[FacialAreaRegion]:
         return resp
 
 
-class YoloClientV8n(YoloClient):
+class YoloDetectorClientV8n(YoloDetectorClient):
     def __init__(self):
         super().__init__(YoloModel.V8N)
 
 
-class YoloClientV11n(YoloClient):
+class YoloDetectorClientV11n(YoloDetectorClient):
     def __init__(self):
         super().__init__(YoloModel.V11N)
 
 
-class YoloClientV11m(YoloClient):
+class YoloDetectorClientV11s(YoloDetectorClient):
+    def __init__(self):
+        super().__init__(YoloModel.V11S)
+
+
+class YoloDetectorClientV11m(YoloDetectorClient):
     def __init__(self):
         super().__init__(YoloModel.V11M)

deepface/models/facial_recognition/Yolo.py

@@ -0,0 +1,44 @@
+# built-in dependencies
+from typing import List
+
+# 3rd party dependencies
+import numpy as np
+
+# project dependencies
+from deepface.models.YoloClientBase import YoloClientBase
+from deepface.models.YoloModel import YoloModel
+from deepface.models.FacialRecognition import FacialRecognition
+from deepface.commons.logger import Logger
+
+logger = Logger()
+
+
+class YoloFacialRecognitionClient(YoloClientBase, FacialRecognition):
+    def __init__(self, model: YoloModel):
+        super().__init__(model)
+        self.model_name = "Yolo"
+        self.input_shape = None
+        self.output_shape = 512
+
+    def forward(self, img: np.ndarray) -> List[float]:
+        return self.model.embed(img)[0].tolist()
+
+
+class YoloFacialRecognitionClientV8n(YoloFacialRecognitionClient):
+    def __init__(self):
+        super().__init__(YoloModel.V8N)
+
+
+class YoloFacialRecognitionClientV11n(YoloFacialRecognitionClient):
+    def __init__(self):
+        super().__init__(YoloModel.V11N)
+
+
+class YoloFacialRecognitionClientV11s(YoloFacialRecognitionClient):
+    def __init__(self):
+        super().__init__(YoloModel.V11S)
+
+
+class YoloFacialRecognitionClientV11m(YoloFacialRecognitionClient):
+    def __init__(self):
+        super().__init__(YoloModel.V11M)

deepface/modules/demography.py

@@ -35,7 +35,7 @@ def analyze(
             Set to False to avoid the exception for low-resolution images (default is True).
 
         detector_backend (string): face detector backend. Options: 'opencv', 'retinaface',
-            'mtcnn', 'ssd', 'dlib', 'mediapipe', 'yolov8', 'yolov11n', 'yolov11m', 'centerface' or 'skip'
+            'mtcnn', 'ssd', 'dlib', 'mediapipe', 'yolov8', 'yolov11n', 'yolov11s', 'yolov11m', 'centerface' or 'skip'
             (default is opencv).
 
         distance_metric (string): Metric for measuring similarity. Options: 'cosine',

deepface/modules/detection.py

@@ -38,7 +38,7 @@ def extract_faces(
             as a string, numpy array (BGR), or base64 encoded images.
 
         detector_backend (string): face detector backend. Options: 'opencv', 'retinaface',
-            'mtcnn', 'ssd', 'dlib', 'mediapipe', 'yolov8', 'yolov11n', 'yolov11m', 'centerface' or 'skip'
+            'mtcnn', 'ssd', 'dlib', 'mediapipe', 'yolov8', 'yolov11n', 'yolov11s', 'yolov11m', 'centerface' or 'skip'
             (default is opencv)
 
         enforce_detection (boolean): If no face is detected in an image, raise an exception.