Feat/batch predict age and gender

deepface/models/Demography.py

@@ -1,4 +1,4 @@
-from typing import Union
+from typing import Union, List
 from abc import ABC, abstractmethod
 import numpy as np
 from deepface.commons import package_utils
@@ -18,5 +18,54 @@ class Demography(ABC):
     model_name: str
 
     @abstractmethod
-    def predict(self, img: np.ndarray) -> Union[np.ndarray, np.float64]:
+    def predict(self, img: Union[np.ndarray, List[np.ndarray]]) -> Union[np.ndarray, np.float64]:
         pass
+
+    def _predict_internal(self, img_batch: np.ndarray) -> np.ndarray:
+        """
+        Predict for single image or batched images.
+        This method uses legacy method while receiving single image as input. 
+        And switch to batch prediction if receives batched images.
+
+        Args:
+            img_batch:
+                Batch of images as np.ndarray (n, x, y, c)
+                    with n >= 1, x = image width, y = image height, c = channel
+                Or Single image as np.ndarray (1, x, y, c)
+                    with x = image width, y = image height and c = channel
+                The channel dimension will be 1 if input is grayscale. (For emotion model)
+        """
+        if not self.model_name: # Check if called from derived class
+            raise NotImplementedError("no model selected")
+        assert img_batch.ndim == 4, "expected 4-dimensional tensor input"
+
+        if img_batch.shape[0] == 1: # Single image
+            # Predict with legacy method.
+            return self.model(img_batch, training=False).numpy()[0, :]
+
+        # Batch of images
+        # Predict with batch prediction
+        return self.model.predict_on_batch(img_batch)
+
+    def _preprocess_batch_or_single_input(
+        self,
+        img: Union[np.ndarray, List[np.ndarray]]
+    ) -> np.ndarray:
+
+        """
+        Preprocess single or batch of images, return as 4-D numpy array.
+        Args:
+            img: Single image as np.ndarray (224, 224, 3) or
+                 List of images as List[np.ndarray] or
+                 Batch of images as np.ndarray (n, 224, 224, 3)
+        Returns:
+            Four-dimensional numpy array (n, 224, 224, 3)
+        """
+        image_batch = np.array(img)
+        # Remove batch dimension in advance if exists
+        image_batch = image_batch.squeeze()
+        # Check input dimension
+        if len(image_batch.shape) == 3:
+            # Single image - add batch dimension
+            image_batch = np.expand_dims(image_batch, axis=0)
+        return image_batch

deepface/models/demography/Age.py

@@ -1,3 +1,7 @@
+# stdlib dependencies
+
+from typing import List, Union
+
 # 3rd party dependencies
 import numpy as np
 
@@ -37,11 +41,29 @@ def __init__(self):
         self.model = load_model()
         self.model_name = "Age"
 
-    def predict(self, img: np.ndarray) -> np.float64:
-        # model.predict causes memory issue when it is called in a for loop
-        # age_predictions = self.model.predict(img, verbose=0)[0, :]
-        age_predictions = self.model(img, training=False).numpy()[0, :]
-        return find_apparent_age(age_predictions)
+    def predict(self, img: Union[np.ndarray, List[np.ndarray]]) -> Union[np.float64, np.ndarray]:
+        """
+        Predict apparent age(s) for single or multiple faces
+        Args:
+            img: Single image as np.ndarray (224, 224, 3) or
+                List of images as List[np.ndarray] or
+                Batch of images as np.ndarray (n, 224, 224, 3)
+        Returns:
+            np.ndarray (age_classes,) if single image, 
+            np.ndarray (n, age_classes) if batched images.
+        """
+        # Preprocessing input image or image list.
+        imgs = self._preprocess_batch_or_single_input(img)
+
+        # Prediction from 3 channels image
+        age_predictions = self._predict_internal(imgs)
+
+        # Calculate apparent ages
+        if len(age_predictions.shape) == 1: # Single prediction list
+            return find_apparent_age(age_predictions)
+
+        return np.array([
+            find_apparent_age(age_prediction) for age_prediction in age_predictions])
 
 
 def load_model(
@@ -65,7 +87,7 @@ def load_model(
 
     # --------------------------
 
-    age_model = Model(inputs=model.input, outputs=base_model_output)
+    age_model = Model(inputs=model.inputs, outputs=base_model_output)
 
     # --------------------------
 
@@ -78,15 +100,16 @@ def load_model(
 
     return age_model
 
-
 def find_apparent_age(age_predictions: np.ndarray) -> np.float64:
     """
     Find apparent age prediction from a given probas of ages
     Args:
-        age_predictions (?)
+        age_predictions (age_classes,)
     Returns:
         apparent_age (float)
     """
+    assert len(age_predictions.shape) == 1, f"Input should be a list of predictions, \
+                                             not batched. Got shape: {age_predictions.shape}"
     output_indexes = np.arange(0, 101)
     apparent_age = np.sum(age_predictions * output_indexes)
     return apparent_age

deepface/models/demography/Emotion.py

@@ -1,3 +1,6 @@
+# stdlib dependencies
+from typing import List, Union
+
 # 3rd party dependencies
 import numpy as np
 import cv2
@@ -43,16 +46,38 @@ def __init__(self):
         self.model = load_model()
         self.model_name = "Emotion"
 
-    def predict(self, img: np.ndarray) -> np.ndarray:
-        img_gray = cv2.cvtColor(img[0], cv2.COLOR_BGR2GRAY)
+    def _preprocess_image(self, img: np.ndarray) -> np.ndarray:
+        """
+        Preprocess single image for emotion detection
+        Args:
+            img: Input image (224, 224, 3)
+        Returns:
+            Preprocessed grayscale image (48, 48)
+        """
+        img_gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
         img_gray = cv2.resize(img_gray, (48, 48))
-        img_gray = np.expand_dims(img_gray, axis=0)
-
-        # model.predict causes memory issue when it is called in a for loop
-        # emotion_predictions = self.model.predict(img_gray, verbose=0)[0, :]
-        emotion_predictions = self.model(img_gray, training=False).numpy()[0, :]
-
-        return emotion_predictions
+        return img_gray
+
+    def predict(self, img: Union[np.ndarray, List[np.ndarray]]) -> np.ndarray:
+        """
+        Predict emotion probabilities for single or multiple faces
+        Args:
+            img: Single image as np.ndarray (224, 224, 3) or
+                List of images as List[np.ndarray] or
+                Batch of images as np.ndarray (n, 224, 224, 3)
+        Returns:
+            np.ndarray (n, n_emotions)
+            where n_emotions is the number of emotion categories
+        """
+        # Preprocessing input image or image list.
+        imgs = self._preprocess_batch_or_single_input(img)
+
+        processed_imgs = np.expand_dims(np.array([self._preprocess_image(img) for img in imgs]), axis=-1)
+
+        # Prediction
+        predictions = self._predict_internal(processed_imgs)
+
+        return predictions
 
 
 def load_model(

deepface/models/demography/Gender.py

@@ -1,3 +1,7 @@
+# stdlib dependencies
+
+from typing import List, Union
+
 # 3rd party dependencies
 import numpy as np
 
@@ -37,11 +41,23 @@ def __init__(self):
         self.model = load_model()
         self.model_name = "Gender"
 
-    def predict(self, img: np.ndarray) -> np.ndarray:
-        # model.predict causes memory issue when it is called in a for loop
-        # return self.model.predict(img, verbose=0)[0, :]
-        return self.model(img, training=False).numpy()[0, :]
+    def predict(self, img: Union[np.ndarray, List[np.ndarray]]) -> np.ndarray:
+        """
+        Predict gender probabilities for single or multiple faces
+        Args:
+            img: Single image as np.ndarray (224, 224, 3) or
+                List of images as List[np.ndarray] or
+                Batch of images as np.ndarray (n, 224, 224, 3)
+        Returns:
+            np.ndarray (n, 2)
+        """
+        # Preprocessing input image or image list.
+        imgs = self._preprocess_batch_or_single_input(img)
+
+        # Prediction
+        predictions = self._predict_internal(imgs)
 
+        return predictions
 
 def load_model(
     url=WEIGHTS_URL,
@@ -64,7 +80,7 @@ def load_model(
 
     # --------------------------
 
-    gender_model = Model(inputs=model.input, outputs=base_model_output)
+    gender_model = Model(inputs=model.inputs, outputs=base_model_output)
 
     # --------------------------
 

deepface/models/demography/Race.py

@@ -1,3 +1,6 @@
+# stdlib dependencies
+from typing import List, Union
+
 # 3rd party dependencies
 import numpy as np
 
@@ -37,10 +40,24 @@ def __init__(self):
         self.model = load_model()
         self.model_name = "Race"
 
-    def predict(self, img: np.ndarray) -> np.ndarray:
-        # model.predict causes memory issue when it is called in a for loop
-        # return self.model.predict(img, verbose=0)[0, :]
-        return self.model(img, training=False).numpy()[0, :]
+    def predict(self, img: Union[np.ndarray, List[np.ndarray]]) -> np.ndarray:
+        """
+        Predict race probabilities for single or multiple faces
+        Args:
+            img: Single image as np.ndarray (224, 224, 3) or
+                List of images as List[np.ndarray] or
+                Batch of images as np.ndarray (n, 224, 224, 3)
+        Returns:
+            np.ndarray (n, n_races)
+            where n_races is the number of race categories
+        """
+        # Preprocessing input image or image list.
+        imgs = self._preprocess_batch_or_single_input(img)
+
+        # Prediction
+        predictions = self._predict_internal(imgs)
+
+        return predictions
 
 
 def load_model(
@@ -62,7 +79,7 @@ def load_model(
 
     # --------------------------
 
-    race_model = Model(inputs=model.input, outputs=base_model_output)
+    race_model = Model(inputs=model.inputs, outputs=base_model_output)
 
     # --------------------------
 

deepface/modules/demography.py

@@ -175,6 +175,7 @@ def analyze(
                     task="facial_attribute", model_name="Age"
                 ).predict(img_content)
                 # int cast is for exception - object of type 'float32' is not JSON serializable
+                print(apparent_age.shape)
                 obj["age"] = int(apparent_age)
 
             elif action == "gender":

tests/test_analyze.py

@@ -135,3 +135,16 @@ def test_analyze_for_different_detectors():
                     assert result["gender"]["Man"] > result["gender"]["Woman"]
                 else:
                     assert result["gender"]["Man"] < result["gender"]["Woman"]
+
+def test_analyze_for_multiple_faces():
+    img = "dataset/img4.jpg"
+    # Copy and combine the same image to create multiple faces
+    img = cv2.imread(img)
+    img = cv2.hconcat([img, img])
+    demography_objs = DeepFace.analyze(img, silent=True)
+    assert len(demography_objs) == 2
+    for demography in demography_objs:
+        logger.debug(demography)
+        assert demography["age"] > 20 and demography["age"] < 40
+        assert demography["dominant_gender"] == "Woman"
+    logger.info("✅ test analyze for multiple faces done")