facenet 512

README.md

@@ -51,7 +51,7 @@ df = DeepFace.find(img_path = "img1.jpg", db_path = "C:/workspace/my_db")
 Deepface is a **hybrid** face recognition package. It currently wraps many **state-of-the-art** face recognition models: [`VGG-Face`](https://sefiks.com/2018/08/06/deep-face-recognition-with-keras/) , [`Google FaceNet`](https://sefiks.com/2018/09/03/face-recognition-with-facenet-in-keras/), [`OpenFace`](https://sefiks.com/2019/07/21/face-recognition-with-openface-in-keras/), [`Facebook DeepFace`](https://sefiks.com/2020/02/17/face-recognition-with-facebook-deepface-in-keras/), [`DeepID`](https://sefiks.com/2020/06/16/face-recognition-with-deepid-in-keras/), [`ArcFace`](https://sefiks.com/2020/12/14/deep-face-recognition-with-arcface-in-keras-and-python/) and [`Dlib`](https://sefiks.com/2020/07/11/face-recognition-with-dlib-in-python/). The default configuration uses VGG-Face model.
 
 ```python
-models = ["VGG-Face", "Facenet", "OpenFace", "DeepFace", "DeepID", "ArcFace", "Dlib"]
+models = ["VGG-Face", "Facenet", "Facenet512", "OpenFace", "DeepFace", "DeepID", "ArcFace", "Dlib"]
 result = DeepFace.verify("img1.jpg", "img2.jpg", model_name = models[1])
 df = DeepFace.find(img_path = "img1.jpg", db_path = "C:/workspace/my_db", model_name = models[1])
 ```

deepface/DeepFace.py

@@ -12,7 +12,7 @@ import pandas as pd
 from tqdm import tqdm
 import pickle
 
-from deepface.basemodels import VGGFace, OpenFace, Facenet, FbDeepFace, DeepID, DlibWrapper, ArcFace, Boosting
+from deepface.basemodels import VGGFace, OpenFace, Facenet, Facenet512, FbDeepFace, DeepID, DlibWrapper, ArcFace, Boosting
 from deepface.extendedmodels import Age, Gender, Race, Emotion
 from deepface.commons import functions, realtime, distance as dst
 
@@ -41,6 +41,7 @@ def build_model(model_name):
 		'VGG-Face': VGGFace.loadModel,
 		'OpenFace': OpenFace.loadModel,
 		'Facenet': Facenet.loadModel,
+		'Facenet512': Facenet512.loadModel,
 		'DeepFace': FbDeepFace.loadModel,
 		'DeepID': DeepID.loadModel,
 		'Dlib': DlibWrapper.loadModel,

deepface/basemodels/Facenet.py

@@ -3,6 +3,24 @@ from pathlib import Path
 import gdown
 from functools import partial
 
+import tensorflow as tf
+tf_version = int(tf.__version__.split(".")[0])
+
+if tf_version == 1:
+	from keras.models import Model
+	from keras.layers import Activation
+	from keras.layers import BatchNormalization
+	from keras.layers import Concatenate
+	from keras.layers import Conv2D
+	from keras.layers import Dense
+	from keras.layers import Dropout
+	from keras.layers import GlobalAveragePooling2D
+	from keras.layers import Input
+	from keras.layers import Lambda
+	from keras.layers import MaxPooling2D
+	from keras.layers import add
+	from keras import backend as K
+else:
 	from tensorflow.keras.models import Model
 	from tensorflow.keras.layers import Activation
 	from tensorflow.keras.layers import BatchNormalization
@@ -20,7 +38,7 @@ from tensorflow.keras import backend as K
 def scaling(x, scale):
 	return x * scale
 
-def InceptionResNetV2():
+def InceptionResNetV2(dimension = 128):
 
 	inputs = Input(shape=(160, 160, 3))
 	x = Conv2D(32, 3, strides=2, padding='valid', use_bias=False, name= 'Conv2d_1a_3x3') (inputs)
@@ -522,7 +540,7 @@ def InceptionResNetV2():
 	x = GlobalAveragePooling2D(name='AvgPool')(x)
 	x = Dropout(1.0 - 0.8, name='Dropout')(x)
 	# Bottleneck
-	x = Dense(128, use_bias=False, name='Bottleneck')(x)
+	x = Dense(dimension, use_bias=False, name='Bottleneck')(x)
 	x = BatchNormalization(momentum=0.995, epsilon=0.001, scale=False, name='Bottleneck_BatchNorm')(x)
 
 	# Create model

deepface/basemodels/Facenet512.py

@@ -0,0 +1,26 @@
+from deepface.basemodels import Facenet
+from pathlib import Path
+import os
+import gdown
+
+def loadModel(url = 'https://github.com/serengil/deepface_models/releases/download/v1.0/facenet512_weights.h5'):
+
+    model = Facenet.InceptionResNetV2(dimension = 512)
+
+    #-------------------------
+
+    home = str(Path.home())
+
+    if os.path.isfile(home+'/.deepface/weights/facenet512_weights.h5') != True:
+        print("facenet512_weights.h5 will be downloaded...")
+
+        output = home+'/.deepface/weights/facenet_weights.h5'
+        gdown.download(url, output, quiet=False)
+
+    #-------------------------
+
+    model.load_weights(home+'/.deepface/weights/facenet512_weights.h5')
+
+    #-------------------------
+
+    return model

deepface/commons/distance.py

@@ -29,6 +29,7 @@ def findThreshold(model_name, distance_metric):
 		'VGG-Face': {'cosine': 0.40, 'euclidean': 0.55, 'euclidean_l2': 0.75},
 		'OpenFace': {'cosine': 0.10, 'euclidean': 0.55, 'euclidean_l2': 0.55},
 		'Facenet':  {'cosine': 0.40, 'euclidean': 10, 'euclidean_l2': 0.80},
+        'Facenet512':  {'cosine': 0.3088582207770799, 'euclidean': 23.564685968740186, 'euclidean_l2': 1.0461709266148662},
 		'DeepFace': {'cosine': 0.23, 'euclidean': 64, 'euclidean_l2': 0.64},
 		'DeepID': 	{'cosine': 0.015, 'euclidean': 45, 'euclidean_l2': 0.17},
 		'Dlib': 	{'cosine': 0.07, 'euclidean': 0.6, 'euclidean_l2': 0.6},

tests/unit_tests.py

@@ -182,7 +182,7 @@ dataset = [
 
 #models = ['VGG-Face', 'Facenet', 'OpenFace', 'DeepFace', 'DeepID', 'Dlib', 'ArcFace']
 metrics = ['cosine', 'euclidean', 'euclidean_l2']
-models = ['VGG-Face', 'Facenet', 'Dlib', 'ArcFace'] #those are robust models
+models = ['VGG-Face', 'Facenet', 'Facenet512', 'Dlib', 'ArcFace'] #those are robust models
 #metrics = ['cosine']
 
 passed_tests = 0; test_cases = 0