From e27971126e88692fba1384500bf22a2d4101622e Mon Sep 17 00:00:00 2001
From: Tiago Freitas Pereira <tiagofrepereira@gmail.com>
Date: Wed, 9 Jun 2021 15:20:42 +0200
Subject: [PATCH] Added new baseline
---
.../resnet50_msceleb_arcface_20210521.py | 15 +++
bob/bio/face/embeddings/tensorflow.py | 113 +++++++++++++++---
setup.py | 2 +
3 files changed, 116 insertions(+), 14 deletions(-)
create mode 100644 bob/bio/face/config/baseline/resnet50_msceleb_arcface_20210521.py
diff --git a/bob/bio/face/config/baseline/resnet50_msceleb_arcface_20210521.py b/bob/bio/face/config/baseline/resnet50_msceleb_arcface_20210521.py
new file mode 100644
index 00000000..b54d2caf
--- /dev/null
+++ b/bob/bio/face/config/baseline/resnet50_msceleb_arcface_20210521.py
@@ -0,0 +1,15 @@
+from bob.bio.face.embeddings.tensorflow import resnet50_msceleb_arcface_20210521
+from bob.bio.face.utils import lookup_config_from_database
+
+annotation_type, fixed_positions, memory_demanding = lookup_config_from_database(
+ locals().get("database")
+)
+
+
+def load(annotation_type, fixed_positions=None, memory_demanding=None):
+ return resnet50_msceleb_arcface_20210521(
+ annotation_type, fixed_positions, memory_demanding
+ )
+
+
+pipeline = load(annotation_type, fixed_positions, memory_demanding)
diff --git a/bob/bio/face/embeddings/tensorflow.py b/bob/bio/face/embeddings/tensorflow.py
index d37243b1..d103f19c 100644
--- a/bob/bio/face/embeddings/tensorflow.py
+++ b/bob/bio/face/embeddings/tensorflow.py
@@ -314,15 +314,15 @@ class Resnet50_MsCeleb_ArcFace_2021(TensorflowTransformer):
def __init__(self, memory_demanding=False):
urls = [
- "https://www.idiap.ch/software/bob/data/bob/bob.bio.face/master/tf2/resnet50_msceleb_arcface_2021.tar.gz",
- "http://www.idiap.ch/software/bob/data/bob/bob.bio.face/master/tf2/resnet50_msceleb_arcface_2021.tar.gz",
+ "https://www.idiap.ch/software/bob/data/bob/bob.bio.face/master/tf2/resnet50-msceleb-arcface_2021-48ec5cb8.tar.gz",
+ "http://www.idiap.ch/software/bob/data/bob/bob.bio.face/master/tf2/resnet50-msceleb-arcface_2021-48ec5cb8.tar.gz",
]
filename = get_file(
- "resnet50_msceleb_arcface_2021.tar.gz",
+ "resnet50-msceleb-arcface_2021-48ec5cb8.tar.gz",
urls,
- cache_subdir="data/tensorflow/resnet50_msceleb_arcface_2021",
- file_hash="1e4b9791669ef79cf8ed80a6fc830205",
+ cache_subdir="data/tensorflow/resnet50-msceleb-arcface_2021-48ec5cb8",
+ file_hash="17946f121af5ddd18c637c4620e54da6",
extract=True,
)
checkpoint_path = os.path.dirname(filename)
@@ -333,13 +333,71 @@ class Resnet50_MsCeleb_ArcFace_2021(TensorflowTransformer):
memory_demanding=memory_demanding,
)
- def inference(self, X):
- if self.preprocessor is not None:
- X = self.preprocessor(tf.cast(X, "float32"))
- prelogits = self.model.predict_on_batch(X)[0]
- embeddings = tf.math.l2_normalize(prelogits, axis=-1)
- return embeddings
+class Resnet50_MsCeleb_ArcFace_20210521(TensorflowTransformer):
+ """
+ Resnet50 Backbone trained with the MSCeleb 1M database. The bottleneck layer (a.k.a embedding) has 512d.
+
+ The difference from this one to :any:`Resnet50_MsCeleb_ArcFace_2021` is the MSCeleb version used to train it.
+ This one uses 100% of the data pruned from annotators.
+
+
+ The configuration file used to trained is:
+
+ .. warning::
+ This configuration file might change in future releases
+
+
+ ```yaml
+ batch-size: 128
+ face-size: 112
+ face-output_size: 112
+ n-classes: 83009
+
+
+ ## Backbone
+ backbone: 'resnet50'
+ head: 'arcface'
+ s: 30
+ bottleneck: 512
+ m: 0.5
+
+ # Training parameters
+ solver: "sgd"
+ lr: 0.1
+ dropout-rate: 0.5
+ epochs: 300
+
+
+ train-tf-record-path: "<PATH>"
+ validation-tf-record-path: "<PATH>"
+
+ ```
+
+
+ """
+
+ def __init__(self, memory_demanding=False):
+
+ urls = [
+ "https://www.idiap.ch/software/bob/data/bob/bob.bio.face/master/tf2/resnet50-msceleb-arcface_20210521-e9bc085c.tar.gz",
+ "http://www.idiap.ch/software/bob/data/bob/bob.bio.face/master/tf2/resnet50-msceleb-arcface_20210521-e9bc085c.tar.gz",
+ ]
+
+ filename = get_file(
+ "resnet50-msceleb-arcface_20210521-e9bc085c.tar.gz",
+ urls,
+ cache_subdir="data/tensorflow/resnet50-msceleb-arcface_20210521-801991f0",
+ file_hash="e33090eea4951ce80be4620a0dac680d",
+ extract=True,
+ )
+ checkpoint_path = os.path.dirname(filename)
+
+ super(Resnet50_MsCeleb_ArcFace_20210521, self).__init__(
+ checkpoint_path,
+ preprocessor=lambda X: X / 255.0,
+ memory_demanding=memory_demanding,
+ )
class Resnet50_VGG2_ArcFace_2021(TensorflowTransformer):
@@ -549,7 +607,7 @@ def resnet50_msceleb_arcface_2021(
annotation_type, fixed_positions=None, memory_demanding=False
):
"""
- Get the Resnet50 pipeline which will crop the face :math:`160 \times 160` and
+ Get the Resnet50 pipeline which will crop the face :math:`112 \times 112` and
use the :py:class:`Resnet50_MsCeleb_ArcFace_2021` to extract the features
Parameters
@@ -572,11 +630,38 @@ def resnet50_msceleb_arcface_2021(
)
+def resnet50_msceleb_arcface_20210521(
+ annotation_type, fixed_positions=None, memory_demanding=False
+):
+ """
+ Get the Resnet50 pipeline which will crop the face :math:`112 \times 112` and
+ use the :py:class:`Resnet50_MsCeleb_ArcFace_20210521` to extract the features
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return resnet_template(
+ embedding=Resnet50_MsCeleb_ArcFace_20210521(memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
def resnet50_vgg2_arcface_2021(
annotation_type, fixed_positions=None, memory_demanding=False
):
"""
- Get the Resnet50 pipeline which will crop the face :math:`160 \times 160` and
+ Get the Resnet50 pipeline which will crop the face :math:`112 \times 112` and
use the :py:class:`Resnet50_VGG2_ArcFace_2021` to extract the features
Parameters
@@ -603,7 +688,7 @@ def mobilenetv2_msceleb_arcface_2021(
annotation_type, fixed_positions=None, memory_demanding=False
):
"""
- Get the MobileNet pipeline which will crop the face :math:`160 \times 160` and
+ Get the MobileNet pipeline which will crop the face :math:`112 \times 112` and
use the :py:class:`MobileNetv2_MsCeleb_ArcFace_2021` to extract the features
Parameters
diff --git a/setup.py b/setup.py
index 12eb3529..f5e5c8e9 100644
--- a/setup.py
+++ b/setup.py
@@ -134,6 +134,7 @@ setup(
"lda = bob.bio.face.config.baseline.lda:pipeline",
"dummy = bob.bio.face.config.baseline.dummy:pipeline",
"resnet50-msceleb-arcface-2021 = bob.bio.face.config.baseline.resnet50_msceleb_arcface_2021:pipeline",
+ "resnet50-msceleb-arcface-20210521 = bob.bio.face.config.baseline.resnet50_msceleb_arcface_20210521:pipeline",
"resnet50-vgg2-arcface-2021 = bob.bio.face.config.baseline.resnet50_vgg2_arcface_2021:pipeline",
"mobilenetv2-msceleb-arcface-2021 = bob.bio.face.config.baseline.mobilenetv2_msceleb_arcface_2021",
"mxnet-tinyface = bob.bio.face.config.baseline.mxnet_tinyface:pipeline",
@@ -176,6 +177,7 @@ setup(
"casia-africa = bob.bio.face.config.database.casia_africa",
"morph = bob.bio.face.config.database.morph",
"resnet50-msceleb-arcface-2021 = bob.bio.face.config.baseline.resnet50_msceleb_arcface_2021",
+ "resnet50-msceleb-arcface-20210521 = bob.bio.face.config.baseline.resnet50_msceleb_arcface_20210521:pipeline",
"resnet50-vgg2-arcface-2021 = bob.bio.face.config.baseline.resnet50_vgg2_arcface_2021",
"mobilenetv2-msceleb-arcface-2021 = bob.bio.face.config.baseline.mobilenetv2_msceleb_arcface_2021",
"iresnet34 = bob.bio.face.config.baseline.iresnet34",
--
GitLab