feat: add rotated linknet_resnet18 tensorflow ckpts (#817)

* feat: add dice loss in linknet

* fix: typing

* feat: add rotated linknet_renset18 clpt

* fix: requested changes

* feat: add aspect ratio for ocr predictor

* fix: requested changes

* fix: tests

* fix: default args

* fix: new ckpts

* fix: linknet

* fix: postprocessing

* fix: isort

doctr/models/detection/linknet/base.py

@@ -129,10 +129,10 @@ def bitmap_to_boxes(
 
             if self.assume_straight_pages:
                 # compute relative polygon to get rid of img shape
+                x, y, w, h = _box
                 xmin, ymin, xmax, ymax = x / width, y / height, (x + w) / width, (y + h) / height
                 boxes.append([xmin, ymin, xmax, ymax, score])
             else:
-                _box = cv2.boxPoints(cv2.minAreaRect(contour))
                 # compute relative box to get rid of img shape
                 _box[:, 0] /= width
                 _box[:, 1] /= height

doctr/models/detection/linknet/tensorflow.py

@@ -19,7 +19,7 @@
 
 from .base import LinkNetPostProcessor, _LinkNet
 
-__all__ = ['LinkNet', 'linknet_resnet18', 'linknet_resnet34', 'linknet_resnet50']
+__all__ = ['LinkNet', 'linknet_resnet18', 'linknet_resnet34', 'linknet_resnet50', 'linknet_resnet18_rotation']
 
 
 default_cfgs: Dict[str, Dict[str, Any]] = {
@@ -29,6 +29,12 @@
         'input_shape': (1024, 1024, 3),
         'url': None,
     },
+    'linknet_resnet18_rotation': {
+        'mean': (0.798, 0.785, 0.772),
+        'std': (0.264, 0.2749, 0.287),
+        'input_shape': (1024, 1024, 3),
+        'url': 'https://github.com/mindee/doctr/releases/download/v0.5.0/linknet_resnet18-a48e6ed3.zip',
+    },
     'linknet_resnet34': {
         'mean': (0.798, 0.785, 0.772),
         'std': (0.264, 0.2749, 0.287),
@@ -286,6 +292,33 @@ def linknet_resnet18(pretrained: bool = False, **kwargs: Any) -> LinkNet:
     )
 
 
+def linknet_resnet18_rotation(pretrained: bool = False, **kwargs: Any) -> LinkNet:
+    """LinkNet as described in `"LinkNet: Exploiting Encoder Representations for Efficient Semantic Segmentation"
+    <https://arxiv.org/pdf/1707.03718.pdf>`_.
+
+    Example::
+        >>> import tensorflow as tf
+        >>> from doctr.models import linknet_resnet18
+        >>> model = linknet_resnet18(pretrained=True)
+        >>> input_tensor = tf.random.uniform(shape=[1, 1024, 1024, 3], maxval=1, dtype=tf.float32)
+        >>> out = model(input_tensor)
+
+    Args:
+        pretrained (bool): If True, returns a model pre-trained on our text detection dataset
+
+    Returns:
+        text detection architecture
+    """
+
+    return _linknet(
+        'linknet_resnet18_rotation',
+        pretrained,
+        resnet18,
+        ['resnet_block_1', 'resnet_block_3', 'resnet_block_5', 'resnet_block_7'],
+        **kwargs,
+    )
+
+
 def linknet_resnet34(pretrained: bool = False, **kwargs: Any) -> LinkNet:
     """LinkNet as described in `"LinkNet: Exploiting Encoder Representations for Efficient Semantic Segmentation"
     <https://arxiv.org/pdf/1707.03718.pdf>`_.

doctr/models/detection/zoo.py

@@ -15,8 +15,8 @@
 
 
 if is_tf_available():
-    ARCHS = ['db_resnet50', 'db_mobilenet_v3_large', 'linknet_resnet18']
-    ROT_ARCHS = []
+    ARCHS = ['db_resnet50', 'db_mobilenet_v3_large', 'linknet_resnet18', 'linknet_resnet18_rotation']
+    ROT_ARCHS = ['linknet_resnet18_rotation']
 elif is_torch_available():
     ARCHS = ['db_resnet34', 'db_resnet50', 'db_mobilenet_v3_large', 'linknet_resnet18', 'db_resnet50_rotation']
     ROT_ARCHS = ['db_resnet50_rotation']