diff --git a/doctr/models/preprocessor/tensorflow.py b/doctr/models/preprocessor/tensorflow.py
index b2321654ce..9d777d623a 100644
--- a/doctr/models/preprocessor/tensorflow.py
+++ b/doctr/models/preprocessor/tensorflow.py
@@ -106,7 +106,9 @@ def __call__(self, x: Union[tf.Tensor, np.ndarray, List[Union[tf.Tensor, np.ndar
                 x = tf.image.convert_image_dtype(x, dtype=tf.float32)
             # Resizing
             if (x.shape[1], x.shape[2]) != self.resize.output_size:
-                x = tf.image.resize(x, self.resize.output_size, method=self.resize.method)
+                x = tf.image.resize(
+                    x, self.resize.output_size, method=self.resize.method, antialias=self.resize.antialias
+                )
 
             batches = [x]
 
diff --git a/doctr/transforms/modules/tensorflow.py b/doctr/transforms/modules/tensorflow.py
index 7d4a4c7731..1cf50480e9 100644
--- a/doctr/transforms/modules/tensorflow.py
+++ b/doctr/transforms/modules/tensorflow.py
@@ -85,6 +85,7 @@ def __init__(
         self.method = method
         self.preserve_aspect_ratio = preserve_aspect_ratio
         self.symmetric_pad = symmetric_pad
+        self.antialias = True
 
         if isinstance(self.output_size, int):
             self.wanted_size = (self.output_size, self.output_size)
@@ -106,7 +107,7 @@ def __call__(
     ) -> Union[tf.Tensor, Tuple[tf.Tensor, np.ndarray]]:
         input_dtype = img.dtype
 
-        img = tf.image.resize(img, self.wanted_size, self.method, self.preserve_aspect_ratio)
+        img = tf.image.resize(img, self.wanted_size, self.method, self.preserve_aspect_ratio, self.antialias)
         # It will produce an un-padded resized image, with a side shorter than wanted if we preserve aspect ratio
         raw_shape = img.shape[:2]
         if self.preserve_aspect_ratio:
diff --git a/tests/tensorflow/test_models_preprocessor_tf.py b/tests/tensorflow/test_models_preprocessor_tf.py
index 96362a5427..941c803229 100644
--- a/tests/tensorflow/test_models_preprocessor_tf.py
+++ b/tests/tensorflow/test_models_preprocessor_tf.py
@@ -40,5 +40,5 @@ def test_preprocessor(batch_size, output_size, input_tensor, expected_batches, e
     assert all(isinstance(b, tf.Tensor) for b in out)
     assert all(b.dtype == tf.float32 for b in out)
     assert all(b.shape[1:3] == output_size for b in out)
-    assert all(tf.math.reduce_all(b == expected_value) for b in out)
+    assert all(tf.math.reduce_all(tf.math.abs(b - expected_value) < 1e-6) for b in out)
     assert len(repr(processor).split("\n")) == 4
diff --git a/tests/tensorflow/test_transforms_tf.py b/tests/tensorflow/test_transforms_tf.py
index bc801f3ddc..6b7e3affbd 100644
--- a/tests/tensorflow/test_transforms_tf.py
+++ b/tests/tensorflow/test_transforms_tf.py
@@ -14,7 +14,7 @@ def test_resize():
     input_t = tf.cast(tf.fill([64, 64, 3], 1), dtype=tf.float32)
     out = transfo(input_t)
 
-    assert tf.reduce_all(out == 1)
+    assert tf.math.reduce_all(tf.math.abs(out - 1) < 1e-6)
     assert out.shape[:2] == output_size
     assert repr(transfo) == f"Resize(output_size={output_size}, method='bilinear')"
 
@@ -24,7 +24,7 @@ def test_resize():
 
     assert not tf.reduce_all(out == 1)
     # Asymetric padding
-    assert tf.reduce_all(out[-1] == 0) and tf.reduce_all(out[0] == 1)
+    assert tf.reduce_all(out[-1] == 0) and tf.math.reduce_all(tf.math.abs(out[0] - 1) < 1e-6)
     assert out.shape[:2] == output_size
 
     # Symetric padding