diff --git a/monai/transforms/spatial/array.py b/monai/transforms/spatial/array.py
index 9e78e18f85..dcbc7aa2f6 100644
--- a/monai/transforms/spatial/array.py
+++ b/monai/transforms/spatial/array.py
@@ -14,7 +14,6 @@
 """
 
 import warnings
-from math import ceil
 from typing import Any, List, Optional, Sequence, Tuple, Union
 
 import numpy as np
@@ -403,7 +402,7 @@ def __call__(
             if not isinstance(self.spatial_size, int):
                 raise ValueError("spatial_size must be an int number if size_mode is 'longest'.")
             scale = self.spatial_size / max(img_size)
-            spatial_size_ = tuple(ceil(s * scale) for s in img_size)
+            spatial_size_ = tuple(int(round(s * scale)) for s in img_size)
         resized = torch.nn.functional.interpolate(  # type: ignore
             input=torch.as_tensor(np.ascontiguousarray(img), dtype=torch.float).unsqueeze(0),
             size=spatial_size_,
diff --git a/tests/test_resize.py b/tests/test_resize.py
index 2f54dcc04f..e5ec5dd1a9 100644
--- a/tests/test_resize.py
+++ b/tests/test_resize.py
@@ -18,11 +18,11 @@
 from monai.transforms import Resize
 from tests.utils import NumpyImageTestCase2D
 
-TEST_CASE_0 = [{"spatial_size": 15}, (6, 11, 15)]
+TEST_CASE_0 = [{"spatial_size": 15}, (6, 10, 15)]
 
-TEST_CASE_1 = [{"spatial_size": 15, "mode": "area"}, (6, 11, 15)]
+TEST_CASE_1 = [{"spatial_size": 15, "mode": "area"}, (6, 10, 15)]
 
-TEST_CASE_2 = [{"spatial_size": 6, "mode": "trilinear", "align_corners": True}, (3, 5, 6)]
+TEST_CASE_2 = [{"spatial_size": 6, "mode": "trilinear", "align_corners": True}, (2, 4, 6)]
 
 
 class TestResize(NumpyImageTestCase2D):
@@ -63,6 +63,11 @@ def test_longest_shape(self, input_param, expected_shape):
         result = Resize(**input_param)(input_data)
         np.testing.assert_allclose(result.shape[1:], expected_shape)
 
+    def test_longest_infinite_decimals(self):
+        resize = Resize(spatial_size=1008, size_mode="longest", mode="bilinear", align_corners=False)
+        ret = resize(np.random.randint(0, 2, size=[1, 2544, 3032]))
+        self.assertTupleEqual(ret.shape, (1, 846, 1008))
+
 
 if __name__ == "__main__":
     unittest.main()
diff --git a/tests/test_resized.py b/tests/test_resized.py
index 6c4f31c9c8..930faf00eb 100644
--- a/tests/test_resized.py
+++ b/tests/test_resized.py
@@ -18,15 +18,15 @@
 from monai.transforms import Resized
 from tests.utils import NumpyImageTestCase2D
 
-TEST_CASE_0 = [{"keys": "img", "spatial_size": 15}, (6, 11, 15)]
+TEST_CASE_0 = [{"keys": "img", "spatial_size": 15}, (6, 10, 15)]
 
-TEST_CASE_1 = [{"keys": "img", "spatial_size": 15, "mode": "area"}, (6, 11, 15)]
+TEST_CASE_1 = [{"keys": "img", "spatial_size": 15, "mode": "area"}, (6, 10, 15)]
 
-TEST_CASE_2 = [{"keys": "img", "spatial_size": 6, "mode": "trilinear", "align_corners": True}, (3, 5, 6)]
+TEST_CASE_2 = [{"keys": "img", "spatial_size": 6, "mode": "trilinear", "align_corners": True}, (2, 4, 6)]
 
 TEST_CASE_3 = [
     {"keys": ["img", "label"], "spatial_size": 6, "mode": ["trilinear", "nearest"], "align_corners": [True, None]},
-    (3, 5, 6),
+    (2, 4, 6),
 ]