diff --git a/horizon/test/unit/utils/test_babel_extract_angular.py b/horizon/test/unit/utils/test_babel_extract_angular.py
index f09b62879e..7aacf8ad2a 100644
--- a/horizon/test/unit/utils/test_babel_extract_angular.py
+++ b/horizon/test/unit/utils/test_babel_extract_angular.py
@@ -135,6 +135,15 @@ class ExtractAngularTestCase(test.TestCase):
             <img alt="{$'some other thing'$}">
             <p>{$'"it\\'s awesome"'|translate$}</p>
             <p>{$"oh \\"hello\\" there"|translate$}</p>
+            <img alt="{$::'hello colon1' | translate $}">
+            <p>{$ ::'hello colon2' |translate$}</p>
+            <p>{$ :: 'hello colon3'| translate$}</p>
+            <img alt="something {$::'hello colon4'|translate$} something
+            {$ ::'hello colon5' | translate$}">
+            <img alt="{::$expr()|translate$}">
+            <img alt="{$::'some other thing'$}">
+            <p>{$:: '"it\\'s awesome"'|translate$}</p>
+            <p>{$ :: "oh \\"hello\\" there" | translate$}</p>
             """
         )
 
@@ -147,6 +156,13 @@ class ExtractAngularTestCase(test.TestCase):
                 (4, u'gettext', 'hello world4', []),
                 (8, u'gettext', '"it\\\'s awesome"', []),
                 (9, u'gettext', 'oh \\"hello\\" there', []),
+                (10, u'gettext', u'hello colon1', []),
+                (11, u'gettext', u'hello colon2', []),
+                (12, u'gettext', u'hello colon3', []),
+                (13, u'gettext', u'hello colon4', []),
+                (13, u'gettext', u'hello colon5', []),
+                (17, u'gettext', u'"it\\\'s awesome"', []),
+                (18, u'gettext', u'oh \\"hello\\" there', []),
             ],
             messages)
 
diff --git a/horizon/utils/babel_extract_angular.py b/horizon/utils/babel_extract_angular.py
index cd9ba5e274..4d06c271ef 100644
--- a/horizon/utils/babel_extract_angular.py
+++ b/horizon/utils/babel_extract_angular.py
@@ -21,7 +21,7 @@ from six.moves import html_parser
 
 # regex to find filter translation expressions
 filter_regex = re.compile(
-    r"""{\$\s*('([^']|\\')+'|"([^"]|\\")+")\s*\|\s*translate\s*\$}"""
+    r"""{\$\s*(::)?\s*('([^']|\\')+'|"([^"]|\\")+")\s*\|\s*translate\s*\$}"""
 )
 
 # browser innerHTML decodes some html entities automatically, so when
@@ -49,6 +49,8 @@ class AngularGettextHTMLParser(html_parser.HTMLParser):
     {$ 'content' | translate $}
         The string will be translated, minus expression handling (i.e. just
         bare strings are allowed.)
+    {$ ::'content' | translate $}
+        The string will be translated. As above.
     """
 
     def __init__(self):
@@ -94,7 +96,7 @@ class AngularGettextHTMLParser(html_parser.HTMLParser):
                 for match in filter_regex.findall(attr[1]):
                     if match:
                         self.strings.append(
-                            (self.line, u'gettext', match[0][1:-1], [])
+                            (self.line, u'gettext', match[1][1:-1], [])
                         )
 
     def handle_data(self, data):
@@ -103,7 +105,7 @@ class AngularGettextHTMLParser(html_parser.HTMLParser):
         else:
             for match in filter_regex.findall(data):
                 self.strings.append(
-                    (self.line, u'gettext', match[0][1:-1], [])
+                    (self.line, u'gettext', match[1][1:-1], [])
                 )
 
     def handle_entityref(self, name):