diff --git a/chainercv/evaluations/eval_semantic_segmentation.py b/chainercv/evaluations/eval_semantic_segmentation.py
index 8e18d5abc8..397e3e1f26 100644
--- a/chainercv/evaluations/eval_semantic_segmentation.py
+++ b/chainercv/evaluations/eval_semantic_segmentation.py
@@ -118,14 +118,12 @@ class :math:`j` by the prediction.
         \\frac \
         {\\sum_{i=1}^k N_{ii}} \
         {\\sum_{i=1}^k \\sum_{j=1}^k N_{ij}}`
-    * :math:`\\text{Accuracy} = \
+    * :math:`\\text{Class Accuracy} = \
         \\frac{N_{ii}}{\\sum_{j=1}^k N_{ij}}`
-    * :math:`\\text{Mean Accuracy} = \\frac{1}{k} \
+    * :math:`\\text{Mean Class Accuracy} = \\frac{1}{k} \
         \\sum_{i=1}^k \
         \\frac{N_{ii}}{\\sum_{j=1}^k N_{ij}}`
 
-    mIoU can be computed by taking :obj:`numpy.nanmean` of the IoUs returned
-    by this function.
     The more detailed description of the above metric can be found in a
     review on semantic segmentation [#]_.
 
@@ -162,9 +160,10 @@ class :math:`j` by the prediction.
             :math:`n\_class` classes. Its shape is :math:`(n\_class,)`.
         * **miou** (*float*): The average of IoUs over classes.
         * **pixel_accuracy** (*float*): The computed pixel accuracy.
-        * **accuracy** (*numpy.ndarray*): An array of accuracies for the \
-            :math:`n\_class` classes. Its shape is :math:`(n\_class,)`.
-        * **mean_accuracy** (*float*): The average of accuracies.
+        * **class_accuracy** (*numpy.ndarray*): An array of class accuracies \
+            for the :math:`n\_class` classes. \
+            Its shape is :math:`(n\_class,)`.
+        * **class_mean_accuracy** (*float*): The average of class accuracies.
 
     """
     # Evaluation code is based on
@@ -174,8 +173,9 @@ class :math:`j` by the prediction.
         pred_labels, gt_labels)
     iou = calc_semantic_segmentation_iou(confusion)
     pixel_accuracy = np.diag(confusion).sum() / confusion.sum()
-    accuracy = np.diag(confusion) / np.sum(confusion, axis=1)
+    class_accuracy = np.diag(confusion) / np.sum(confusion, axis=1)
 
     return {'iou': iou, 'miou': np.nanmean(iou),
             'pixel_accuracy': pixel_accuracy,
-            'accuracy': accuracy, 'mean_accuracy': np.nanmean(accuracy)}
+            'class_accuracy': class_accuracy,
+            'mean_class_accuracy': np.nanmean(class_accuracy)}
diff --git a/tests/evaluations_tests/test_eval_semantic_segmentation.py b/tests/evaluations_tests/test_eval_semantic_segmentation.py
index 0c614f2aa9..a0d626523f 100644
--- a/tests/evaluations_tests/test_eval_semantic_segmentation.py
+++ b/tests/evaluations_tests/test_eval_semantic_segmentation.py
@@ -16,13 +16,13 @@
      'gt_labels': iter(np.repeat([[[1, 0, 0], [0, -1, 1]]], 2, axis=0)),
      'iou': np.array([4 / 6, 4 / 6]),
      'pixel_accuracy': 4 / 5,
-     'accuracy': np.array([2 / 3, 2 / 2]),
+     'class_accuracy': np.array([2 / 3, 2 / 2]),
      },
     {'pred_labels': np.array([[[0, 0, 0], [0, 0, 0]]]),
      'gt_labels': np.array([[[1, 1, 1], [1, 1, 1]]]),
      'iou': np.array([0, 0]),
      'pixel_accuracy': 0 / 6,
-     'accuracy': np.array([np.nan, 0])
+     'class_accuracy': np.array([np.nan, 0])
      }
 )
 class TestEvalSemanticSegmentation(unittest.TestCase):
@@ -32,11 +32,11 @@ def test_eval_semantic_segmentation(self):
             self.pred_labels, self.gt_labels)
         np.testing.assert_equal(result['iou'], self.iou)
         np.testing.assert_equal(result['pixel_accuracy'], self.pixel_accuracy)
-        np.testing.assert_equal(result['accuracy'], self.accuracy)
+        np.testing.assert_equal(result['class_accuracy'], self.class_accuracy)
 
         np.testing.assert_equal(result['miou'], np.nanmean(self.iou))
         np.testing.assert_equal(
-            result['mean_accuracy'], np.nanmean(self.accuracy))
+            result['mean_class_accuracy'], np.nanmean(self.class_accuracy))
 
 
 class TestCalcSemanticSegmentationConfusion(unittest.TestCase):